Discovery of Highly Potent and Selective Small-Molecule Reversible Factor D Inhibitors Demonstrating Alternative Complement Pathway Inhibition in Vivo

Article abstract of DOI:10.1021/acs.jmedchem.7b00425

The highly specific S1 serine protease factor D (FD) plays a central role in the amplification of the complement alternative pathway (AP) of the innate immune system. Genetic associations in humans have implicated AP activation in age-related macular degeneration (AMD), and AP dysfunction predisposes individuals to disorders such as paroxysmal nocturnal hemoglobinuria (PNH) and atypical hemolytic uremic syndrome (aHUS). The combination of structure-based hit identification and subsequent optimization of the center (S)-proline-based lead 7 has led to the discovery of noncovalent reversible and selective human factor D (FD) inhibitors with drug-like properties. The orally bioavailable compound 2 exerted excellent potency in 50% human whole blood in vitro and blocked AP activity ex vivo after oral administration to monkeys as demonstrated by inhibition of membrane attack complex (MAC) formation. Inhibitor 2 demonstrated sustained oral and ocular efficacy in a model of lipopolysaccharide (LPS)-induced systemic AP activation in mice expressing human FD.

Full text of DOI:10.1021/acs.jmedchem.7b00425

Subscriber access provided by EAST TENNESSEE STATE UNIV
Article
Discovery of highly potent and selective small-molecule reversible Factor D
inhibitors demonstrating alternative complement pathway inhibition in vivo
Edwige Lorthiois, Karen Anderson, Anna Vulpetti, Olivier Rogel, Frederic Cumin, Nils
Ostermann, Stefan Steinbacher, Aengus Mac Sweeney, Omar Delgado, Sha-Mei Liao,
Stefan Randl, Simon Rüdisser, Solene Dussauge, Kamal Fettis, Laurence Kieffer, Andrea de
Ekernez, Louis Yang, Constanze Hartwieg, Upendra A Argikar, Laura R. LaBonte, Ronald
Newton, Viral Kansara, Stefanie Flohr, Ulrich Hommel, Bruce Jaffee, and Jürgen Maibaum
J. Med. Chem., Just Accepted Manuscript • Publication Date (Web): 16 Jun 2017
Downloaded from http://pubs.acs.org on June 17, 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 77
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
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 77
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
Discovery of highly potent and selective smallꢀ
molecule reversible Factor D inhibitors
demonstrating alternative complement pathway
inhibition in vivo
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
a
a
a
Edwige Lorthiois, Karen Anderson, Anna Vulpetti, Olivier Rogel, Frederic Cumin, Nils
a
c
a
b
b
Ostermann, Stefan Steinbacher, Aengus Mac Sweeney, †† Omar Delgado, ShaꢀMei Liao,
a
a
a
a
a
Stefan Randl, † Simon Rüdisser, Solene Dussauge, Kamal Fettis, Laurence Kieffer, Andrea
b
b
a
b
b
de Erkenez, Louis Yang, Constanze Hartwieg, Upendra A. Argikar, Laura R. La Bonte,
b
b
a
a
b
Ronald Newton, Viral Kansara, Stefanie Flohr, Ulrich Hommel, Bruce Jaffee, Jürgen
a*
Maibaum
a
Novartis Pharma AG, Novartis Institutes for BioMedical Research, Novartis Campus, CHꢀ4056
b
Basel, Switzerland, Novartis Pharma AG, Novartis Institutes for BioMedical Research,
c
Cambridge, Massachusetts 02139, USA, Proteros Biostructures GmbH, Dꢀ82152 Planeggꢀ
Martinsried, Germany
KEYWORDS: Alternative Complement Pathway, Factor D inhibitors, S1 serine protease,
structureꢀbased drug design, MAC deposition, ageꢀrelated macular degeneration.
