Structure-based optimization of non-peptidic Cathepsin D inhibitors

Article abstract of DOI:10.1016/j.bmcl.2014.07.054

We discovered a novel series of non-peptidic acylguanidine inhibitors of Cathepsin D as target for osteoarthritis. The initial HTS-hits were optimized by structure-based design using CatD X-ray structures resulting in single digit nanomolar potency in the

Full text of DOI:10.1016/j.bmcl.2014.07.054

Bioorganic & Medicinal Chemistry Letters xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Structure-based optimization of non-peptidic Cathepsin D inhibitors
Ulrich Grädler a, , Paul Czodrowski , Christos Tsaklakidis , Markus Klein , Daniela Werkmann ,
⇑
a
a
a
a
a
b
a
Sven Lindemann , Klaus Maskos , Birgitta Leuthner
a
Merck KGaA, Merck Serono Research, Small Molecule Platform, Frankfurter Str. 250, 64293 Darmstadt, Germany
Proteros Biostructures GmbH, Bunsenstrasse 7a, 82152 Martinsried, Germany
b
a r t i c l e i n f o
a b s t r a c t
Article history:
We discovered a novel series of non-peptidic acylguanidine inhibitors of Cathepsin D as target for osteo-
arthritis. The initial HTS-hits were optimized by structure-based design using CatD X-ray structures
resulting in single digit nanomolar potency in the biochemical CatD assay. However, the most potent ana-
logues showed only micromolar activities in an ex vivo glycosaminoglycan (GAG) release assay in bovine
cartilage together with low cellular permeability and suboptimal microsomal stability. This new scaffold
can serve as a starting point for further optimization towards in vivo efficacy.
Received 30 April 2014
Revised 15 July 2014
Accepted 18 July 2014
Available online xxxx
Keywords:
Cathepsin D
Ó 2014 Elsevier Ltd. All rights reserved.
Non-peptidic inhibitor
X-ray structure
Structure-based design
Cathepsin D (CatD, EC 3.4.23.5) belongs to the large family of
aspartic proteases involved in peptide bond hydrolysis via an
acid–base mechanism mediated by two catalytic aspartates.
(1, Fig. 1) which contains the unusual amino acid (3S,4S)-
4-amino-3-hydroxy-6-methylheptanoic acid (statine) mimicking
the tetrahedral transition state of peptide catalysis. Pepstatin A
(1) was discovered in various strains of actinomyces and inhibits
1
–3
CatD is found in lysosomes of most mammalian tissues and ini-
tially synthesized as preprocathepsin in the rough endoplasmic
several aspartic proteases including pepsin (IC50 ꢀ0.1 nM), Cathep-
4
,5
12,13
reticulum. Further processing by removal of an N-terminal signal
peptide results in the 52 kDa procathepsin D (pCD) representing an
inactive zymogene at neutral pH due to propeptide segments
blocking the active site. The proenzyme is auto activated under
acidic conditions in the lysosome by removal of the 44 amino acid
N-terminal propeptide leading to a 48 kDa single chain intermedi-
ate active enzyme form. Final proteolytic cleavage results in a mix-
ture of single and two-chain molecules composed of heavy
sins D (IC50 < 1 nM) and E (IC50 < 7.5 nM).
The therapeutical
potential of pepstatin A (1) is hindered by its low intracellular pen-
etration, but recently reported conjugates demonstrated anti-
1
4
proliferative effects on tumour cell cultures. Highly potent CatD
inhibitors containing a hydroxyethylamine aspartyl protease iso-
i
stere as in 2 (K = 0.7 nM, IC50 = 85 nM) were identified by solid-
1
5
phase synthesis of a small combinatorial library. Non-peptidic
CatD inhibitors with alternative chemical scaffolds such as the
benzophenone rhodanine 3 (IC50 = 210 nM) or the diarylsulfona-
mide 4 (IC50 = 250 nM) were published by Eli Lilly and Bayer,
(
34 kDa) and light (14 kDa) chains linked by non-covalent interac-
6–9
tions in a globular structure.
1
6,17
Cathepsin D is a target for cancer, inflammation, and osteoar-
thritis. In osteoarthritis (OA), CatD seems to be directly involved
in pathophysiological events resulting in degeneration of the hya-
line cartilage. Cathepsin D is expressed on mRNA and protein level
in the chondrocytes and its catalytic activity correlates directly
with the severity of OA.10 Cathepsin D cleaves aggrecan at unique
sites resulting in a beginning degradation of the extracellular
matrix, which is necessary for the unique biomechanical properties
respectively.
