Development of a Series of Kynurenine 3-Monooxygenase Inhibitors Leading to a Clinical Candidate for the Treatment of Acute Pancreatitis

Article abstract of DOI:10.1021/acs.jmedchem.7b00055

Recently, we reported a novel role for KMO in the pathogenesis of acute pancreatitis (AP). A number of inhibitors of kynurenine 3-monooxygenase (KMO) have previously been described as potential treatments for neurodegenerative conditions and particularly for Huntington’s disease. However, the inhibitors reported to date have insufficient aqueous solubility relative to their cellular potency to be compatible with the intravenous (iv) dosing route required in AP. We have identified and optimized a novel series of high affinity KMO inhibitors with favorable physicochemical properties. The leading example is exquisitely selective, has low clearance in two species, prevents lung and kidney damage in a rat model of acute pancreatitis, and is progressing into preclinical development.

Full text of DOI:10.1021/acs.jmedchem.7b00055

Subscriber access provided by University of Newcastle, Australia
Article
Development of a series of Kynurenine 3-Monooxygenase Inhibitors
leading to a Clinical Candidate for the Treatment of Acute Pancreatitis
Ann L. Walker, Nicolas Ancellin, Benjamin Beaufils, Marylise Bergeal, Margaret Binnie,
David Clapham, Carl P. Haslam, Duncan S. Holmes, Jonathan P. Hutchinson, John
Liddle, Andrew McBride, Olivier Mirguet, Christopher G. Mowat, Paul Rowland, Nathalie
Tiberghien, Lionel Trottet, Iain Uings, Scott P Webster, Xiaozhong Zheng, and Damian J. Mole
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.7b00055 • Publication Date (Web): 28 Mar 2017
Downloaded from http://pubs.acs.org on March 30, 2017
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 80
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Development of a series of Kynurenine 3ꢀ
Monooxygenase Inhibitors leading to a Clinical
Candidate for the Treatment of Acute Pancreatitis
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
a
b
b
c
Ann L. Walker *, Nicolas Ancellin , Benjamin Beaufils , Marylise Bergeal , Margaret
d
c
c
a
c
Binnie , David Clapham , Carl P. Haslam , Duncan S. Holmes , Jonathan P. Hutchinson ,
a
d
b
e
c
John Liddle , Andrew McBride , Olivier Mirguet , Christopher G. Mowat , Paul Rowland ,
b
b
a
d
f
Nathalie Tiberghien , Lionel Trottet , Iain Uings , Scott P. Webster , Xiaozhong Zheng ,
f
Damian J. Mole .
a) Discovery Partnerships with Academia, GSK: b) Flexible Discovery Unit, GSK: c)
Platform Technology Sciences, GSK: d) Centre for Cardiovascular Science,
University of Edinburgh: e) EastChem School of Chemistry, University of Edinburgh:
f) MRC Centre for Inflammation Research, University of Edinburgh.
* Corresponding author
RECEIVED DATE (to be automatically inserted after your manuscript is accepted if
required according to the journal that you are submitting your paper to)
ABSTRACT
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 80
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Recently, we reported a novel role for KMO in the pathogenesis of acute pancreatitis (AP). A
number of inhibitors of kynurenine 3ꢀmonooxygenase (KMO) have previously been
described as potential treatments for neurodegenerative conditions and particularly for
Huntington’s disease. However the inhibitors reported to date have insufficient aqueous
solubility relative to their cellular potency to be compatible with the intravenous (i.v.) dosing
route required in AP. We have identified and optimised a novel series of high affinity KMO
inhibitors with favourable physicochemical properties. The leading example is exquisitely
selective, has low clearance in two species, prevents lung and kidney damage in a rat model
of acute pancreatitis and is progressing into preclinical development.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
INTRODUCTION
Acute pancreatitis (AP) is an inflammatory disease of the pancreas that is usually triggered by
gallstones or excessive alcohol consumption. In the majority of sufferers, symptoms resolve
within a few days, but for approximately 20% of patients with AP, the disease progresses to
multiple organ dysfunction syndrome (APꢀMODS) requiring treatment in intensive care. The
1
individual case fatality rate in APꢀMODS is 21%. Currently, there is no specific therapy for
APꢀMODS and treatment is solely supportive. Until now, no diseaseꢀmodifying treatments
for APꢀMODS have been identified and there are no specific treatments in routine clinical
practice. Therefore, it is clear that AP represents a significant unmet medical need. We have
2
recently demonstrated a key role for KMO in the pathogenesis of APꢀMODS. In
experimental models of AP, mice that were genetically engineered to have absent KMO
activity, and rats treated with a small molecule KMO inhibitor tool compound, were
2
protected from APꢀMODS.
ACS Paragon Plus Environment

