Structure-based optimisation of orally active & reversible MetAP-2 inhibitors maintaining a tight ‘molecular budget’

Article abstract of DOI:10.1016/j.bmcl.2020.127533

Structure-based led optimisation of orally active reversible Methionine Aminopeptidase-2 (MetAP-2) inhibitors utilising a ‘molecular budget’ medicinal chemistry strategy is described. The key physicochemical parameters of target molecules (cLogP, molecula

Full text of DOI:10.1016/j.bmcl.2020.127533

Journal Pre-proofs
Structure-based optimisation of orally active & reversible MetAP-2 inhibitors
maintaining a tight ‘molecular budget’
David J. Hirst, Martin Brandt, Gordon Bruton, Erica Christodoulou, Leanne
Cutler, Nigel Deeks, Jonathan D. Goodacre, Torquil Jack, Matthew Lindon,
Afjal Miah, Kevin Page, Nigel Parr, Lena Shukla, Martin Sims, Pamela
Thomas, James Thorpe, Duncan S. Holmes
PII:
S0960-894X(20)30644-2
DOI:
Reference:
https://doi.org/10.1016/j.bmcl.2020.127533
BMCL 127533
To appear in:
Bioorganic & Medicinal Chemistry Letters
Received Date:
Revised Date:
Accepted Date:
28 May 2020
22 August 2020
29 August 2020
Please cite this article as: Hirst, D.J., Brandt, M., Bruton, G., Christodoulou, E., Cutler, L., Deeks, N., Goodacre,
J.D., Jack, T., Lindon, M., Miah, A., Page, K., Parr, N., Shukla, L., Sims, M., Thomas, P., Thorpe, J., Holmes,
D.S., Structure-based optimisation of orally active & reversible MetAP-2 inhibitors maintaining a tight
‘molecular budget’, Bioorganic & Medicinal Chemistry Letters (2020), doi: https://doi.org/10.1016/j.bmcl.
2020.127533
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover
page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version
will undergo additional copyediting, typesetting and review before it is published in its final form, but we are
providing this version to give early visibility of the article. Please note that, during the production process, errors
may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
© 2020 Published by Elsevier Ltd.

Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com
Structure-based optimisation of orally active & reversible MetAP-2 inhibitors
maintaining a tight ‘molecular budget’
David J. Hirsta,, Martin Brandtb, Gordon Brutona, Erica Christodouloua, Leanne Cutlera, Nigel Deeksa,
Jonathan D. Goodacrea, Torquil Jacka, Matthew Lindona, Afjal Miaha, Kevin Pagea, Nigel Parra, Lena
Shuklaa, Martin Simsa, Pamela Thomasa, James Thorpea and Duncan S. Holmesa.
aGlaxoSmithKline, Medicines Research Centre, Gunnels Wood Road, Stevenage, Hertfordshire, SG1 2NY, U.K.
bGlaxoSmithKline, 1250 South Collegeville Rd., Collegeville, PA 19426, USA.
ARTICLE INFO
ABSTRACT
Article history:
Received
Revised
Accepted
Available online
Structure-based led optimisation of orally active reversible Methionine Aminopeptidase-2
(MetAP-2) inhibitors utilising a ‘molecular budget’ medicinal chemistry strategy is described.
The key physicochemical parameters of target molecules (cLogP, molecular size and H-bond
donor count) were monitored through straightforward and intuitive use of atom count and
distribution. The balance between structure-based design and an awareness of the physico-
chemical properties of the compounds synthesised enabled the rapid identification of a potent
molecule with good oral pharmacokinetic (PK) characteristics by making fewer, higher quality
compounds. The resulting candidate quality molecule was validated in a mechanistic cellular
assay and a rodent secondary immunisation model.
Keywords:
Methionine Aminopeptidase-2
MetAP-2
Molecular budget
Structure-based drug design
Indole carboxamide
2020 Elsevier Ltd. All rights reserved.
———
 Corresponding author. Tel.: +44-1438-762658; fax: +44-1438-768302; e-mail: david.j.hirst@gsk.com

Human methionine aminopeptidase-2 (MetAP-2) is
a
MetAP-2 is a metalloprotease with two metal ion binding
regions bound by five active site residues: His331, Asp262,
Asp251, Glu459 and Glu364.14 This provides 4-point coordination
of each metal with a water molecule contributing a fifth.