ACS Paragon Plus Environment

Page 3 of 77
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
ABSTRACT: The highly specific S1 serine protease Factor D (FD) plays a central role in the
amplification of the complement alternative pathway (AP) of the innate immune system. Genetic
associations in humans have implicated AP activation in ageꢀrelated macular degeneration
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
(
AMD), and AP dysfunction predisposes individuals to disorders such as paroxysmal nocturnal
hemoglobinuria (PNH) and atypical hemolytic uremic syndrome (aHUS). The combination of
structureꢀbased hit identification and subsequent optimization of the center (S)ꢀprolineꢀbased
lead 7 has led to the discovery of nonꢀcovalent reversible and selective human Factor D (FD)
inhibitors with drugꢀlike properties. The orally bioavailable compound 2 exerted excellent
potency in 50% human whole blood in vitro and blocked AP activity ex vivo after oral
administration to monkeys as demonstrated by inhibition of membrane attack complex (MAC)
formation. Inhibitor 2 demonstrated sustained oral and ocular efficacy in a model of
lipopolysaccharide (LPS)ꢀinduced systemic AP activation in mice expressing human FD.
Introduction
The alternative pathway (AP) of the complement system, when triggered by pathogenꢀ or
damageꢀassociated molecular patterns on nonꢀhost and injured host cells, provides an innate
1
,2
immune defense mechanism in all vertebrates. The AP also functions as an amplification
mechanism for both the classical and the lectin pathways. A low level of steady complement
activation by AP “tickꢀover” results in the generation of a small fraction of soluble C3ꢀ
convertase in plasma (Figure 1). Upon activation, the initial C3ꢀconvertase cleaves complement
C3 into the anaphylatoxin C3a and the larger fragment C3b, which binds to cell surfaces
(opsonization) and recruits Factor B (FB), a proꢀenzyme of the S1 serine protease family. FD
cleaves FB in its complex with C3b thereby generating the active surfaceꢀbound C3ꢀconvertase
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 77
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
(
C3bBb). C3bBb stabilized by properdin (P) cleaves additional molecules of C3, thus triggering
a local amplification of AP activation, and moreover by recruitment of additional C3b generates
the C5ꢀconvertase (C3bC3bBb). C3ꢀ and C5ꢀconvertases are key players in the alternative
complement pathway ultimately leading to the formation of the membrane attack complex
(MAC) and cell lysis. Dysregulation of AP and activation against the host has been connected to
a variety of diseases such as ageꢀrelated macular degeneration (AMD), atypical hemolytic
uremic syndrome (aHUS), membranoproliferative glomerulonephritis type II (MPGNII) and
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
,3
paroxysmal nocturnal hemoglobinuria (PNH). The S1 protease FD plays a central function in
AP amplification by catalyzing the first and rateꢀlimiting proteolytic step. FD is highly substrate
specific due to a unique active site architecture that requires an inducedꢀfit conformational
transition to acquire full proteolytic activity. The extremely high specificity of FD is essential as
it circulates in the blood as a mature enzyme, albeit adopting a latent active site conformation,
4
ꢀ6
and is not regulated by any known endogenous inhibitor.
Figure 1: Schematic drawing of the classical, the lectin and the alternative complement
pathways, highlighting the pivotal role of the S1 serine protease Factor D. An inhibitor of Factor
ACS Paragon Plus Environment

Page 5 of 77
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
D has the potential to block the initiating step of the alternative pathway (AP) and the
amplification loop that augments the other two pathways, and prevent the formation of the AP
C3 and C5 convertases. The key effector mechanisms of innate immunity contributed by
complement are opsonization by C3 fragments and lysis by deposition of the membrane attack
complex (MAC) on cell surfaces.
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
As a consequence, FD inhibition appears to be an attractive mechanism to block the complement
AP with high specificity and efficiency. For this reason, we had embarked in 2006 on an
extensive drug discovery effort in the quest to develop orally effective, reversible nonꢀcovalent
7
,8
FD inhibitors useful for clinical therapy. When we initiated our project, the only reported FD
inhibitors
were
nonꢀselective
and
covalentꢀirreversible,
such
as
2ꢀ
9
carbamimidoylbenzo[b]thiophenꢀ6ꢀyl thiopheneꢀ2ꢀcarboxylate (BCXꢀ1470), properties that
were considered as incompatible with selective and sustained AP inhibition required for disease
4
a
therapy, in particular in view of the short plasma halfꢀlife of FD. Furthermore, no small
1
0,11
molecule AP inhibitor has become available to date for clinical use.