In a high-throughput screening campaign conducted on our
corporate compound library, we identified a series of acylguani-
dine 5–14 as potent inhibitors of human CatD with IC50s in the
1
8
range between 440 nM and 29 lM (Table 1). This chemo type
served as interesting starting point because of its non-peptidic
scaffold with preferred lead-like properties and ligand binding effi-
ciencies (LE ꢀ0.3). In addition, the HTS-hits were inactive below
11
of hyaline cartilage of diarthrodial joints. Several CatD inhibitors
are published bearing peptidic scaffolds such as pepstatin A
30 lM against the structural closely related aspartic protease
1
9
renin. Renin controls the first step of the renin–angiotensin–
aldosterone system (RAAS) regulating the blood pressure and
2
0
was considered as an off-target. The acylguanidine moiety has
been reported as specific binder forming key H-bond interactions
⇑
Corresponding author. Tel.: +49 6151 725975; fax: +49 6151 72915975.
E-mail address: ulrich.graedler@merckgroup.com (U. Grädler).
http://dx.doi.org/10.1016/j.bmcl.2014.07.054
960-894X/Ó 2014 Elsevier Ltd. All rights reserved.
0
Please cite this article in press as: Grädler, U.; et al. Bioorg. Med. Chem. Lett. (2014), http://dx.doi.org/10.1016/j.bmcl.2014.07.054

2
U. Grädler et al. / Bioorg. Med. Chem. Lett. xxx (2014) xxx–xxx
Figure 1. Published Cathepsin D inhibitors.
to the two catalytic aspartic acids of the structurally related aspar-
tic protease BACE-1.
For further structure-based optimization, we co-crystallized a
recombinant human form of CatD (rhCatD) with the acylguanidine
improvement of binding affinity by targeting additional subpock-
ets of the large CatD binding cleft. The X-ray structure of human
CatD in complex with pepstatin A (1) revealed the specific occupa-
tion of six subpockets by the various amino acid side chains of the
inhibitor each contributing to the observed subnanomolar binding
2
1
7
and solved the structure at 2.9 Å resolution (Supplementary
2
2,23
26
Table 1).
An X-ray structure analysis revealed three specific
affinity (Fig. 2b). We determined the crystal structure of rhCatD
H-bonds between the mono substituted and presumably proton-
ated amino group of the acylguanidine 7 and the side chains of
in complex with the reference inhibitor 2 at 2.6 Å resolution
(Supplementary Table 1) for additional guidance of our structure-
based design campaign. The inhibitor 2 forms several H-bonds
within the CatD binding cleft between the following residues: (1)
statine OH-group and the catalytic Asp231/Asp33, (2) both amide
carbonyl O-atoms and Gly79/Ser80 of the flap region and (3) both
methoxy O-atoms and Ser235 (Fig. 3a). The lipophilic substituents
of the hydroxyethylamine derivative 2 are placed into six subpock-
the catalytic Asp33 and Asp231 in the active site cleft of CatD
1
(
Fig. 2a). The N -acylated amino group of 7 forms an additional
2
H-bond to Asp231, while the N -guanidine atom is involved in
an intramolecular H-bond interaction with the carbonyl O-atom
of the acyl group. Consequently, the acylguanidine moiety adopts
1
2
a pseudo 6-membered ring with a z(N ) and e(N ) configuration.
This spatial arrangement positions the 4-methylbenzyl group of
1 3
ets: the 3-phenoxy-benzyl ring in S and S , the 2-bromo-4,
0
7
into the S
3
-subpocket of CatD with lipophilic interactions to
5-dimethoxy benzamide group in S
2
and S
4
, the phthalimide ring
0
0
Ile311, Ile320 and Tyr205. The binding interactions correspond
with the initial SAR indicating a preference for lipophilic substitu-
ents in the 4-position of the benzyl group such as the most active
in S
2
and the 2,5-dichlorophenyl moiety in S
3
(Fig. 3a). The struc-
tural superimposition of CatDꢁ2 and CatDꢁ7 indicates four available
0
subpockets (S , S , S
2 3 4
and S
2
) for the introduction of additional
3
,4-chloro derivative 5 (IC50 = 440 nM). In contrast, a significant
drop in binding affinity is observed for the 4-amino derivative 14
IC50 = 29 M). The 3,4-dimethoxyphenyl substituent is bound into
the S - and S -subpockets with lipophilic contacts to Thr125,
substituents to the acylguanidine scaffold (Fig. 3b). The benzyl
0
2
position of the P -residues adjacent to the N -guanidine atom
3
(
l
(‘branching position’) offers suitable vectors to address the S
2
or
0
1
3
S
2
subpocket, which was further underlined by SZMAP calcula-
Phe131 and Ile134 (Fig. 2a). The benzyl linker of 7 is involved in
van-der-Waals interactions to Tyr78 and derivatives with an ethyl
or n-propyl linker like 12 (IC50 = 18 lM) and 13 (IC50 = 25 lM)
tions (Fig. 4). This method predicts the occurrence of water sites
based on implicit solvent Poisson Boltzmann calculations.27 SZMAP
revealed water sites with a hydrophilic character in the vicinity of
were lower active (Table 1).
the ‘branching position’ and Ser80 of the flap region, suggesting the
introduction of an amide substituent as linker to target the S or S
2 3
subpocket. At the elongation of the amide linker, a pocket formed
by Thr234, Met309 and Ile320 is found which accommodates
water sites with a hydrophobic character.