Modelling of the endogenous substrate indicated that the P1’
pocket accommodates the Gly or Ala amino acid terminus released
after successful removal of the methionine by the enzyme. We
were fortunate enough to have access to robust ligand-protein
crystalography, the structures of which were shown to closely
match those published previously.14 In the case of ligand 9, the
indole itself extends into the hydrophobic P1 pocket in which the
methionine alkyl thioether chain of the endogenous substrate is
believed to bind. The 2-carbonyl of the ligand binds the first metal
directly while a water bridge exists between the NH of the indole
and the second metal. Meanwhile the 3-aryl ring exhibits two
oblique edge-to-face interactions above and below its plane with
Tyr444 and His339 (Figure 3). This binding is analogous to that
of published ligands 6 & 8 (Figure 3b,c); and in particular the latter
manganese containing metalloprotease. Its function within the cell
is to remove the N-terminal methionine from nascent H-Met-Gly
and H-Met-Ala proteins.1-3 This cleavage function is shared by
another protein in the same family, human methionine aminopepti-
dase-1 (MetAP-1). These proteins are also linked via further
downstream processing (e.g. myristoylation), to the targeting of
proteins toward their final destination and other cellular ‘house-
keeping’ activities. It has been reported that MetAP-2 has a
selective and high expression in germinal centre B-cells.4
Subsequently it has also been demonstrated that PPI-2458 1, a
selective nanomolar inhibitor of MetAP-2, inhibits antibody
production
in-vitro
through
interleukin-21
mediated
differentiation of B-cells to plasma cells.5 Furthermore, treatment
with PPI-2458 1 in a marmoset immunisation model resulted in
inhibition of T-cell dependant antibody production and depletion
of B-cells in the germinal centre. These findings generated an
interest in the role of MetAP-2 in B-cell dependent immuno-
inflammatory diseases such as rheumatoid arthritis.
O
H
H
N
H
O
CONH2
The MetAP-2 protein was isolated and characterised in 1995,
and was subsequently identified as the cellular target of the known
anti-angiogenic agent TNP470 2.6 Both 1 & 2 are analogues of the
natural product Fumagilin 3, a relative of which (Ovalicin) is also
a known immunosuppressive agent.7 The generation of a crystal
structure showing the covalently linked Fumagilin 3-MetAP-2
complex confirmed that these compounds inhibited MetAP-2
covalently through opening of the labile C-4 spiro-epoxide.8
N
CO2H
4
R =
Cl
OMe
R
O
O
1
PPI-2458,
2
TNP470,
Fumagilin,
3
O
pyrazoloindoles, the crystal structures of which superpose very
well with that of 9.11-13
Figure 1. Structures of irreversible MetAP-2 inhibitors 1-3.
TNP470 2 has been shown to induce severe neurotoxicity in-
vivo; beleived to be caused by irreversible inhibition of the Met-
AP enzymes within the central nervous system (CNS).9 Indeed,
Fumagilin analogues with reduced CNS penetration have
continued to be developed as potential oncology targets (PPI-2458
O
NEt2
S
O
O
N
N
S
O
N
N
N
H
N
H
N
N
H2N
N
Cl
O
OH
F
H
OH
Abbott
O
4
5
6
Merck
Abbott
F
F
N
O
HO
N
N
NH
NH
O
Cl
NH2
O
N
Figure 3: (a) Space-filling depiction of X-ray crystal structure of 9 in MetAP-2
(pdb7A12); (b) Superposed ligand-MetAP-2 X-ray crystal structure of 9
(orange) and analogue of 6 (blue pdb5LYW); (c) Superposed ligand-MetAP-2
X-ray crystal structure of 9 and analogue of 8 (magenta pdb5JFR).