Our multiple parallel
approaches for initial hit identification included structureꢀbased S1 target family scaffold
hopping and tailored library design, focused fragmentꢀbased screening (FBS), screening of
diverse Novartis corporate fragment libraries and functionꢀbased highꢀthroughput screening
using assays for FD thioesterolysis and for APꢀmediated MAC deposition. Subsequent merging
of newly discovered structural motifs essential for binding to the unique conformation of the
latent FD active site together with further structure activity relationship (SAR) optimization
7
eventually culminated in the center (S)ꢀprolineꢀbased FD inhibitors 1 and 2 (Chart 1). These
compounds were shown to efficiently block alternative pathway (AP) activation and to prevent
7
lysis of PNH erythrocytes in vitro. We describe herein a more detailed account of our work
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 77
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
leading to the novel inhibitor 2 with promising in vitro selectivity and safety profiles, distribution
into ocular tissues and demonstrating in vivo and ex vivo efficacy after oral administration in
animal models of complement activation. Smallꢀmolecule potent and selective FD inhibitors
were very recently reported by researchers at Achilion, however, without disclosing the origin
and chemical structure of these compounds. ACHꢀ4471 which has entered into early clinical
investigation was shown to block ex vivo induced AP activity in cynomolgus monkeys after oral
dosing, to reduce complementꢀmediated hemolysis of PNH erythrocytes and to block
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
2
dysregulated AP activity in vitro in aHUS patient serum.
Cl
F
H
N
O
H
N
O
N
N
N
Br
O
N
N
O
N
N
O
O
N
H N
H2N
2
1
2
S1'
H
H
N
N
O
N
H
Merging with S2'
binding motif from 3
N
N
HN
HN
HN
N
OCF3
OCF3
O
O
O
O
O
O
N
N
S2'
O
S1
4 : IC50 = 5.8
M
5 : IC50 = 20
M
3
: IC50 = 14
M
BEI: 11.6 / LLE : 1.6
BEI: 10.5 / LLE : 1.3
BEI: 10.5 / LLE : 1.4
CO H
2
O
H N
2
N
O
H
S1
6
: NMR KD = 1600
M
BEI: 9.0 / LLE : 1.1
Chart 1.Chemical structures of compounds 1–6.
ACS Paragon Plus Environment

Page 7 of 77
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Results and Discussion
7
,8
As previously reported, the center (S)ꢀprolineꢀbased FD inhibitor 3 (Chart 1), was identified
from a structureꢀbased library design approach inspired by a structurally related analog binding
to kallikrein 7, a member of the S1 protease family, and demonstrated moderate potency in
blocking the thioesterolytic activity of latent human FD. Inhibitor 3 was shown by Xꢀray
crystallography to bind in a bent conformation to the S1, S1’ and S2’ pockets of FD, in which the
salt bridge between Asp182 and Arg218 at the bottom of the S1 pocket is present (closed
conformation). Initial broad SAR exploration of the various portions of molecule 3, in part by
iterative small library design, failed to deliver any analogs with significantly improved in vitro
potencies in the FD thioesterolysis assay (data not shown). Gratifyingly, the Nꢀurea (S)ꢀproline
analog 4 (Chart 1) was uncovered by structureꢀsimilarity searches in the corporate compound
collection and found to be a modest inhibitor of FD in the thioesterolysis assay. Computational
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
7
modeling predicted the Nꢀmethylꢀ1Hꢀindolꢀ3ꢀyl pharmacophore to bind into the S1 pocket.
However, the generation of an Xꢀray crystal structure of 4 in complex with recombinant human
FD by both crystal soaking and coꢀcrystallization remained unsuccessful in our hands, despite
significant efforts. We therefore prepared its close analog 5, in which the bulky
diphenylmethaneꢀamido residue is replaced by the less hydrophobic metaꢀOCF ꢀanilide motif of
3
3 (Chart 1). Although compound 5 was ~4ꢀfold less potent toward FD, as compared to 4, the Xꢀ
ray crystal structure of 5 bound to the active site of FD could be eventually resolved (Figure 2A).