Page 3 of 80
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
AP patients are treated in an intensive care setting and therefore an intravenous dosing route
is strongly preferred. Where a fast onset of treatment is desirable, as it is in AP, initial drug
loading with a bolus dose is advantageous to achieve the target drug plasma concentration
rapidly. This plasma concentration may then be maintained by infusion dosing for as long as
required. The initial bolus dose volume is limited in practise, to 50 mL, which in turn
requires that the drug have high cellular potency, a significant free fraction and good
solubility in aqueous solution at a biologically acceptable pH in order to achieve a
therapeutically relevant free drug plasma concentration in vivo.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For reasons of storage and dosing the stability of the compound in solution is also important,
with a necessity for excellent solution stability at relevant pHs and to light exposure.
3
ꢀ10
A number of series of KMO inhibitors have been described including most recently those
3
from CHDI. The majority of the molecules described share a common pharmacophore
containing both a carboxylic acid / acidic moiety and a mono or 1,2ꢀdichloro substitution of
2
,11
the core phenyl ring. In our previous papers , we have shown that one such molecule,
2
3
‐(5,6‐dichloro‐2‐oxo‐2,3‐dihydro‐1,3ꢀbenzoxazol‐3‐yl)propanoic acid
1
(GSK180 )
2
protected animals from secondary organ dysfunction in a rat model of experimental AP.
However, the cellular potency of 1 was relatively weak so high exposures were required to
inhibit KMO activity in vivo resulting in effects on the pathway that were not related to KMO
inhibition. Moreover, while 1 has solubility in excess of 0.2mg/ml (Table 4), the compound
showed significant instability precluding further development of this compound for an i.v.
dosing modality.
We therefore sought to expand our existing series to discover better molecules that have
optimised characteristics that will accelerate the translation of KMO inhibitors towards the
clinic.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 80
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
RESULTS AND DISCUSSION
We first prepared a number of other analogues from the chemical series represented by 1;
1
1
3
‐(5‐chloro‐6‐ethyl‐2‐oxo‐2,3‐dihydro‐1,3ꢀbenzoxazol‐3‐yl)propanoic acid 2 (GSK428 )
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
12
retained potency in our KMO enzymatic activity assay , had improved stability compared
with the 5,6ꢀdichloro example and had excellent aqueous solubility as the TRIS salt (>120
mg/mL). Further expansion of the 6ꢀmethyl group to ethyl 3 or isoꢀbutyl 4 was also well
tolerated, as was replacement of the methyl by a small neutral alkoxy group 5 – 10; however
basic side chains, for example, compound 11 significantly reduced KMO inhibitory activity.
2
During these changes, the activity in a cell based assay of KMO activity remained at c.a.
1ꢁM IC50, except for more polar side chains as in 10 where potency in the cellular assay
declined. The artificial membrane permeability for the compounds was low, typically at
around or below the lower level of quantification; this was therefore not a reliable
discriminator of cellular behaviour. We also considered other predictors of permeability; the
number of hydrogen bond donors and acceptors are largely unchanged for these molecules, as
is the pKa and the molecular weights of the molecules are within a narrow distribution range.
1
3
However, the lipophilicity (chromLogD ) did correlate reasonably with the relative cellular
activity for these and other small halogenated, alkyl and alkoxy examples and also with the
14
human serum albumin (HSA) binding ; suggesting both HSA binding and passage of the
compounds into the cell were largely driven by lipophilicity (Figure 1). This is a potentially
problematic scenario in an in vivo situation as any reduction in the target plasma free
concentration due to increased cellular potency is likely be offꢀset by increased the protein
binding so that the total plasma drug concentration required is not reduced.
a)
ACS Paragon Plus Environment

Page 5 of 80
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
b)
Figure 1 a) Dropꢀoff from human KMOenzyme to human KMOcell pIC50 plotted against the
Chrom LogD for examples of the series with a small 6ꢀsubstituent including compounds 1 ꢀ
12. b) % binding to HSA compared to Chrom LogD for the same compound set.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 80
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Addition of a benzyloxy group at the 6ꢀposition as in 12 followed previous trends with
modest cell activity and high HSA binding (98.3%), however its 2ꢀpyridylmethoxy analogue
1
3 showed increased KMO potency (pIC 8.5, which is close to the tight binding limit of the
50
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
assay), but with a significantly lower level of plasma protein binding (HSA 87.4%).
However, this increased potency did not translate into potency in the cellular assay with 13
having slightly weaker cellular activity than the more lipophilic 12.
Surprisingly addition of a methyl group onto the benzylic position of 13 to give 14 resulted in
a substantial (~30x) jump in cellular potency (pIC 6.9) and a higher free drug fraction than
50
would be expected from the logD based on the previous trends. This substantial increase in
2
cell potency was confirmed in primary human hepatocytes (pIC50 7.0).
Eg.
No.
R
Salt KMO KMO HEK chro
HS hPPB
AM
mean
pIC50
LE
mean mlog
cell
pIC50 pH7.4
n)
A
(%) permeab
ility (nm
/LLE
D
(%)
(n)
/ sec )
(
1
2
Cl
8.2
6)
7.9
0.66 /
5.97
5.7
(4)
6.2
1.61 96.3 98.2
24
(
Me
Tris
0.64 /
1.26 95.0 97.6
<3 ꢀ <10
ACS Paragon Plus Environment