N
F
H
N
N
H
MetAP-2 pIC50: 6.5
HN
9
7
8
Takeda
Merck
1 entered Phase 1 clinical trials in 2004). Our aim was to overcome
potential toxicity risks by investigating reversible inhibition of
MetAP-2, which would offer the prospect of maintaining the
beneficial effects (inhibition of primary B-cell function) yet
provide a significantly improved safety profile resulting in an
improved therapeutic index. Since undertaking this research
several other reversible inhibitors have subsequently been
reported; primarily discovered as part of progams targeting
oncology and weight-loss.10-13 Nonetheless, at the time we were
unaware of any such molecules and conducted a comprehensive
screen of the GSK collection against MetAP-2. This identified a
novel series of indoles which, after follow-up focused library
synthesis and early hit-to-lead chemistry, led to indole 9 (Figure
2), an interesting start-point for lead optimisation.
Our ability to readily generate protein-ligand crystal structures
meant early optimisation could be guided by structure-based drug
design. Utilising this information we devised a focused strategy
targeting areas of the molecule that we believed would deliver
maximum potency enhancement (high ligand efficiency) while
allowing modulation of the physicochemical properties. As such,
we decided to concentrate on optimisation of the amido indole core
and 3-aryl ring by: (a) investigation of the carboxamide metal
binding motif; (b) judicious replacement of ring carbons with
nitrogen; (c) exploring the effect of incorporating indole halogens;
(d) addition of ortho-substituents to bind within the P1’ pocket; (e)
incorporation of para-substituents primarily to tune molecular
properties (Figure 4).
Figure 2. Selected reversible MetAP-2 inhibitors 4-8 and new ‘hit’ 9.

phenyl ring of the indole. Optimisation of this interaction would
ensure maximum ligand efficiency and provide as much room for
maneuver with later modifications as possible.
Potential surface interactions
- Explore SAR of potency
enhancing para-substituents
Additive effect?
Table 1. Indole 2-amide variations.
P1' pocket
- Target from the ortho position
Ph
Cl
R
X
NH2
O
Metal binding group
- Investigate alternative
metal binding groups
N
H
N
H
Cmpd
9
R
cLogP
3.5
MWt
271
CMR
7.7
# HBD
MetAP-2
pIC50
P1 pocket
- Subn of ring carbons for nitrogen to lower cLogP
- Optimize interaction in the P1 pocket through
addition of F at C-4, C-6 or C-7
2
1
6.5
CONH2
10
O
N
4.3
295
8.3
<4.6
Figure 4. Summary of medicinal chemistry strategy.
Although ligand efficient, we believed compound 9 had a
potency ~1.5 to 2 orders of magnitude below that required to
observe efficacy. However, we were aware that historically drug
discovery programs have proved extremely successful at
increasing potency within lead series; albeit often to the detriment
of the molecular properties.15 Numerous articles proposing
various ‘rules’ for the ‘ideal’ physicochemical space for orally
bioavailable compounds have been published.16 Based on these
reports, we chose to concentrate on MWt, HBD count and cLogP;
suitable control of these easily predicted parameters should ensure
a greater chance of designing and synthesising molecules with
superior in vivo PK profiles. MWt, HBD count and cLogP are all
intimately related to the number, ratio and particular presentation
of the heavy atoms within a structure. On this premise, ‘ideal’
ranges for physicochemical parameters directly translate to ‘ideal’
ranges for heavy atom count and distribution i.e. an atomic or
‘molecular budget’ exists within which we should work.17 Once
the upper limits of the budget are approached, redistribution rather
than continual addition of atom count is necessary. As a result, the
number of suitable changes for a given position is reduced and the
effort becomes more focused on making smaller numbers of high
quality compounds. We set ourselves a ‘molecular budget’ of
cLogP 2-4, MWt 200-450 (ideally <400 or CMR 7-12), #HBD 0-
3. Following these guidelines while optimising lead molecule 9
there existed the option of adding a maximum of ~6-10 additional
heavy atoms (ca. 80-130 MWt) and ~0.5 log unit of additional
cLogP (meaning additional heavy atoms would need to be
carefully balanced between carbon and heteroatoms), and ideally
no additional hydrogen bond donors.