0
The biochemically most active compounds from the HTS-series
were also tested in an ex vivo assay measuring the glycosamino-
glycan (GAG) release in bovine cartilage explants. Briefly, bovine
cartilage explants were prepared from the metacarpal phalangeal
joint of 1–2 year old cows at the day of slaughter, provided from
a local slaughter house. The GAG content of the explants was
determined with a spectrophotometric 1,9-dimethylmethylene
blue shift assay and revealed high potency for pepstatin A
We prepared a library of follow-up acylguanidines according to
2
8
Scheme 1 (see Supplement).
Additional substituents in the
‘branching position’ such as phthalimide or 3,4-dimethoxybenzyl
were inspired by superimposition with the reference compound
2 and prior evaluated by docking (data not shown). Initial deriva-
tives bearing benzyl- and N-phthalimide ethyl amide substituents
in the ‘branching position’ were tested as separated enantiomers
2
4,25
(
1, IC50 = 1.5 nM).
the hydroxyethylamine reference inhibitor 2 did not show inhibi-
tion of GAG-release below 1 M. Our next goal was therefore the
However, the acylguanidines 5–9 as well as
l
Please cite this article in press as: Grädler, U.; et al. Bioorg. Med. Chem. Lett. (2014), http://dx.doi.org/10.1016/j.bmcl.2014.07.054

U. Grädler et al. / Bioorg. Med. Chem. Lett. xxx (2014) xxx–xxx
3
Table 1
Structure–activity relationship of acylguanidine HTS-series against human Cathepsin D and renin
Compound
Structure
Cathepsin D
lM)
Renin
IC50c
(lM)
IC50a
LEb
(
O
O
O
NH
5
0.44
0.3
>30
Cl
Cl
N
H
N
H
+
N
O
O
O
O
O
NH
6
1.5
0.28
>30
N
N
H
H
+
N
O
O
Cl
O
O
O
NH
7
1.9
0.31
>30
N
H
N
H
O
N
O
O
O
O
NH
8
2.7
0.25
>30
N
N
H
H
+
O
O
O
O
O
NH
9
3.0
0.27
>30
N
N
H
H
+
N
O
O
Cl
O
O
NH
1
0
6.1
0.31
>30
N
H
N
H
Cl
O
O
O
NH
1
1
10
0.26
>30
N
N
H
H
+
N
O
O
(
continued on next page)
Please cite this article in press as: Grädler, U.; et al. Bioorg. Med. Chem. Lett. (2014), http://dx.doi.org/10.1016/j.bmcl.2014.07.054

4
U. Grädler et al. / Bioorg. Med. Chem. Lett. xxx (2014) xxx–xxx
Table 1 (continued)
Compound
Structure
Cathepsin D
lM)
Renin
IC50c
(lM)
IC50a
LEb
(
O
O
O
NH
1
2
18
0.22
>30
N
H
N
H
+
N
O
O
O
NH
O
1
1
3
4
N
N
25
29
0.25
0.25
20
H
H
O
O
O
NH
>30
O
N
N
H
H
N
a
b
c
Cathepsin D biochemical assay.18
Ligand binding efficiencies (LE) based on the biochemical IC50
.
19
Renin biochemical assay.
Figure 2. (a) X-ray structure of human CatD in complex with 7 (magenta, PDB-ID: 4obz). H-bonds to the catalytic Asp231 and Asp33 are represented as dashed lines.
b) Superimposition of CatD X-ray structures in complex with 7 (magenta, PDB-ID: 4obz) and pepstatin A (yellow, PDB-ID: 1lyb, 1) and subpocket positions.