Page 7 of 80
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
(13)
8.0
5.76
(3)
(free
acid)
3
4
5
6
7
8
9
1
1
1
Et
Tris
Tris
ꢀ
0.61 /
5.33
6.5
(4)
6.1
(3)
5.3
(2)
6.4
(7)
5.9
(8)
5.8
(11)
5.9
(2)
4.3
(5)
nd
1.94 97.4 98.7
<10ꢀ22
<10
<3
(
7)
7.9
5)
7.9
8)
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
iBu
0.54 /
4.31
3.03 99.2
nd
nd
(
OMe
OEt
OcPr
OCH
OCH
0.60 /
6.43
nd
89.0
(
Tris
Tris
Tris
Tris
Tris
8.3
(5)
8.5
0.60 /
6.30
1.49 88.7 95.8
1.64 94.3 96.9
2.09 96.2 96.0
2.90 98.1 98.1
0.88 44.2 70.7
<10
15
0.58 /
6.45
(5)
2
cPr
8.0
(5)
7.6
0.52 /
5.56
19
2
cBu
0.47 /
4.60
12
(2)
0
1
2
O(CH
O(CH
2
)
)
2
OMe
7.5
(5)
6.1
(5)
7.6
0.49 /
6.14
<3
2
2
pyrrolid Tris
Tris
0.35 /
6.13
ꢀ0.26 15.1
2.71 98.3
nd
nd
<3
inꢀ1ꢀyl
OBn
0.43 /
4.37
5.8
(2)
nd
(
5)
8.5
7)
8.5
13
OCH
2
pyridꢀ2ꢀyl Tris
0.49 /
6.76
5.3
(11)
6.9
1.32 87.4 92.6
1.86 87.1 81.7
11
(
14
OCH(Me)pyridꢀ Tris
0.47 /
<3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 80
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
2ꢀyl
(7)
8.6
6.45
0.47 /
6.55
(9)
6.7
1
5
(R)ꢀ
Tris
Tris
Tris
Tris
ꢀ
1.71 93.3 88.1
<3ꢀ20
<3
OCH(Me)pyridꢀ
(12)
(15)
2
ꢀyl
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
16
(S)ꢀ
6.6
(7)
0.36 /
4.55
4.2
(3)
1.79 91.6
1.11 77.3
0.83 59.1
nd
nd
nd
OCH(Me)pyridꢀ
2
ꢀyl
1
7
8
(R)ꢀ
8.5
(5)
0.49 /
7.28
5.1
(3)
<3
OCH(Me)oxazol
ꢀ2ꢀyl
1
(R)ꢀ
8.3
(5)
0.45 /
7.21
5.5
(3)
<3
OCH(Me)pyrimi
dinꢀ2ꢀyl
19
(R)ꢀ
8.9
(5)
0.49 /
8.08
6.2
(3)
0.83 59.1 68.1
<3
OCH(Me)pyrida
zinꢀ3ꢀyl
20
21
22
(R)ꢀOCH(Me)ꢀ
4ꢀ
Tris
8.4
(5)
0.44 /
5.85
4.5
(2)
2.15 94.4
nd
<3
methylpyridinꢀ
2
ꢀyl
(R)ꢀOCH(Me)ꢀ
Tris
Tris
8.5
(6)
0.45 /
5.95
8.2
2.02
nd
97.8
22
60
5
ꢀ
(12)
methylpyridinꢀ
ꢀyl
(R)ꢀOCH(Me)ꢀ
2
8.5
0.45 /
7.0
2.65 97.9 99.0
ACS Paragon Plus Environment

Page 9 of 80
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
5ꢀchloropyridinꢀ
ꢀyl
(R)ꢀOCH(Me)ꢀ
ꢀfluoropyridinꢀ
(6)
5.66
(3)
2
2
3
4
Tris
Tris
8.6
(3)
0.45 /
6.33
7.6
(9)
1.98 94.0
nd
nd
31
21
5
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2ꢀyl
2
(R)ꢀOCH(Me)ꢀ
8.3
(5)
0.44 /
5.75
6.5
(6)
2.12 93.2
6
ꢀ
methylpyridinꢀ
2ꢀyl
Table 1
The Xꢀray crystal structure of human KMO has not been reported. However KMO from the
bacterium Pseudomonas fluorescens (Pf KMO) shows high levels of similarity to the human
protein in the active site. All the residues directly lining the active site are conserved with
human KMO, except His320 in Pf KMO which is a phenylalanine in human.
Modelling of these compounds based on the Xꢀray structure previously described for 1 in Pf
2
KMO , gave a good fit of the smaller alkoxy groups into the available space around the 6ꢀ
position, but the larger alkoxy and in particular the benzyl and pyridꢀ2ꢀylmethyl examples 12
and 13 could only be modelled into a pocket formed by Asn54, Glu195, Thr234 and Thr236
with displacement of the resident waters. This gave no explanation of the potency jump seen
for 13 as the nitrogen of the pyridine appeared to make no additional contacts.
We were able to obtain a crystal structure of 14 in Pf KMO (Figure 2). The crystal structure
showed only the Rꢀenantiomer 15 bound into the active site, with the pyridyl moiety
occupying an unexpected binding mode (Figure 2). In the Xꢀray structure of 14 / 15, the
bicyclic core and acid side chain are bound as we had seen previously for 1, although as with
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 80
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1 the exact interactions of the carboxylic acid vary between the A and B chains in the crystal
structure (a direct hydrogen bond between carboxylate and Tyr98 hydroxyl in the A chain
and a water mediated hydrogen bond in the B chain). However the tricyclic group of the
catalytic FAD has rotated to accommodate 15, allowing the (R)ꢀ1ꢀ(pyridineꢀ2ꢀyl)ethoxy to sit
underneath forming a face to face π interaction between the pyridine of 15 and FAD. The
pyridine nitrogen makes Hꢀbonding interactions with a water network extending towards
Pro318 and Val317, whilst the pyridine ring in addition to the face to face π interaction with
FAD also forms two edge to face π interactions from the 5ꢀposition of the pyridine into
Tyr193, and from Phe238 into the face of the pyridine ring.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
A similar binding mode for the Sꢀenantiomer 16 would result in obvious protein clashes, and
indeed this compound is much less potent.
Whilst there is high homology between the Pf KMO and human KMO and the SAR from the
human KMO is consistent with the Pf KMO crystal structures, these structures and the FAD
movement must be interpreted with some caution. 15 inhibited Pf KMO with a similar
potency as the human enzyme (Pf KMO pIC50 8.5, h KMO pIC50 8.6), however the h KMO
potency may be an underestimate as we are close to the tight binding limit of the assay.
Subsequent crystal structures on a small set of intermediately sized inhibitors 6, 7 and 9,
showed 6 and 7 to bind in the same mode as 1 with the sideꢀchain accommodated into a space
in front of the FAD tricycle. 9 however shows significant differences in binding mode in the
A and B chains of the protein crystal structure (Figures 3a / 3c). In the A chains, where the
ligand carboxylic acid is directly Hꢀbonded to Tyr98, the protein environment around the
ligand alkyl groups is the same, including the position of the FAD tricycle. The different
ACS Paragon Plus Environment