11
4.7
3.2
270
289
7.8
7.7
1
1
5.8
O
12**
<4.6
S
O
O
13
14
15
5.7
4.0
3.7
300
272
285
8.4
7.5
8.1
1
1
2
<4.6*
4.7
CO2Et
CO2H
<4.6*
CONHMe
*Data generated in a different (but corrolatory) assay; **5-Fluoro analogue.
Good surface complementarity was observed between the 6-
membered ring of the indole and the P1 pocket of MetAP-2; and
the 5-chloro substituent present in 9 appeared to match the shape
of the P1 pocket particularly well. This observation was supported
by the reduction in potency of differentially chloro-substituted
analogues 16-18 and the non-chlorinated analogue 20 (Table 2).
Table 2. Effect of indole substutution.
Ph
R1
4
NH2
5
6
N
O
R2
7
H
Cpd
R1
R2
cLogP
MWt
CMR
# HBD
MetAP-2
pIC50
9
5-Cl
4-Cl
6-Cl
H
H
H
3.5
3.5
3.5
271
271
271
7.7
7.7
7.7
2
2
2
6.5
16
17
<4.6
<4.6
We were eager to identify the most efficient metal-binding
motif from the outset. The crystallographic information indicated
that the amide carbonyl oxygen bound directly to one of the active
site metals. Thus, changes to this region could not only have a
significant impact on potency but also the HBD count of 2. We
believed the indole NH to be essential for activity since it made
important positive interactions with the structural water network
in the active site. This was demonstrated by a >80 fold drop in
potency for the indole N-methylated analogues of compound 9
(data not shown). Therefore, we were keen to explore the
possibility of reducing #HBD with amido mimetics. A range of
analogues were synthesised, which either maintained or reduced
the HBD count of the core (Table 1). Given the confined space
around the protein bound amide, it was perhaps unsurprising that
significant alterations such as oxazole 10 and sulphone 12 resulted
in complete loss in potency. Nonetheless, we were somewhat
disappointed to find that all other modifications investigated also
led to a significant drop in potency relative to compound 9. On a
positive note, we were aware of the lipophilicity gain in most of
these analogues and we would not need to compensate for this at
a later stage. Based on these data we opted to retain the primary
amide and concentrate on the other key area of core, namely the
18
19
20
21
22
23
24
25
26
7-Cl
5-F
5-H
H
H
H
3.5
2.9
2.7
2.2
2.0
2.7
3.7
3.7
3.7
271
254
236
237
237
272
289
289
289
7.7
7.2
7.2
7.0
7.0
7.5
7.7
7.7
7.7
2
2
2
2
2
2
2
2
2
5.1*
6.2
5.6
<4.6
5.2
5.9
5.4
7.2
6.2
H
4-aza
6-aza
H
5-Cl 7-aza
5-Cl 4-F
5-Cl 6-F
5-Cl 7-F
*Data generated in a different (but corrolatory) assay.

Examination of the structural data led us to believe replacement
of the indole 6-membered ring carbons with nitrogen, or
substitution with fluorine, could be tolerated and improved shape
complementarity might be possible. Unfortunately, all aza-
analogues (examples 21, 22 and 23) did not maintain the potency
a negative effect on ligand efficiency). However, the addition of
a piperazine group to give compound 39 was of note since, despite
adding significant MWt, it combined a significant decrease in
cLogP with an increase in potency. Structural data indicated the
N-methyl of the piperazine was directed towards solvent and the
group as a whole did not introduce any specific new interations
with the enzyme. Rather, the increased affinity was attributed to;
a) non-specific hydrophobic surface interactions between the
piperazine and enzyme; b) electron-donation from the piperazine
leading to increased electron-density on the phenyl, thus
enhancing the aforementioned edge-to-face interactions with
Tyr444 and His339. (Note: electron-donating ortho-substituents
likely have a similar effect).
of the corresponding all-carbon compounds.