The electron density map was consistent with a single binding conformation of 5, revealing the
indole scaffold to be positioned in the S1 pocket and the Nꢀmethyl group pointing toward the
side chain of Arg218 at the bottom of this binding site. The (S)ꢀproline and the metaꢀOCF ꢀanilide
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 77
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
portion of 5 occupied the S1’ꢀS2’ specificity pockets in a very similar manner as observed for the
7
parent inhibitor 3 (PDB code: 5FBE).
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
Figure 2. A. Crystal structure of the complex between human FD and 5 (in pink) (PDB code:
5NAT) overlaid with FDꢀbound 6 (in cyan) (PDB code: 5FBI). Hydrogenꢀbond interactions of
the urea carbonyl to the Gly193 NH (oxyanion hole) and the amide NH of compound 5 to the
backbone carbonyl of Leu are highlighted, as well as the Hꢀbonds of the terminal carboxamide
4
1
of 6 to Thr214 of the selfꢀinhibitory loop and to Arg218 at the bottom of the S1 pocket, and the Hꢀ
bond of the indoleꢀNH with the Arg218 backbone carbonyl. Note that the benzoic acid portion of
6
is positioned in solvent space, and hence is not resolved in the crystal structure. B. Crystal
structure of the complex between human FD and 7 (in orange) (PDB code: 5NAR). The
imidazole side chain of the catalytic His is in an ”out” conformation. The network of Hꢀbonds
5
7
between the Nꢀcarbamoyl residue of 7 and Thr , Arg and a water molecule bound deeply in
2
14
218
the S1 pocket is highlighted.
ACS Paragon Plus Environment

Page 9 of 77
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
The indoleꢀcarboxamide fragment 6 (Chart 1), which had previously been identified through
1
15
screening of a focused set of fragments, was shown by Hꢀ N HSQCꢀNMR to bind with very
weak affinity to FD, and by Xꢀray crystallography to be accommodated by the S1 pocket of
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
7,8
latent FD in its closed conformation. Remarkably, the indole moiety of both 6 and 4 adopted a
very similar planar orientation within the S1 pocket and sandwiched between the Lys192 side
chain on one side and the selfꢀinhibitory loop on the other side (Figure 2A). Moreover, the
closely overlapping positions of the indole nitrogen atom of 4 and the C2 atom of 6, as well as
the respective bond vectors, suggested combining these structural motifs (Figure 2). As a
consequence, compound 7 bearing an Nꢀcarbamoyl substituted indole was synthesized and
demonstrated a remarkable 40ꢀfold increase in inhibitory activity in the FD thioesterolysis assay,
1
3ꢀ16
as compared to 5 (Table 1). The Binding Efficiency Index (BEI)
and the Lipophilic Ligand
17ꢀ20
Efficiency (LLE)
for 7 (BEI: 13.3; LLE: 3.54) were improved compared to that calculated
for compounds 3–6 (Chart 1). Importantly, compound 7 revealed similar potency with an IC50
value of 0.8 ꢀM in the FD proteolysis assay using human FB in complex with cobra venom
factor (CVF:FB), which is an ortholog of the activated human C3, as the endogenous surrogate
7
substrate. The (2R)ꢀconfigured enantiomer 8 was 60ꢀfold less active in the FD thioesterolytic
assay (Table 1). The Xꢀray crystal complex of 7 bound to FD (Figure 2B) was determined at 1.55
Å resolution using the soaking method. The Nꢀcarbamoyl group of 7, being key for strong
binding affinity, is positioned deeply in the S1 pocket, as expected, and is engaged in direct Hꢀ
bond interactions with the terminal guanidine nitrogen of the Arg218 side chain and the backbone
carbonyl of Thr214 in a similar manner as was previously observed in the crystal structure of
fragment 6 bound to FD (Figure 2A). In addition, one of the amide NH is in Hꢀbond distance to
an enzymeꢀbound water molecule positioned in a polar environment comprising the Ile227 amide
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 77
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
backbone and the hydroxyl of the Thr214 side chain. An additional interaction between the NH of
the urea S1ꢀS1’ spacer in 7 and the Ser215 backbone carbonyl is also observed. The
corresponding (S)ꢀproline scaffold and the OCF ꢀanilide moiety of both 7 and 3 are closely
3
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
superpositioned in the S1’ꢀS2’ prime sites (not shown).