Page 11 of 80
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
sized alkyl groups are accommodated by small adjustments in the inhibitor bicycle position,
with the largest, cyclobutyloxy, group adjacent to the FAD in a perpendicular position.
For the B chains, where the carboxylic acid is bonded to Tyr98 through a water molecule, the
protein environment around the ligand alkyl groups is the same for the ethyl and cyclopropyl
groups as in the three A chain structures, however the cyclobutyl group induces a change in
the orientation of the FAD tricycle, and by changing its own conformation by approximately
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
90° it is able to adopt a conformation parallel to the FAD aromatic ring. This is the first
example of the FAD tricycle adopting an intermediate conformation between the two
extremes seen with 1 and 14/15 (Figure 3c, d). Although 9 is able to cause a change in FAD
tricycle conformation, in this case it has a lower potency than 6 and 7, suggesting that the
conformational change results in overall less favourable interactions in the KMO active site.
It appears that only when the larger and planar pyridyl group of 14/15 is present can the FAD
tricycle achieve the full positional shift required to allow the favourable aromatic πꢀstacking
and edge to face π interactions to be made.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 80
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Figure 2 Crystal structures of 1 (orange) and 14/15 (green) bound to Pf KMO protein (A
Chain). The tricycle of the catalytic FAD has rotated in the structure of 14 to accommodate
the pyridine.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a
b
c
d
Figure 3 a, b Comparison of binding modes of 9 (blue) with 6 (gold) and 7 (magenta)
bound to Pf KMO in the A chain (a) and in the B chain (b). In the A chain, the largest,
cyclobutyl, group shifts the inhibitor slightly towards Tyr98. In contrast, in the B chain,
whilst the FAD is twisted upon binding of the cyclobutyl substitutent, there is very little
difference in the actual positions of the inhibitor bicycles. c, d Comparison of binding
modes of 1 (orange), 9 (blue) and 14/15 (green) bound to Pf KMO in the B chain
exhibiting the three different observed FAD conformations. (c) front view (d) side view.
ACS Paragon Plus Environment

Page 13 of 80
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Three approaches were taken to further improve on 15 based on knowledge from the crystal
structure: to modify the 1ꢀ(pyridineꢀ2ꢀylethoxy) side chain, to substitute the pyridine ring and
to explore alternative bicyclic cores.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Analogues where the pyridine was replaced by an alternative heteroaromatic (17 – 19)
retained potent enzyme inhibition but had weaker cell potency, possibly due to a potentially
negative impact of the additional polarity on cell permeability, although this was not
confirmed. As expected the increased polarity correlated with lower protein binding.
From the crystal structure of 15 bound to Pf KMO, it was hypothesised that substitution of
the pyridine would be tolerated at any of the 4ꢀ, 5ꢀ and 6ꢀpositions, but that a 5ꢀposition
substituent would have the potential to form an interaction with the π cloud of Tyr193.
Substitution at the 3ꢀposition was excluded as being likely to clash with the αꢀmethyl group
in the exemplified binding mode. In the KMO primary assay, substitution of the pyridine
appeared to be tolerated at all of these positions, but only substitution at the 5ꢀposition was
beneficial in the cell with 21, 22 and 23, all having enhanced cell potency compared with 15.
The 1ꢀ(pyridinꢀ2ꢀylethoxy) was introduced onto a number of 6/5 and 6/6 cores, including
1
1
some which we had previously examined
and in others chosen based on the crystal
structures we had generated (Table 2). All examples demonstrated KMO pIC50s at around
the theoretical tight binding limit of the KMO enzyme assay. The cellular potency was
generally higher than for 15 with the exception of the indazole 27. The benzisoxazole 28 was
of particular interest with an approximately 60ꢀfold increase in cell potency over 15, although
with a ~5ꢀfold reduced free fraction (2.2% vs 11.9%). This cellular activity was confirmed in
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 80
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
primary human hepatocytes (pIC50 8.0). The thioheteroaryl 29 was also very potent with a
HEK cell pIC50 of 7.5.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Eg. No. PQR
Salt KMO mean
pIC50 (n)
HEK
mean cell
pIC50 (n)
6.7 (15)
7.2 (11)
7.6 (2)
chromlog hPPB
D
pH 7.4
1.71
1.56
2.17
1.70
1.99
1.81
(%)
15
OC(=O)ꢀN
OCH ꢀC(=O)ꢀN
Tris
Tris
ꢀ
8.6 (12)
8.3 (6)
8.3 (5)
8.5 (5)
8.3 (7)
8.7 (5)
88.1
92.5
nd
25
26
27
2
CH=CHꢀN
CH=NꢀN
OꢀN=C
Tris
Tris
ꢀ
7.1 (5)
90.1
97.8
nd
28
8.5 (12)
7.5 (1)
29
SꢀC(=O)ꢀN
Table 2
Translation of the substituted pyridines onto the more potent benzisoxazole core gave some
extremely cell potent examples (30, 31, 32), although with lower free fractions than the less
lipophilic benzoxazolones (Table 3). The (6ꢀmethylpyridazinꢀ3ꢀyl)ethoxy example 33 was
also prepared on the benzisoxazole core, aiming to combine the large free fraction of 19 with
the more lipophilic core and the additional methyl substituent (as in 30) to achieve a logD
consistent with good cellular potency. This combination, as desired, gave enhanced cellular
ACS Paragon Plus Environment