This was
disappointing due to the potential advantage they held in reduction
of the inherent cLogP. However, the increased potency observed
with compound 25 provided confidence in our structural analysis
and design while the fluorine had only a minor impact on overall
MWt, cLogP and HBD count (Figure 5).
Concomitantly, we investigated modification of the 3-aryl
group. The crystallographic data indicated positive interactions
could be obtained from the P1’ region of the protein and that a
vector to access this area was provided from the ortho-position of
the 3-aryl ring.18 Maintaining our strict control of predicted
properties, we designed and synthesised a small number of
compounds to explore this vector (Table 3).
Table 4. Effect of 3-aryl para-substitution.
R
Cl
NH2
O
N
H
Table 3: Effect of 3-aryl ortho-substitution.
Cpd
R
cLogP
MWt
CMR
# HBD
MetAP-2
pIC50
R1
Cl
R2
NH2
O
3.5
4.2
3.4
3.7
2.9
2.6
3.0
3.5
3.3
3.7
271
305
301
314
296
368
356
369
328
337
7.7
8.2
8.3
9.0
8.2
10.2
9.9
10.6
9.4
9.4
2
6.5
6.4*
6.9
7.0
6.3*
6.6
7.3
7.6
6.0
6.6
9
H
Cl
N
H
Cpd
R1
R2
cLogP
MWt
CMR
# HBD
MetAP-2
pIC50
33
34
35
36
37
38
39
40
41
2
OMe
NMe2
CN
2
2
9
H
H
H
3.5
2.9
271
301
7.7
8.3
2
2
6.5
27
OMe
7.0
2
28
Me
H
3.7
285
8.1
2
6.5
O
N
2
29 OCF3
H
H
F
4.0
3.7
3.0
3.9
355
329
319
347
8.3
9.2
8.3
9.2
2
2
2
2
6.4
7.6
8.2
8.3
O
N
2
30
31
32
OiPr
OMe
OiPr
N
2 (3)
2 (3)
3
N
CH2NMe2
F
N
HN
Although ortho-methoxy compound 27 provided an increase in
potency over that of the unsubstituted analogue 9, structural data
indicated it did not fill the P1’ pocket optimally.19 We postulated
an increase in steric bulk at this position might offer enhanced
potency but that, due to the lipophilic nature of the P1’ pocket, this
would likely require enlarged alkyl substituents. Compound 27
had a molecular weight of 301, leaving approximately 100 amu
(~8-9 heavy atoms) available before moving outside our ‘budget’.
The cLogP of 2.9 was of greater concern as the addition of just two
carbon atoms would risk moving out of the desired range. This
restricted the choice of potential groups at this position to ethyl, n-
propyl and iso-propyl. iso-Propyl was felt to offer optimal space-
filling and had the added advantage of reducing the potential for
dealkylation. Satisfyingly, the 2-iso-propyl ether had the expected
beneficial effect on potency, and combination with the 5-chloro-6-
fluoro indole gave compound 32 (pIC50 = 8.3).19 Unfortunately,
this promising result was offset by the increase cLogP which was
approaching the upper boundary of the desirable range.
*Data generated in a different (but corrolatory) assay.
At this point we were keen to ascertain whether the potency
enhancements observed independantly via addition of ortho-
isopropoxy and para-piperazine groups was additive; and if so,
could we control the MWt and cLogP and remain within ‘budget’.
Cogniscent of this, we opted to combine these two groups whilst
removing the piperazine methyl; we hypothesised this would have
little effect on potency while minimising the impact on the overall
physicochemical properties. To our satisfaction, the fully
elaborated compounds 42 & 43 demonstrated additive SAR and
high potency at MetAP-2 (Table 5). This was corroborated by the
X-ray crystal structure of 43 bound to MetAP-2 which illustrated
the minimal effect of the distal optimised groups on each other
(Figure 5).19 Based on these promising data, we tested 43 in a
cellular setting where we saw significant inhibition of IgG
secretion in murine B-cells.