Compound 7 showed excellent selectivity against a panel of 10 human S1 proteases including
kallikrein 7, chymotrypsin, trypsin, coagulation factor Xa, thrombin, plasmakallikrein and
neutrophil elastase (all IC values >30 ꢀM), and furthermore against a standard safety panel of
50
2
6 enzymes, receptors, ion channels and transporters (IC values >30 ꢀM; weak affinities for
5
0
adenosine 3 receptor (Ad3: 8.8 ꢀM), adenosine transporter (AdT: 4.2 ꢀM) and cycloꢀoxygenase
(COXꢀ1: 6.4 ꢀM)). Also, 7 did not inhibit the hERG channel (IC >30 ꢀM; dofetilide binding
1
5
0
2
1
22
assay) and showed no signs for in vitro genotoxicity (miniꢀAMES and Micronucleus test ). No
inhibition of the cytochrome P450 enzymes CYP3A4 and CYP2D6 (IC >30 ꢀM) and only
5
0
weak inhibitory activity against CYP2C9 (IC = 3.4 ꢀM) was observed. Inhibitor 7 displayed
5
0
modest solubility, medium permeability in the Cacoꢀ2 assay (Table 4) and moderate in vitro
clearance in rat and human liver microsomes (not shown). Importantly, the compound displayed
a promising in vivo pharmacokinetic profile in SpragueꢀDawley (SD) rats (dosed at 1 mg/kg
intravenously (IV) and at 10 mg/kg orally (PO)), characterized by moderate clearance (CL = 21
mL/min/kg), a terminal halfꢀlife (T ) of 1.4 h, sustained micromolar exposure levels in plasma
1
/2
(AUC0ꢀ8 h = 4990 nMh; Cmax = 1167 nM) and 25% oral bioavailability. Based on its overall
properties, we considered compound 7 as a reasonable starting point for further SAR exploration.
Our initial efforts were aimed at optimizing the binding interactions to the S1, S1’ and S2’
8
pockets considered as hot spots in the active site of latent FD, rather than expanding into the
13ꢀ16
17ꢀ20
S1β region, with the objective to improve potency not at the expense of BEI
and LLE.
ACS Paragon Plus Environment

Page 11 of 77
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
We reasoned that additional contacts to the S1’ site, in particular to the small indentation defined
by the His backbone, the Cys ꢀCys disulfide bridge and the Ser side chain (Figure 2),
57
42
58
215
would improve FD binding. Accordingly, alterations of the (S)ꢀproline scaffold were envisaged
by introducing small hydrophobic substituents to C4 and/or C5 of the center ring with
appropriate absolute configuration (Table 1). Introduction of a methyl group to the 4ꢀposition of
the proline, either with absolute (R)ꢀ or (S)ꢀconfiguration (compounds 9 and 10, respectively),
and likewise the incorporation of a fused transꢀ3,4ꢀcyclopropylꢀproline (11), did not significantly
influence the in vitro potency, as compared to 7, when tested in the biochemical FD assay.
Replacing the pyrrolidine by a piperidine center ring resulted in a dramatic loss in potency (12).
In contrast, and very gratifyingly, attaching a fused transꢀ4,5ꢀcyclopropyl ring to the proline
scaffold gave a 10ꢀfold improvement in potency with 13 displaying an IC value of 50 nM. The
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
5
0
potency of 13 was retained when measured in the CVF:FB proteolytic ELISA assay (IC = 0.21
5
0
ꢀM), and moreover 13 inhibited AP activation in vitro in a functional membrane attack complex
7
(
MAC) deposition assay using 50% human whole blood, albeit with a 20ꢀfold decrease in
activity (Table 2).