Page 15 of 80
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
activity (pIC50 8.5) and a higher plasma free fraction than the equivalent pyridine 30 (7.1% vs
0.7%).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Eg.
No.
Ar
R
Salt
KMO
mean
HEK
mean
cell
chromlog HSA hPPB
D
(%)
(%)
pIC50 (n)
pH 7.4
pIC50
(n)
30
5ꢀ
Me
Me
Me
Me
Tris
8.5 (6)
(30a HCl
salt)
8.8 (5)
8.4 (4)
7.5 (6)
8.5 (19)
2.42
2.91
2.47
2.97
98.4 99.7
methylpyridi
nꢀ2ꢀyl
3
1
2
5ꢀ
Tris
8.5 (5)
99.0 99.8
chloropyridin
ꢀ2ꢀyl
3
5ꢀ
ꢀ
ꢀ
8.5 (9)
8.7 (4)
98.3
nd
fluoropyridin
ꢀ2ꢀyl
33
6ꢀ
89.2 92.9
methylpyrida
zinꢀ3ꢀyl
Table 3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 80
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
The enhanced cellular activity of 14 and later analogues compared with the smaller alkyl
examples and 12 /13 and movement of the FAD in Pf KMO led us to examine the kinetics of
interaction between these molecules and hKMO in an attempt to understand the increase in
potency. We had previously shown that 2 had a fast dissociation rate. These kinetics of
interaction of a range of inhibitors with human KMO were studied and are reported in detail
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
15
elsewhere. In “recovery of activity” experiments 13 was found to have a dissociation T1/2 in
the region of 30min and a K in the order of 1nM, whereas the T1/2 of 15 was ~2h and its K is
i
i
approximately 100pM. For 28 the association kinetics were measured in onset of inhibition
timeꢀcourses using the mass spectrometry assay. Observed rate constants increased linearly
5
ꢀ1 ꢀ1
with [28], indicative of second order kinetics, with k = 6.2 x 10 M s . Due to the high
on
potency of 28, the recovery of activity approach could not be used to measure koff. Instead the
timeꢀcourse was analysed by numerical fitting to a model of competitive inhibition with kon
5
ꢀ1 ꢀ1
ꢀ5 ꢀ1
fixed at 6.2 x 10 M s , giving a value of koff = 1.94 x 10 s . The calculated K
i
of 28 is thus
50pM and the T1/2 was around 10h.
A number of the small alkoxy examples 1, 2, 6, 7, 8, chiral pyridyls 15, 28, 30 and the chiral
1
6
pyridazinyl 33 were profiled in i.v. rat DMPK (Table 4) and for stability to light and pH
under oxidative and nonꢀoxidative conditions and for solubility in saline. The examples with
a small alkyl or alkoxy substituent had a small volume of distribution, low to moderate
clearance and, except for 1 and 2, a short halfꢀlife. By contrast 15 had a long halfꢀlife and a
much higher volume of distribution. The pyridyl benzisoxazole examples, 28, 30 and 31 had
low volumes and clearance with moderate to long half lives, whereas pyridazinyl example 33
had a slightly higher volume, moderate clearance and moderate T1/2
.
Eg. No.
rPPB
Rat i.v.dmpk (~1 mg/kg)
Stability
Saline
ACS Paragon Plus Environment