In parallel we also explored the para-position of the 3-aryl
group with a small set of substituents designed to explore chemical
space, with a view to increasing potency whilst remaining within
our ‘molecular budget’ (Table 4). Although all modifications
generated molecules within our preferred physicochemical
criteria, a majority had little impact on potency (and consequently

The 2-isopropoxy-4-Boc-piperazinyl pinacol boronate ester
side-chain 51 was synthesised from the commercially available 3-
hydroxyphenylpiperazine 49. Boc-protection followed by iodine
monochloride iodination selectively provided the 4-iodo-3-
hydroxyphenylpiperazine 50. Alkylation of the phenol oxygen
and subsequent palladium catalysed borylation afforded 51. All
other molecules mentioned herein were synthesised by an
analogous route.
Figure 5. X-ray crystal structure of 43 in MetAP-2 (pdb7A16).
In 43 we now had a molecule of sufficient potency with reduced
cLogP and acceptable HBD count. The MWt was at the upper
limit, but exaggerated by the presence of two halogen atoms.
However, the more accurate CMR was within the appropriate
range, consequently we were confident the overall
physicochemical profile fulfilled the criteria we had set for an oral
drug molecule. Pleasingly, this resulted in a good in vitro and
mouse in vivo PK profile (Table 5); with moderate to low blood
clearance (45 mL/min/kg), a half-life of 4.5 hours and a moderate
volume of distribution (as expected with a compound containing a
basic amine). The oral bioavailability was almost quantitative
(89% at 10 mg/kg) while solubility was high; approaching or
exceeding >1 mg/mL in all media tested.
F
F
F
I
OH
b
a
c, d
N
H
O
F
F
NH2
F
NH2
45
46
44
H
Boc
N
N
N
N
I
F
F
e
f, g
OMe
O
O
O
N
H
F
F
F
F
NH2
OMe
O
47
N
H
43
N
H
48
O
NBoc
NBoc
NH
N
Table 5. Key data for compound 42 & 43.
N
j, k
h, i
N
O
H
HO
H
HO
B
N
O
N
b)
O
I
Cpd
N
N
49
50
51
Scheme 1. Synthetic route to compound 43. Reagents and conditions: a) I2,
NaHCO3, 92%; b) pyruvic acid, Pd(OAc)2, DABCO, 96%; c) H2SO4, MeOH,
38%; d) NIS, 91%; e) 51, Pd2(dba)3, DavePhos, K2CO3, 74%; f) NH3, LiCl,
75%; g) TFA, HCl, 85%; h) Boc2O, NEt3, 46%; i) ICl, 65%; j) 2-iodopropane,
K2CO3, 52%; k) Pinacol borane, Pd(OAc)2, cHex-JohnPhos, NEt3, 47%.
O
O
Cl
F
F
NH2
NH2
N
O
N
H
F
O
H
42
43
MetAP-2 pIC50
Mouse B-Cell IgG pIC50
MWt / CMR / #HBD
cLogP / ChromLogD7.4
Solubility (ug/mL)
8.5
nt
8.7
7.4
The in vivo efficacy of compound 43 was tested in a mouse
model of B-cell adaptive immunity using the antigen KLH-TNP;
allowing us to measure the effect of MetAP-2 inhibition on the
secondary antibody response of differentiated memory B-cells.
Compound 43, dosed twice daily from day 21 to day 31, inhibited
this secondary response in a concentration dependent manner
(Figure 6).
431 / 11.7 / 3
414 / 11.2 / 3
2.8 / 2.8
629
3.4 / 3.1
nt
<1, <1
nt
CLi [R/H] (mL/min/kg)
CLb [mouse] (ml/min/kg)
Vdss (L/kg), T1/2 (h)
F% (10 mg/kg)
<1, <1
45
1400
nt
7.7, 4.5
89
Pre-Boost
1200
1000
800
600
400
200
0
nt
7 days Post-Boost
Compound 43 itself was synthesised from the commercially
available 3,4-difluoroaniline 44. Iodination followed by
palladium catalysed intramolecular ring-closure afforded the
indole-2-carboxylic acid 46 in high yield. Following
esterification, selective iodination at the 3-position of the indole
was achieved with N-iodosuccinimide (NIS) to give the 3-iodo
indole-2-carboxylate 47. Coupling of this intermediate with the
separately prepared 2-isopropoxy-4-Boc-piperazinyl pinacol
boronate ester 51 gave the 3-aryl indole 48. Amidation of the
methyl ester and Boc-deprotection afforded compound 43 in 13%
overall yield over 7 linear steps.