To inform further medicinal chemistry design, we determined the crystal structure of 13 in
complex with FD (Figure 3A). The overall binding pose of 13 emulating the Uꢀshaped active site
conformation of latent FD extending from S1 to S2’ was confirmed, with the His57 again
adopting a ‘nonꢀcanonical’ catalyticꢀtriad conformation. The (S)ꢀproline ring portion of the
bicyclic azaꢀbicyclo[3.1.0]hexane scaffold adopted a very similar bound conformation as
observed for the unsubstituted analog 7 (Figure 3A). The axial methylene group of the
cyclopropyl ring is oriented toward the backꢀsurface of the S1’ site and forms close van der
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 77
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
Waals contacts by nicely filling the small hydrophobic indentation which is formed as a
consequence of the flipped conformation of the His side chain.
57
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
Table 1. In vitro Potency Data for Compounds 7–18, using a FD Thioesterolytic Fluorescent
Assay.
FD
IC (ꢀM)
Entry
Cpd
7
a,b
5
0
1
2
3
4
5
6
7
8
0.50
8
30
9
0.85
0.30
0.65
22
10
11
12
13
14
0.05
0.05
ACS Paragon Plus Environment

Page 13 of 77
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
9
15
16
17
18
0.70
0.45
1.4
10
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
1
1
2
0.50
a
Half–maximal inhibition of recombinant human complement FD catalytic domain, as
b
determined in a thioesterolysisꢀbased assay using ZꢀLysꢀthiobenzylester as the substrate. Data
represent mean values of duplicate measurements.
To explore the impact of the C4ꢀsubstitution and stereochemistry by introducing a fluorine atom
to the center pyrrolidine, the (R)ꢀ and (S)ꢀconfigured monoꢀfluoro analogs 14 and 15, as well as
the 4,4ꢀdifluoro derivative 16, were synthesized. A 10ꢀfold improvement in potency was
accomplished for (4R)ꢀ14, whereas both the (4S)ꢀdiastereomer 15 and the difluoro analog 16
were found to be equipotent to the parent compound 7 (Table 1). The stereoꢀdependent effects of
a fluorine and other heteroatoms substituting the C3 and C4 positions of Nꢀacylated prolineꢀ
23
carboxamides on ring puckering is well documented. A fluorine at position C4 of the proline
alters the relative energies of the ring puckers, which can lead to the nearly exclusive population
of a single conformer. The C ꢀendo conformation is preferred for 4(S)ꢀfluoroꢀproline monoꢀ
4
substituted derivatives, whereas 4(R)ꢀfluoroꢀprolines adopt the C ꢀexo conformation which in
4
case of (4R)ꢀ14 is expected to direct the fluorine atom toward the S1’ backꢀsurface, thereby
enhancing the van der Waals interactions to the protein. Moreover, the presence of the (4R)ꢀ
fluorine results in a preference for the transꢀpeptidic bond of the Nꢀacyl proline resembling the
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 77
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
23
FD bound conformation observed for inhibitors 7 and 13. The Xꢀray crystal structure of 14
bound to FD (Figure 3A) indeed revealed the proline ring in a C ꢀexo pucker, thereby positioning
4
the Fꢀatom close to the Cys ꢀCys disulfide bridge and the His backbone carbonyl. The antiꢀ
4
8
52
57
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
parallel dipoleꢀdipole interaction observed between the large CꢀF dipole and the His carbonyl
5
7
dipole may further contribute to the improved binding affinity of 14 (Figure 3B). Vice versa, the
fluorine of (4S)ꢀ15 in its anticipated C ꢀendo conformation is directed toward solvent space
4
without being involved in direct binding interactions.