Page 17 of 80
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
(%)
solubility
24 h
T
1/2 (h)
Clb
mL/
Vdss
pH6
Light (%
(
(l/kg)
oxidative remaining
(ug/mL)
min/k
g)
(T1/2
)
at 12h)
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
2
6
7
8
91.7
95.5
81.6
90.3
90.2
78.5
97.8
99.1
99.7
93.3
4.3
3.7
4.3
0.28
0.52
0.36
0.29
0.26
2.96
0.42
0.31
0.26
1.10
334 h
41
100
270
>127000
>1880
>2400
>2060
>9500
>7440
>860
4.7
>1000 h
>1000 h
>1000 h
>1000 h
>1000 h
>1000 h
>1000 h
>1000 h
>1000 h
0.43
0.39
0.68
5.6
16.0
9.4
99.5
100
4.1
104.9
99.2
97.7
96.5
99.5
85.2
15
16.8
0.8
28
6.6
30
31
3.6
1.1
8.7
0.35
12.0
>620
33
1.5
230
Table 4
All examples showed excellent stability to both light and relevant pH and all had good levels
of solubility in saline, particularly as the TRIS salts. We had identified the TRIS salt as a
means of significantly improving the aqueous solubility of the smaller alkyl / alkoxy
substituted molecules including 2 and for these the TRIS salts were well behaved solids.
However, for molecules containing a chiral pyridine or chiral heteroaryl side chain, the TRIS
salts were hygroscopic gums. We therefore returned to preparing the free acid for later
examples and compound batches.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 80
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Following consideration of all available data compound 28 was chosen for further profiling.
This compound showed no inhibitory effects on the other enzymes in the tryptophan
metabolism pathway (IDO, TDO, KATI / II, kynureninase) and had remarkable selectivity in
GSK’s liability panel (57 assays) with a window of >20,000 fold over all the targets
examined.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Compound 28 was evaluated in a rat PK/PD study using a loading i.v. bolus followed by a 3h
i.v. infusion formulated as the ethanolamine salt formed inꢀsitu in the dosing vehicle. The
potency of 28 against rat KMO was lower than observed with human enzyme. When rat
KMO was expressed in HEK cells the apparent cellular K was calculated as (45nM) and so
i
three dose levels were tested that delivered steady free drug levels ranging from 10x above to
10x below this level (Table 5). 28 had no impact on circulating TRP levels, but showed a
dose dependent effect on the concentration of 3ꢀhydroxykynurenine (the product of KMO)
and a corresponding increase in kynurenine (the substrate of KMO), with a half maximal
effect at concentrations around the rat cell K (Figure 6). As expected, dose dependent
i
increases were also observed in the diversionary products of kynurenine metabolism –
2
kynurenic acid and anthranilic acid – consistent with the effects previously reported.
Dose of 28 (free acid)
Free plasma
concentration
3nM
Concentration relative to
cellular Ki
0.025mg/kg & 0.0025mg/kg/hr
0.07x
0.25mg/kg & 0.025mg/kg/hr
50nM
1.1x
9.0x
2.5mg/kg & 0.25mg/kg/hr
405nM
ACS Paragon Plus Environment

Page 19 of 80
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 5
1
00
8
10
8
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
8
6
4
2
0
0
0
0
0
6
4
2
0
6
4
2
0
4
3
2
00
00
00
6
4
2
00
00
00
0
100
0
Figure 6 a) Effect of 28 on components of the kynurenine pathway measured at 3hrs after
initiation of dosing with doses of 2.5mg/kg bolus & 0.25mg/kg/hr infusion (high), 0.25mg/kg
bolus & 0.025mg/kg/hr infusion (med) and 0.025mg/kg bolus & 0.0025mg/kg/hr infusion
(lo) doses (Table 5).
17
Compound 28 was assessed in an experimental model of AP in rats dosing as an initial 20
mg/kg bolus followed by a 5h, 2.2 mg/kg/hr infusion therapeutically, commencing 1 hour
after the induction of AP. This regime delivered free concentration in plasma of 50x the
pIC for rat KMO and resulted in a decrease in 3ꢀhydroxykynurenine and an increase in
5
0
,
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 80
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
kynurenine as expected (Figure 7). The extent of pancreatic injury in this model (measured
by serum amylase levels) was only marginally impacted by treatment with 28. The stress
response during experimental APꢀMODS caused a decrease in plasma glucose concentration,
which was less pronounced in rats with AP treated with 28. Importantly, 28 showed
significant protective effects against the secondary organ injury to lung as assessed by
histology and measured by analysing protein leak into the alveolar space (in bronchoalveolar
lavage fluid). Furthermore, kidney injury was reduced with 28 treatment, measured by a
reduction in APꢀinduced plasma creatinine rise and a reduction in apoptotic cell count in the
outer medullary stripe of the kidney assessed by TUNEL staining (Figure 7).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
a)
b)
ACS Paragon Plus Environment