*
**
Vehicle
GSK2229238A GSKC2p2d2942338A GSKC2p2d2942338A
Cpd 43
mg/kg
30 mg/kg
10mmgg//kkgg
33mmgg//kkgg
Figure 6. In vivo efficacy data for compound 43. Graph depicting levels of
anti-TNP specific IgG in vehicle and compound-treated (30, 10, 3 mg/kg) mice.
Dosing at 30, 10 & 3 mg/kg inhibited the secondary response
by 89%, 39% & 15% respectively. Similar results were obtained
when measuring anti-KLH specific IgG (data not shown). To the
best of our knowledge this is first demonstration of a reduction in
the secondary immune response in any species by a reversible
inhibitor of MetAP-2.

6.
Griffith, E. C.; Su, Z.; Niwayama, S.; Ramsay, C. A.; Chang, Y.-H;
Liu, J. O. Proc. Natl. Acad. Sci. U.S.A. 1998, 95, 15183.
Arrenbrecht, S.; Lazary, S. Agents & Actions, 1970, 1, 221.
Liu, S.; Widom, J.; Kemp, C. W.; Crews, C. M.; Clardy, J. Science,
1998, 282, 1324.
A
retrospective analysis of the number of molecules
synthesised and their properties indicated that our application of a
‘molecular budget’, coupled with detailed structure-based design
had provided an effective framework by which to focus our efforts
(Figure 7). By synthesising fewer yet well-designed targets, in
four months we were able to identify a first-class molecule (43)
suitable for oral in vivo dosing. In doing so we synthesised <50
compounds, almost all of which had physico-chemical properties
within our desired space.
7.
8.
9.
Turk, B. E.; Su, Z.; Liu, J. O. Bioorg. Med. Chem. 1998, 6, 1163.
10. Yin, S.-Q.; Wang. J.-J.; Zhang, C.-M.; Liu, Z.-P. Curr. Med. Chem.
2012, 19, 1021.
11. Heinrich, T.; Buchstaller, H.-P.; Cezanne, B.; Rohdich, F.; Bomke,
J.; Friese-Hamim, M.; Krier, M.; Knöchel, T.; Musil, D.; Leuthner,
B.; Zenke, F. Bioorg. Med. Chem. Lett. 2017, 27, 551.
12. (a) Heinrich, T.; Seenisamy, J.; Blume, B.; Bomke, J.; Calderini,
M.; Eckert, U.; Friese-Hamim, M.; Kohl, R.; Lehmann, M.;
Leuthner, B.; Musil, D.; Rohdich, F.; Zenke, F. T. J. Med. Chem..
2019, 62, 5025. (b) Heinrich, T.; Seenisamy, J.; Becker, B.; Blume,
B.; Bomke, J.; Dietz, M.; Eckert, U.; Friese-Hamim, M.; Gunera,
J.; Hansen, K.; Leuthner, B.; Musil, D.; Pfalzgraf, J.; Rohdich, F.;
Siegl, C.; Spuck, D.; Wegener, A.; Zenke, F. T. J. Med. Chem..
2019, 62, 11119.
a)
13. (a) Cheruvallath, Z.; Tang, M.; McBride, C.; Komandla, M.; Miura,
J.; Ton-Nu, T.; Erikson, P.; Feng, J.; Farrell, P.; Lawson, J. D.;
Vanderpool, D.; Wu, Y.; Dougan, D. R.; Plonowski, A.; Holub, C.;
Larson, C. Bioorg. Med. Chem. Lett. 2016, 26, 2774. (b) McBride,
C.; Cheruvallath, Z.; Komandla, M.; Tang, M.; Farrell, P.; Lawson,
J. D.; Vanderpool, D.; Wu, Y.; Dougan, D. R.; Plonowski, A.;
Holub, C.; Larson, C. Bioorg. Med. Chem. Lett. 2016, 26, 2779.