(
Please cite this article in press as: Grädler, U.; et al. Bioorg. Med. Chem. Lett. (2014), http://dx.doi.org/10.1016/j.bmcl.2014.07.054

U. Grädler et al. / Bioorg. Med. Chem. Lett. xxx (2014) xxx–xxx
5
Figure 3. (a) X-ray structure of CatD in complex with the reference inhibitor 2 (blue, PDB-ID: 4oc6, H-bonds represented as dashed lines). (b) Superimposition of CatD X-ray
structures with the HTS-hit 2 (magenta, PDB-ID: 4oc6) and the reference inhibitor 7 (cyan, PDB-ID: 4obz). The red arrows mark the substitution position at the benzyl
0
substituent of 2 for targeting the S
2
or S
2
subpockets to improve binding affinity.
with assigned stereochemistry (Table 2). Interestingly, both enan-
tiomers revealed significant potency differences in the CatD assay
binding mode, the amide group at the ‘branching position’ forms
additional H-bonds with Gly79 of the flap region and Thr234 in
0
(22: IC50 = 320 nM; 23: IC50 = 8.3
l
M) as well as against renin
the S
3
subpocket (Fig. 5a). The 3,4-methoxybenzyl substituent
(
22: IC50 = 27 M; 23: IC50 = 8.4 M). For further clarification, we
l
l
induces a side chain rotation of Met309 to accommodate the aro-
matic ring sandwiched between Met309 and the backbone amide
of Gly79 of the flap hairpin loop. Protein–inhibitor interactions
calculated with the software tool ViewContacts revealed van-
co-crystallized the analogue 24 with an improved CatD potency
(
IC50 = 58 nM) and inactivity against renin below 30 M in CatD.
l
The X-ray structure of rhCatDꢁ24 revealed an R-configuration in
0
the ‘branching position’ with the benzyl ring bound into the S
3
der-Waals contacts between the 3,4-methoxybenzyl substituent
0
29
subpocket (Fig. 5a). To our surprise, the S
2
subpocket was unoccu-
and Met309 in addition to H-donor–p contacts to Gly79 (Fig. 5a).
pied and the 3,4-methoxybenzyl amide substituent was oriented
This new crystal structure in combination with SZMAP calculations
served as template for further optimization of the acylguanidines.
0
towards the interface of the S
2
and S
3
subpockets (Fig. 5b). In this
Please cite this article in press as: Grädler, U.; et al. Bioorg. Med. Chem. Lett. (2014), http://dx.doi.org/10.1016/j.bmcl.2014.07.054

6
U. Grädler et al. / Bioorg. Med. Chem. Lett. xxx (2014) xxx–xxx
Figure 4. SZMAP calculation for the HTS hit 7. Yellow sites correspond to hydrophilic sites, whereas pink sites correspond to hydrophobic sites. Our design strategy was to
0
introduce suitable substituents into the branching position (blue circle) to fill the S
3
/S
2
pockets.
MeO
MeO
OH
MeO
MeO
MeO
MeO
Cl
O
O
NH
NH
(
a)
(b)
O
N
H
O
O
15
16
17
(
c)
F
O
S
F
F
MeO
MeO
O
N
O
O
N
H
NH
1
8
F
O
F
O
HO
O
(
f)
HN
O
(
g)
O
HN
O
(
d), (e)
NHBOC
NHBOC
+
NH3
19
20
21
Cl
O
F
O
O
NH
H
N
x TFA
O
N
H
N
H
Br
O
3
8
Scheme 1. Reagents and conditions: (a) (COCl)
2
, DCM, rt, 96%; (b) N-BOC-guanidine, 82%; DCM, DIPEA, ꢂ10 °C; (c) (CF
3
SO
2
)
2
O, TEA, ꢂ70 °C, N21) , 59%; (d) THF, DIPEA, rt;
(
e) TFA, DCM, rt, 49% for steps (d and e); (f) 4-Fluoro-3-methoxy-benzylamine xHCl, N-methyl-morpholine, EDCI, HOBt, DMF, rt, 100%; (g) 4 N HCl in dioxane, rt, 100%.