Page 21 of 80
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
*
*
*
*
2
1
0000
5000
20
15
10
5
10000
5000
0
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
0
*
*
*
*
3
00
1
1
50
00
200
1
00
0
5
0
0
*
*
3
2
1
0
0
0
0
c)
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 80
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Lung sham
Lung AP
Lung AP + 28
Figure 7 a) Reduced plasma levels of 3ꢀHK and increased levels of kynurenine compared
with both sham operated and AP control animals at 5 hours after start of administration of 28.
b) Therapeutic administration of 28 1 hour after the induction of experimental AP in rats
protects against secondary organ damage in lung and kidney. Each point represents data from
an individual rat, with group sizes as follows: sham n=5, AP n=6, AP + 28 n=7. For the lung
TUNEL figure panel, one extreme outlier data point was omitted from the sham group and
one extreme outlier from the AP + 28 group after careful consideration. Both excluded data
points were at least 7x standard deviations from the mean of the remainder of the
corresponding group and were felt to be due to technical error. Statistical comparison
between groups was by oneꢀway ANOVA with postꢀhoc StudentꢀNeumanꢀKeuls, *P < 0.05.
c) Lung histology representative photomicrographs. Haematoxylin and eosin staining of
formalinꢀfixed sections of lung tissue. Alveolarꢀepithelial thickening and inflammatory
infiltrates are visible on images in the AP group, which is less marked in the AP + 28 group.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
8
Dog i.v. infusion PK of the free acid of 28 dosed at 1 mg/kg showed a similar profile to rat
i.v. PK with long halfꢀlife and low volume and clearance (Dog: T1/2 5.8h, Vd 0.36, Clb 0.93).
Based on scaling of PK parameters from rat and dog, combined with the high cellular potency
and moderated free fraction, we estimate that a daily dose of less than 50mg should provide
substantial suppression of 3ꢀhydroxykynurenine levels in man.
SYNTHESIS
2
/19
The synthesis of 1 and 2 have been described previously
and examples 3 to 7 were
prepared using the same synthetic sequence (Scheme 1). The benzo[d]oxazolꢀ2(3H)ꢀones
ACS Paragon Plus Environment

Page 23 of 80
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
35a – g were prepared by cyclisation of the relevant 2ꢀaminophenol (34a-g) using either CDI
or triphosgene. The bicyclic product 35a-g was then alkylated with 3ꢀbromopropanoic acid
and the target molecule was then optionally converted to the TRIS salt in aqueous methanol
or ethanol.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 1
(i) CDI, THF, 60 ꢀ 80 °C or Triphosgene, Et
3 2 3
N, 0°C – rt (ii) 3ꢀbromopropanonic acid, K CO , MeCN, 60 ꢀ 80
°C (iii) TRIS, EtOH or MeOH, H O, 80 °C.
2
The remaining alkoxy substituted examples 8 to 24 were prepared as shown in Scheme 2
from the commercially available 6ꢀbromoꢀ5ꢀchlorobenzo[d]oxazolꢀ2(3H)ꢀone 36. This was
alkylated with methyl 3ꢀbromopropanoate to give 37.
Treatment of 37 with
bis(pinacolato)diboron and palladium catalysis give the pinnacolborane 38 which was
converted to the phenol 39 with hydrogen peroxide. The 6ꢀalkoxy group was installed either
under Mitsonubu conditions with the alcohol or by alkylation using either the appropriate
bromide or an activated alcohol. Acid mediated hydrolysis of the esters 40a-o was generally
preferred to avoid ring opening, which gave the target compounds 8 to 24.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 80
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 2
(
i) methyl 3ꢀbromopropanoate, K
00 °C (iii) 30% H , AcOH, THF, rt (iv) ROH, DEAD, PPH
80 °C or 6N HCl 70ꢀ100 °C or LiOH, dioxane, 0 °C ꢀ rt (vi) TRIS, MeOH, rt.
2
CO
3
, MeCN, 70 °C (ii) bis(pinacolato)diboron, PdCl (dppf), KOAc, dioxane,
2
1
2
O
2
3
or RX / ROMs, K CO (v) 0.5 N HCl, dioxane,
2
3
Compound 25 was synthesised in a similar manner to the equivalent benzoxazolone from 7ꢀ
bromoꢀ6ꢀchloroꢀ2Hꢀbenzo[b][1,4]oxazinꢀ3(4H)ꢀone 42, but using acrylonitrile in place of
methyl 3ꢀbromopropionate for the alkylation which installs the side chain (Scheme 3). The
nitrile 43 was converted to the ester 44 by hydrolysis with hydrochloric acid in methanol,
before conversion of the 7ꢀbromo to the 7ꢀphenol 46 via the pinnacol borane 45.
subsequent Mitsonubu coupling and ester hydrolysis gave 25.
A
ACS Paragon Plus Environment

Page 25 of 80
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 3
(
i) 2ꢀchloroacetyl chloride, KOAc, DMF, rt (ii) acrylonitrile, K
iv) bis(pinacolato)diboron, PdCl (dppf), KOAc, dioxane 100 °C, Ar (v) H
ꢀyl)ethanol, PPh , DEAD, toluene, rt (vii) 0.5N HCl, dioxane, 80 °C.
2
CO
3
, MeCN, 70 °C (iii) c. HCl, MeOH, 90 °C
(
2
2 2
O
, AcOH, THF, rt (vi) (S)ꢀ1ꢀ(pyridꢀ
2
3
The choice of alkylation conditions was found to be highly sensitive to the core bicyclic, for
ꢀchloroꢀ5ꢀmethoxyꢀ1Hꢀindole 48, ethyl acrylate was preferred as the alkylating group
Scheme 4). Conversion of the 5ꢀmethoxy indole 49 to the 5ꢀphenol 50 was followed by a
Mitsonubu coupling and ester hydrolysis to give 26.
6
(
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 80
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Scheme 4
(i) ethyl acrylate, K
3
PO
4
, MeCN, rt (ii) AlCl
O, rt.
3 3
, toluene, 100 °C (iii) (S)ꢀ1ꢀ(pyridꢀ2ꢀyl)ethanol, PPh , DEAD,
toluene, rt (iv) LiOH, THF, H
2
Commercially available 5ꢀbromoꢀ6ꢀchloroindazole 52 was also alkylated with ethyl acrylate
to give ester 53. This was then converted from the 5ꢀbromo compound in the same manner as
the analogous benzoxazolone to give compound 27.
Scheme 5
(i) ethyl acrylate, DBU, MeCN, rt (ii) bis(pinacolato)diboron, PdCl
2
(dppf), KOAc, dioxane 100 °C, Ar (iii)
O, rt.
H
2
O
2
, THF, AcOH, rt (iv) (S)ꢀ1ꢀ(pyridꢀ2ꢀyl)ethanol, PPh , DEAD, toluene, rt (v) LiOH, THF, H
3
2
ACS Paragon Plus Environment