Next, we introduced polar substituents to C4 of the proline ring with the aim to improve aqueous
solubility while retaining in vitro potency. Prolineꢀamides bearing a (4S)ꢀhydroxy or (4S)ꢀamino
2
3
group have been shown to preferentially adopt a C ꢀendo ring pucker, in line with our design
4
concept to place a solubilizing group close to solvent space. Compound 17 bearing a (4S)ꢀOH
showed a ~3ꢀfold drop in biochemical potency, whereas the more soluble (4S)ꢀNH analog 18
2
exhibited similar affinity for FD compared to 7, despite the slightly reduced van der Waals
contacts of the proline endoꢀpucker to the S1’ hydrophobic surface observed in the FD/18ꢀ
complex Xꢀray structure (depicted in Figure 3A).
ACS Paragon Plus Environment

Page 15 of 77
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Figure 3.A. Superpositions of compounds 7 (in orange color), 13 (magenta) (PDB code:
5NAW), 14 (cyan) (PDB code: 5NBA) and 18 (yellow) (PDB code: 5NB6) observed in their Xꢀ
ray crystal structures in complex to human FD. The conformational flexibility of the His and
5
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
Ser215 side chains are highlighted, while the Cys ꢀCys disulfide bridge conformation is largely
42
58
conserved. B. Closeꢀup view highlighting the antiꢀparallel dipoleꢀdipole interaction observed
between the CꢀF dipole in 14 and the FD His carbonyl dipole.
5
7
We next focused our efforts on exploring the S2’ region identified as hydrophobic hotspot
8
defined by Gly142 and the side chains of Ile143 and Arg . The trifluoromethoxyꢀphenyl residue
151
24ꢀ26
in its energetically favoured perpendicular conformation
positions the highly lipophilic CF3
moiety deeply into the S2’ pocket, thereby tightly filling the small hydrophobic indentation of
the enzyme (Figure 4A). Replacement of the OCF group in 13 by a OCHF (19) decreased
3
2
potency by 3ꢀfold, whereas the OCH analog 20 showed a more pronounced 6ꢀfold potency drop
3
2
4ꢀ26
likely due to the favoured inꢀplane conformation of the methoxyꢀanilide portion
(Table 2).
Halogen replacements were tolerated with a preference for bromine over a chlorine atom (22 vs.
21), reflecting the increase in strength of the halogen bond to the backbone carbonyl of Trp141 as
2
7ꢀ30
a result of the increase in halogen size (vide infra for a more detailed discussion).
Table 2. In vitro Potency Data for Compounds 13, 19–26, using a FD Thioesterolytic
Fluorescent Assay and a MAC Deposition Assay.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 77
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
FD
IC (ꢀM)
MAC
1
Entry
Cpd
13
19
20
21
22
23
24
25
26
Y
R
e
a,c
b,c
hPPB (%)
IC (ꢀM)
50
50
1
2
3
4
5
6
7
8
9
CH
CH
CH
CH
CH
CF
CF
N
OCF3
OCHF2
OCH3
Cl
0.05
0.15
1.25
nd
d
nd
nd
nd
0.32
nd
0.15
nd
nd
Br
0.045
0.012
0.023
0.014
0.006
nd
nd
OCF3
CF3
0.250
0.587
0.022
0.009
nd
>99
95.2
96.4
CF3
N
Br
a
Half–maximal inhibition of recombinant human complement FD catalytic domain, as
b
determined in a thioesterolysisꢀbased assay using ZꢀLysꢀthiobenzylester as the substrate; Half–
maximal inhibition of soluble MAC complex formation, as determined in an ELISA assay using
alkalineꢀphosphataseꢀconjugatedꢀantiꢀC9ꢀneoepitope in normal human whole blood diluted with
ACS Paragon Plus Environment

Page 17 of 77
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
c
an equal amount of gelatine veronal buffer; Data represent mean values of duplicate
d
e
measurements; Not determined; Compound fraction in % bound to human plasma protein
.
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
The concept of introducing an additional fluorine atom at the 2ꢀposition of the P2’ anilide moiety
of 13 derived from SAR data for structurally related prolineꢀbased FD inhibitors, such as 1,
7
bearing orthoꢀF substituted benzylamides at P2’. Compound 23 was found to be 4ꢀtimes more
potent than the desꢀfluoro analog 13 in both the thioesterolysis assay and in the MAC deposition
assay (Table 2). The gain in potency could be explained by the conformational preference of the
anilide moiety induced by the presence of the orthoꢀF adjacent to the amidocarbonyl spacer,
2
3,31
properly directing the metaꢀOCF into the S2’pocket.