The improved biochemical CatD potency of 24 was already suf-
ficient to reach submicromolar inhibition in the GAG-assay
IC50 = 330 nM), but we also observed high in vitro clearance (Clint
improve the in vitro clearance. Substitution of the benzyl group
0
in 24 by cyclohexyl or isobutyl to more optimally fill the S
3
sub-
(
)
pocket retained CatD-potency (25: IC50 = 120 nM; 26: IC50 = 67
nM) and slightly improved human and rat Clint-values for 26
(262 and 143 lL/min/mg). Further reduction of the polar surface
values in human and rat liver microsomes (>1000 and 384 lL/min/
mg) for this derivative (Table 3).3
0–32
A sufficient stability in
human liver microsomes (Clint < 10
l
L/min/mg) was one major
area (PSA) by subsequent replacement of one or both methoxy
0
parameter in our hit-optimization strategy towards a lead candi-
date for targeting human CatD against OA. Therefore, our next
attempt was to replace metabolically labile groups in the molecule,
which are e.g. prone to demethylation or benzyl oxidation to
groups of the S
/S
2 3
benzyl amide substituent by fluorine as in
27–30 resulted in enhanced CatD-potency (30: IC50 = 24 nM). This
0
SAR well corresponds with the observed contacts of 24 in the S
/S
2 3
pocket suggesting improved van-der-Waals and/or H-donor–
p
Please cite this article in press as: Grädler, U.; et al. Bioorg. Med. Chem. Lett. (2014), http://dx.doi.org/10.1016/j.bmcl.2014.07.054

U. Grädler et al. / Bioorg. Med. Chem. Lett. xxx (2014) xxx–xxx
7
Table 2
SAR of optimized acylguanidines against human Cathepsin D and renin
Compound
Structure
Cath D
GAG (bovine)
Renin
a
IC50b
IC50c
IC50
(l
M)
(l
M)
(lM)
H
N
O
O
HN
O
NH
O
O
N
O
O
2
2
2
2
3
4
N
H
0.32
ND
27
O
O
O
O
H
N
HN
O
NH
N
N
H
8.3
ND
8.4
H
HN
N
O
O
O
NH
O
N
H
O
0.058
0.33
>30
O
O
H
N
HN
O
O
O
NH
O
2
2
2
5
6
7
N
H
O
O
0.12
ND
ND
0.5
ND
ND
16
H
N
HN
O
O
O
NH
O
O
0.067
0.170
N
H
O
O
H
N
HN
O
O
F
NH
N
H
O
H
N
HN
O
F
NH
O
O
O
2
8
N
H
O
O
O
0.098
ND
13
O
O
O
O
H
N
HN
O
F
F
NH
2
3
3
9
0
1
N
H
0.170
0.024
0.092
ND
ND
8.7
ND
H
N
HN
O
O
F
NH
0.62
ND
N
H
H
H
HN
N
N
N
N
O
NH
O
O
N
N
H
O
O
O
H
N
HN
O
NH
3
2
N
H
0.044
1.8
ND
H
N
O
N
N
N
(
continued on next page)
Please cite this article in press as: Grädler, U.; et al. Bioorg. Med. Chem. Lett. (2014), http://dx.doi.org/10.1016/j.bmcl.2014.07.054

8
U. Grädler et al. / Bioorg. Med. Chem. Lett. xxx (2014) xxx–xxx
Table 2 (continued)
Compound
Structure
Cath D
GAG (bovine)
Renin
a
IC50b
IC50c
IC50
(l
M)
(l
M)
(lM)
H
N
HN
O
NH
O
3
3
N
H
O
0.081
ND
ND
H
N
O
N
N
F
N
H
N
H
N
HN
N
N
O
N
NH
O
O
O
3
3
4
5
0.370
0.280
ND
ND
ND
ND
N
H
O
O
O
O
O
O
H
N
H
N
HN
N
N
F
O
N
NH
N
H
H
N
HN
O
NH
3
3
3
6
7
8
N
H
0.510
0.690
ND
ND
0.9
ND
ND
>30
H
N
N
N
N
N
N
H
N
HN
O
NH
O
N
H
O
H
N
O
F
N
Br
O
H
HN
N
O
O
F
NH
O
N
H
O
0.0086
Br
H
N
H
N
HN
N
O
N
NH
O
3
4
9
0
N
N
H
O
0.140
0.240
ND
ND
ND
ND
O
Br
H
N
H
N
HN
N
N
O
N
NH
O
N
H
O
O
a
b
c
Cathepsin D biochemical assay.18
25
Glycosaminoglycan (GAG) release in bovine cartilage explants (ex vivo). ND = not determined.
19
Renin biochemical assay. ND = not determined.
interactions to Met309 by fluorine atoms replacing the methoxy
groups. However, the enhanced biochemical CatD affinity of the
best fluorine derivatives did not result in potency gains in the
Met309 and Gly79 (Table 2). Interestingly, a reversed SAR was
observed for the meta- or para-substituted tetrazoles with a cyclo-
hexyl (meta 31: IC50 = 92 nM; para 32: IC50 = 44 nM) or isobutyl
GAG-assay (27: IC50 = 0.5
the microsomal stabilities of 27 and 30 were not further improved
in comparison to 24 and we still detected significant renin activity
l
M; 30: IC50 = 0.62
l
M). Furthermore,
group (meta 34: IC50 = 370 nM; para 36: IC50 = 510 nM) bound to
0
the adjacent S
3
subpocket. But still the most potent tetrazole
derivative 32 (CatD IC50 = 44 nM) showed only weak inhibition in
the GAG-assay (IC50 = 1.8 M).