Page 27 of 80
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
2
0
The benzisoxazole core can be synthesised by several routes , the shorter improved route is
shown in Scheme 6. Acid 57 was converted to the ester with thionyl chloride and methanol
and then the methoxyl adjacent to the ketone was selectively demethylated with aluminium
trichloride and sodium iodide to give phenol 59. The ketone was converted to the
hydroxylamine with ammonium hydroxide and the isoxazole closed by heating in a mixture
of acetic anhydride and pyridine. Demethylation of 61 using aluminium chloride gave phenol
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
6
2. This was then converted to ether 64a-d by addition of the chiral pyridine alcohol 63a
under Mitsonubu conditions or by alkylation with the relevant mesylate 63b-d. Subsequent
lithium hydroxide hydrolysis of the esters gave compounds 28, 30, 31, 32 and 33.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 80
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
O
O
O
O
O
O
(ii)
(i)
O
O
Cl
OH
Cl
O
5
7
5
8
O
OH
O
O
OH
N
O
(iii)
O
(iv)
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Cl
O
Cl
O
5
9
OH
60
O
O
N
(v)
HO
Cl
(vi)
O
N
Cl
R
61
X
O
R2
N
O
O
O
OR1
63a - e
6
6
2 R = Me
2a R = Et
63a R= H, X = CH, R = H
3b R= Me, X = CH, R = Ms,
63c R= Cl, X = CH, R = Ms
63d R= F, X = CH, R = H
1
6
1
1
1
R
63e R = Me, X = N, R1 = H
R
X
X
N
N
O
O
(vii)
O
O
N
N
Cl
Cl
R2
28, 30, 31, 32, 33
64a - e
O
OH
O
O
64a R= H, R = Me
64b R= Me, R = Et
4c R= Cl, R = Et
4d R= F, R = Me
2
2
6
6
6
2
2
4e R = Me, R = Me
2
Scheme 6
(
2 3 2
i) SOCl , MeOH, 0 °C ꢀ rt (ii) NaI, AlCl , MeCN, 80 °c (iii) pyridine, MeOH, NH OH.HCl, 100 °C (iv)
pyridine, Ac O, 110 °C to 120 °C, (v) AlCl , DCM, rt (vi) 63a: ROH, PPh , DEAD, THF or 63bꢀd: ROMs,
2
3
3
K
2
CO
3
, DMF or MeCN, 70 ꢀ 80 °C (vii) LiOH, THF, H
2
O, MeOH or EtOH, 0 °C ꢀ rt
The 5ꢀchloroꢀ6ꢀmethoxybenzo[d]thiazolꢀ2(3H)ꢀone 66 was prepared from the 5ꢀchloroꢀ6ꢀ
methoxybenzo[d]thiazolꢀ2ꢀamine 65 by diazotization and treatment with acetic acid (Scheme
7
). Alkylation with ethyl acrylate and demethylation with boron tribromide gave phenol 68.
ACS Paragon Plus Environment

Page 29 of 80
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Treatment of 68 with the mesylate of the chiral pyridine under alkylation conditions and
subsequent acid mediated hydrolysis of the ester 69 gave 29.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
O
O
O
S
(ii)
S
N
S
N
(i)
O
O
NH2
N
H
Cl
Cl
Cl
6
5
66
6
7
O
O
N
HO
Cl
O
(iii)
S
(iv / v)
S
O
O
N
N
Cl
68
6
9 R = Et
29 R = H
O
OR
O
O
Scheme 7
(
i) HCO H, AcOH, c. HCl, NaNO , H O, ꢀ10 ꢀ 100 °C (ii) ethylacrylate, Si(OEt) , CsF, toluene, 110 °C (iii)
2
2
2
4
BBr
3
, DCM, rt (iv) (S)ꢀ1ꢀ(pyridꢀ2ꢀyl)ethanol, PPh
3
, DEAD, toluene, rt (v) 0.5N HCl, dioxane, 90 °C.
CONCLUSION
We report here the development and optimisation of a series of KMO inhibitors arising from
a Discovery Partnership with Academia collaboration between the University of Edinburgh
and GSK. Using novel Pf KMO crystal structures, we have rationalised some unexpected
SAR on the human KMO enzyme. This has enabled further optimisation using the principles
of structureꢀdriven design to generate molecules with very high KMO inhibitory potential. In
the process we have also converted the series from compounds with a rapid offꢀrate (2) to
examples with very slow dissociation kinetics (15, 28). The resulting molecules have potent
cellular activity, excellent aqueous solubility and a DMPK profile suitable for i.v. dosing.
The effect on plasma concentrations of kynurenine pathway metabolites after administration
of 28 recapitulates the previously observed effect using tool KMO inhibitor compounds and
ACS Paragon Plus Environment