14. Sheppard, G. S.; Wang, J.; Kawai, M.; Fidanze, S. D.; BaMaung,
N. Y.; Erickson, S. A.; Barnes, D. M.; Tedrow, J. S.; Kolaczkowski,
L.; Vasudevan, A.; Park, D. C.; Wang, G. T.; Sanders, W. J.;
Mantei, R. A.; Palazzo, F.; Tucker-Garcia, L.; Lou, P.; Zhang, Q.;
Park, C. H.; Kim, K. H.; Petros, A.; Olejniczak, E.; Nettesheim, D.;
Hajduk, P.; Henkin, J.; Lesniewski, R.; Davidsen, S. K.;Bell, R. L.;
J. Med. Chem. 2006, 49, 3832.
15. Hann, M. M. MedChemComm, 2011, 2, 349.
16. a) Leeson, P. D.; Springthorpe, B. Nat. Rev. Drug. Disc. 2007, 6,
881; b) Hughes, J. D.; Blagg, J.; Price, D. A.; Bailey, S.;
DeCrescenzo, G. A.; Devraj, R. V.; Ellsworth, E.; Fobian, Y. M.;
Gibbs, M. E.; Gilles, R. W.; Greene, N.; Huang, E.; Krieger-Burke,
T.; Loesel, J.; Wager, T.; Whiteley, L.; Zhang, Y. Bioorg. Med.
Chem. Lett. 2008, 18, 4872; c) Gleeson, M. P. J. Med. Chem. 2008,
51, 817; d) Leeson, P. D.; Young, R. J. ACS Med. Chem. Lett. 2015,
6, 722; e) Waring, M. J. Expert Opin. Drug Discov. 2010, 5, 235.
17. Stemp, G. Chemistry Today, 2008, 26, 32.
Figure 7. Data for all synthesised compounds [compound 43 highlighted in
red]: a) potency vs registration date [by month]; b) distribution of CMR vs
cLogP [‘molecular budget’ space indicated].
18. Additional X-ray crystallographic data are available from the
RSBC: 27 – pdb7A13; 32 – pdb7A14; 42 – pdb7A15.
In conclusion, optimisation of a series of indole MetAP-2
inhibitors has been described. We have shown that, not only were
we able to improve the potency of the lead by >150-fold (9 to 43,
pIC50 = 6.5 to = 8.7 respectively) through judicious use of X-ray
crystal structure data; but in doing so, we could also maintain good
physico-chemical properties by controlling our ‘molecular
budget’. This resulted in the rapid identification of a molecule
with good oral mouse PK, which elicited a positive effect in a
secondary immunisation animal study.
Acknowledgements
The authors would particularly like to acknowledge the efforts
of Colin Eldred, Brian Evans, Markus Koglin, Onkar Singh,
Tadeusz Skarzynski, Chris Smethurst and Michael Wall; and also
Prof. Ed Tate for insightful scientific discussions.
References and notes
1. Arfin, S. T.; Kendall, R. L.; Hall, L.; Weaver, L. H.; Stewart, A. T.;
Mathews, B. W.; Bradshaw, R. A. Proc. Natl. Acad. Sci. U.S.A.
1995, 92, 7714.
2. Li. X.; Chang, Y.-H. Proc. Natl. Acad. Sci. U.S.A. 1995, 92, 12357.
3. Serero, A.; Giglione, C.; Sardini, A.; Martinez-Sanz, J.; Meinnel, T.
J. Biol. Chem. 2003, 278, 52953.
4. Kanno, T.; Endo, H.; Takeuchi, K.; Morishita, Y.; Fukuyama, M.;
Mori, S. Lab Invest., 2002, 82, 901.
5. Priest, R. C.; Spaull, J.; Buckton, J.; Grimley, R. L.; Sims, M.;
Binks, M.; Malhotra, R. Clin. Exp. Immunol. 2008, 155, 514.

Click here to remove instruction text...