1 3
Finally, we focussed our optimization efforts on the S /S sub-
(30: IC50 = 8.3
lM). A major step forward in Clint optimization was
l
the introduction of a tetrazole in the meta- or para-position of the
benzyl amide substituent targeting the S
0
/S
2 3
subpocket as exem-
pockets and achieved single digit nanomolar CatD inhibition (38:
IC50 = 8.6 nM) by the introduction of a 2-bromo-4,5-dimethoxy-
phenyl substituent as present in the reference inhibitor 2. How-
ever, 38 revealed only moderate activity in the GAG-assay
plified by 31–37. We obtained tetrazoles such as 36 with already
low mouse and rat Clint-values (<10 L/min/mg) and 37 with a
medium clearance (54 L/min/mg) in human liver microsomes
Table 3). All tetrazole derivatives revealed submicromolar CatD
l
l
(
(IC50 = 0.9
823
l
M) and high human and rat Clint-values (>1000 and
0
activity in agreement with the rhCatDꢁ24 X-ray structure indicat-
l
L/min/mg). The introduction of a tetrazole in the S
2
/S
3
sub-
ing available space for the tetrazole ring in the interface between
pocket as in 39 and 40 again improved the microsomal clearance
Please cite this article in press as: Grädler, U.; et al. Bioorg. Med. Chem. Lett. (2014), http://dx.doi.org/10.1016/j.bmcl.2014.07.054

U. Grädler et al. / Bioorg. Med. Chem. Lett. xxx (2014) xxx–xxx
9
Figure 5. (a) X-ray structure of human CatD in complex with an optimized acylguanidine 24 (green, PDB-ID: 4od9). Favourable interactions between protein and inhibitor
atoms were calculated with the software tool ViewContacts and represented as dashed, coloured lines (black: H-bonds, red: van-der-Waals contacts, blue: H-donor–
p
interactions). (b) Superimposition of CatD X-ray structures with the HTS-hit 2 (magenta, PDB-ID: 4oc6), the reference inhibitor 7 (cyan, PDB-ID: 4obz) and the optimized
acylguanidine 24 (green, PDB-ID: 4od9).
Table 3
Physicochemical properties and in vitro ADME parameters for selected compounds
Compound
Microsomal stability ClInta
Rat
(
l
L/min/mg)
Caco-2 cell permeability
app a?b (cm/s) Efflux ratio
tPSAb (Å2)
Human
Mouse
P
24
26
27
30
32
36
37
38
39
40
>1000
262
>1000
966
76
69
54
>1000
240
107
384
143
241
148
38
<10
<10
823
33
ND
ND
ND
ND
<10
<10
ND
ND
40
ND
ND
ND
ND
26
ND
34
ND
ND
ND
ND
ND
131
131
122
122
167
167
167
122
167
167
5ꢁ10e-6
ND
0.25ꢁ10e-6
ND
ND
ND
ND
17
38
ND
a
The in vitro clearance was measured in human, mouse and rat liver microsomes (ND = not determined).
The topological surface area (tPSA) was calculated with MOE.
b
41
Please cite this article in press as: Grädler, U.; et al. Bioorg. Med. Chem. Lett. (2014), http://dx.doi.org/10.1016/j.bmcl.2014.07.054

1
0
U. Grädler et al. / Bioorg. Med. Chem. Lett. xxx (2014) xxx–xxx
(
ꢀ
10–50
l
L/min/mg in mouse and rat), but also resulted in a
substrate to release fluorescence, which can then easily be quantified using a
Perkin Elmer Envision multilabel reader at Ex/Em = 340/450 nm. As assay
buffer 100 mM sodium acetate buffer pH 5.5, 0.25% CHAPS was used. The assay
20-fold drop in CatD potency. The final show stopper for further
in vivo profiling of the most active derivatives in the GAG-assay
was the low permeability. The compounds 27 (Papp a?b = 5 10e-6
cm/s; efflux ratio = 26) and especially 32 (Papp a?b = 0.25 10e-6
cm/s; efflux ratio = 34) were unable to show sufficient cell penetra-
tion, which might explain the observed drops between enzymatic
and cellular potencies (27: 3-fold; 32: 40-fold, Tables 2 and 3).
In summary, our optimization strategy resulted in acylguani-
dines with low nanomolar CatD potency together with an aspired
selectivity profile against renin. However, the most potent ana-
logues showed only micromolar activities in the explant assay
together with low microsomal stability and permeability. The polar
surface area of the most potent derivatives need to be further
reduced (<120) to improve cellular potency and intestinal perme-
was performed in
compound.
a kinetic mode. Pepstatin A was used as reference
1
9. Renin enzymatic activity was determined in a fluorescence- based assay that
utilized the quenched peptide substrate Dabcyl-g-Abu-IHPFHLVIHT-Edans
(5 lM, Anaspec, Fremont CA). 10 nM human recombinant renin (Proteos,
Kalamazoo, MI) was incubated with the substrate in 50 mM MOPS buffer, 0.1%
Igepal, 0.5% BSA at 37 °C. The assay was run in a kinetic mode. The cleavage of
the substrate was determined by release in fluorescence intensity with a
Perkin Elmer Envision multilabel reader at Ex/Em = 340/495 nm. As reference
compound renin inhibitor 2 (Sigma–Aldrich, Taufkirchen) was used.
0. Zaman, M. A.; Oparil, S.; Calhoun, D. A. Nat. Rev. Drug Disc. 2002, 1, 621.
1. Fobare, W. F.; Solvibile, W. R.; Robichaud, A. J.; Malamas, M. S.; Manas, E.;
Turner, J.; Hu, Y.; Wagner, E.; Chopra, R.; Cowling, R.; Jin, G.; Bard, J. Bioorg.
Med. Chem. Lett. 2007, 17, 5353.
2. Human recombinant Cathepsin D in complex with inhibitors was crystallized
in 35% PEG2000, 0.1 M Na-Acetat (pH 5.5) in hanging drop by vapour diffusion
at 20 °C. Crystals were cryoprotected in the mother liquor and X-ray diffraction
data were collected at the PXIII beam line equipped with a Pilatus detector at
the Paul Scherrer Institute in Villingen, Switzerland (see Supplemental
Table 1). With the detector set at 250 mm, data were collected in
2
2
2
3
3
ability. This might be achieved by diminishing the number of
nitrogen and oxygen atoms in the overall molecule including the
acylguanidine group, which forms a pseudo 6-membered ring in
the bioactive conformation. A meaningful cyclization of the
acylguanidine might result in heterocycles maintaining the crucial
H-bond interactions to the catalytic aspartates, but with favourable
pharmacokinetic properties as shown for BACE1.34
720 contiguous 0.25° oscillation images at 1 Å wavelength. Diffraction
images were reduced, integrated and the intensities were merged and scaled
using XDS.35 The structures were solved by molecular replacement using
36
Phaser version 2.3.0 in CCP4 with a previously published human CatD
structure (PDB-ID: 1LYB)37 as search model. After molecular replacement,
38
manual rebuilding in Coot followed by restrained refinements cycles using
BUSTER version 2.11.139 yielded the final structures. In the final model
building/refinement cycles, water molecules were inserted at stereo chemi-
cally reasonable sites that had features in electron density, and individual
restrained B-values were refined. The quality of the final structures was
assessed using ProCheck.4
Supplementary data
0
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.bmcl.2014.07.
0
2
3. The atomic coordinates and structure factors for the human CatD structures in
complex with 2, 7 and 24 have been deposited in the Protein Data Bank
54.
(www.rcsb.org) under the accession codes 4OC6, 4OBZ and 4OD9, respectively.
2
2
4. Klompmakers, A. A.; Hendriks, T. Anal. Biochem. 1986, 153, 80.
5. To activate Cathepsin D in the cartilage, the explants were incubated for 3 days
Reference and notes
2
at pH 5.5 at 37 °C and 7.5% CO . The pH in DMEM medium (+50 lg/ml ascorbic
acid) was adjusted with MES (2-Morpholinethan–sulfonic acid Monohydrat).
During the 3 days incubation time, the explants were treated with or without
the potential CatD inhibitors. After 3 days the amount of sulphated GAGs as a
major part of the cartilage were determined in the supernatant of explants.
Increasing amount of GAG in the supernatant is a sign of cartilage destruction
and should be reduced by the treatment of CatD inhibitors.
1
2
3
4
5
6
7
8
9
.
.
.
.
.
.
.
.
.
Barrett, A. J. Adv. Exp. Med. Biol. 1977, 95, 291.
Davies, D. R. Annu. Rev. Biophys. Biophys. Chem. 1990, 19, 189.
Tang, J.; Wong, R. N. J. Cell. Biochem. 1987, 33, 53.
Hasilik, A.; Neufeld, E. F. J. Biol. Chem. 1980, 255, 4937.
Kornfeld, S. Biochem. Soc. Trans. 1990, 18, 367.
Conner, G. E.; Richo, G. Biochemistry 1992, 31, 1142.
Erickson, A. H.; Conner, G. E.; Blobel, G. J. Biol. Chem. 1981, 256, 11224.
Lee, A. Y.; Gulnik, S. V.; Erickson, J. W. Nat. Struct. Biol. 1998, 5, 866.
Richo, G. R.; Conner, G. E. J. Biol. Chem. 1994, 269, 14806.
26. Majer, P.; Collins, J. R.; Gulnik, S. V.; Erickson, J. W. Protein Sci. 1997, 6, 1458.
27. SZMAP, version 2.1.0, OpenEye Scientific Software, Santa Fe, USA.
28. Feichtinger, K.; Sings, H. L.; Baker, T. J.; Matthews, K.; Goodman, M. J. Org. Chem.
1998, 63, 8432.
1
0. Konttinen, Y. T.; Mandelin, J.; Li, T. F.; Salo, J.; Lassus, J.; Liljestrom, M.;
29. Kuhn, B.; Fuchs, J. E.; Reutlinger, M.; Stahl, M.; Taylor, N. R. J. Chem. Inf. Model.
2011, 51, 3180.
Hukkanen, M.; Takagi, M.; Virtanen, I.; Santavirta, S. Arthritis Rheum. 2002, 46,
9
53.
30. Barter, Z. E.; Bayliss, M. K.; Beaune, P. H.; Boobis, A. R.; Carlile, D. J.; Edwards, R.
J.; Houston, J. B.; Lake, B. G.; Lipscomb, J. C.; Pelkonen, O. R.; Tucker, G. T.;
Rostami-Hodjegan, A. Curr. Drug Metab. 2007, 8, 33.
1
1. Handley, C. J.; Mok, M. T.; Ilic, M. Z.; Adcocks, C.; Buttle, D. J.; Robinson, H. C.
Matrix Biol. 2001, 20, 543.
1
1
2. Marciniszyn, J., Jr.; Hartsuck, J. A.; Tang, J. J. Biol. Chem. 1976, 251, 7088.
3. Umezawa, H.; Aoyagi, T.; Morishima, H.; Matsuzaki, M.; Hamada, M. J. Antibiot.
31. Davies, B.; Morris, T. Pharm. Res. 1993, 10, 1093.
32. Iwatsubo, T.; Suzuki, H.; Sugiyama, Y. J. Pharmacol. Exp. Ther. 1997, 283, 462.
33. Clark, D. E. Future Med. Chem. 2011, 3, 469.
(
Tokyo) 1970, 23, 259.
1
4. Maynadier, M.; Vezenkov, L. L.; Amblard, M.; Martin, V.; Gandreuil, C.; Vaillant,
O.; Gary-Bobo, M.; Basile, I.; Hernandez, J. F.; Garcia, M.; Martinez, J. J.
Controlled Release 2013, 171, 251.
34. Oehlrich, D.; Prokopcova, H.; Gijsen, H. J. Bioorg. Med. Chem. Lett. 2014, 24,
2033.
35. Kabsch, W. Acta Crystallogr., Sect. D 2010, 66, 125.
1
1
5. Lee, C. E.; Kick, E. K.; Ellman, J. A. J. Am. Chem. Soc. 1998, 120, 9735.
6. Whitesitt, C. A.; Simon, R. L.; Reel, J. K.; Sigmund, S. K.; Phillips, M. L.; Kevin
Shadle, J.; Heinz, L. J.; Koppel, G. A.; Hunden, D. C.; Lifer, S. L.; Berry, D.; Ray, J.;
Little, S. P.; Xiadong, L.; Marshall, W. S.; Panetta, J. A. Bioorg. Med. Chem. Lett.
36. McCoy, A. J.; Grosse-Kunstleve, R. W.; Adams, P. D.; Winn, M. D.; Storoni, L. C.;
Read, R. J. J. Appl. Crystallogr. 2007, 40, 658.
37. Baldwin, E. T.; Bhat, T. N.; Gulnik, S.; Hosur, M. V.; Sowder, R. C.; Cachau, R. E.;
Collins, J.; Silva, A. M.; Erickson, J. W. Proc. Natl. Acad. Sci. U.S.A. 1993, 90, 6796.
38. Emsley, P.; Cowtan, K. Acta Crystallogr., Sect. D 2004, 60, 2126.
39. Bricogne, G.; Blanc, E.; Brandl, M.; Flensburg, C.; Keller, P.; Paciorek, W.;
Roversi, P.; Sharff, A.; Smart, O.S.; Vonrhein, C.; Womack, T.O., In: 2011.
40. Laskowski, R. A.; MacArthur, M. W.; Moss, D. S.; Thornton, J. M. J. Appl.
Crystallogr. 1993.
1
996, 6, 2157.
1
1
7. Dumas, J.; Brittelli, D.; Chen, J.; Dixon, B.; Hatoum-Mokdad, H.; König, G.;
Sibley, R.; Witowsky, J.; Wong, S. Bioorg. Med. Chem. Lett. 1999, 9, 2531.
8. Cathepsin
utilized the preferred CatD substrate sequence GKPILFFRLK(Dnp)-D-R-NH2
labeled with MCA (5 M, Enzo Life Sciences, Lörrach). CatD isolated from
human liver (1.4 nM, Sigma–Aldrich, Taufkirchen) cleaves the synthetic
D activity was determined in a fluorescence-based assay that
l
41. Molecular Operating Environment (MOE), 2. 1.; Chemical Computing Group, 1.
S. St. W. S. #. M. Q. C. H. 2. 2.
Please cite this article in press as: Grädler, U.; et al. Bioorg. Med. Chem. Lett. (2014), http://dx.doi.org/10.1016/j.bmcl.2014.07.054