The conformation with the fluorine cis
3
to the amide carbonyl is destabilized by electrostatic repulsion, while the preferred trans
conformation is favoured by the formation of the NH/CF 1,3ꢀantiperiplanar dipolar interaction.
In addition, the improvement in potency observed upon introduction of the orthoꢀF might also be
the result of the increased aniline NꢀH acidity, thereby strengthening the Hꢀbond with the Leu41
backbone carbonyl.
Replacement of the 2ꢀfluoroꢀaniline by a 2ꢀaminopyridine residue was designed to have a similar
effect on the ligand conformation due to a repulsion between the nitrogen atom of the pyridine
and the carbonyl of the amide P1’ꢀP2’ linker. Compounds 24 bearing a 2ꢀF,3ꢀCF ꢀaniline P2’
3
motif and 25 bearing a 6ꢀCF ꢀpyridine were found to be equipotent in the biochemical assay
3
(Table 2). Notably, in the MAC deposition assay performed using 50% human whole blood, a
lower ‘IC shift’ was observed for 25 compared to 24, which could be attributed to the lower %
50
fraction of compound 25 bound to human plasma protein (hPPB) as compared to 24 (Table 2).
Most strikingly, the 6ꢀbromoꢀpyridine analog 26 exhibited singleꢀdigit nanomolar potency in
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 77
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
both assays in line with the moderate hPPB of 26 and consistent with our more general
observation that pyridyl P2’ꢀresidues resulted in reduced ‘IC shifts’ between the biochemical
50
and MAC deposition assays. We reasoned that the pyridyl nitrogen further increases the size of
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
29,32
the σꢀhole of the bromine atom
leading to an increase in the strength of the halogen bond
with the carbonyl of Trp141 (vide infra).
With the attractive P1’ cyclopropyl proline and the 2ꢀaminoꢀ6ꢀbromoꢀpyridine S2’ꢀbinding motif
in hand, we next turned our attention to the optimization of the P1 indole carboxamide portion.
The contact between the NH of the urea P1ꢀP1’ spacer in 7 and the Ser215 backbone carbonyl in
the crystal structure complex with FD (Figure 2B) was recognized to be subꢀoptimal. In addition,
the dihedral angle between the indole plane and that of the urea group has to be almost
perpendicular in order to allow for the accommodation of the indole scaffold in the S1 pocket.
As such a conformation is likely less favoured energetically in solution, we sought to properly
modify the spacer. Compound 27 harboring a potentially more beneficial amide spacer, however,
was found to be 30ꢀfold less potent than 23 (Table 3). In contrast, the invertedꢀindole analog 28
demonstrated improved potency with only a 10ꢀfold loss compared to 26 in the biochemical
assay. In the case of 28, the slightly more acidic protons of the amide linker may weakly
contribute to a favorable contact with the Ser215 backbone carbonyl. Surprisingly, a dramatic loss
in potency was observed in the MAC deposition assay for 28. In this case, % hPPB would not
entirely explain the larger ‘IC shift’ observed for 28 as compared to compound 26 with slightly
5
0
lower %PPB (Table 3). Most gratifyingly, a 10ꢀfold boost in potency in both the biochemical and
the MAC deposition assays was observed when the indole in 28 was replaced by an indazole
moiety leading to compound 2 which is equipotent to 26 in the thioesterolysis assay and 5ꢀfold
less potent in the MAC deposition assay despite similar % hPPB. Notably, replacing the Nꢀ
ACS Paragon Plus Environment

Page 19 of 77
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
carbamoyl group (2) by an Nꢀacetyl group (29) had only a marginal impact on potency, which
was unexpected in view of the observed Hꢀbonding interactions between both hydrogen atoms of
3
3
the terminal NH deep in the S1 pocket (Figures 2B, 4A), but which could possibly result from
2
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
differences in desolvatation energies.
Table 3. In vitro Potency Data for Compounds 2, 23, 26–30, using a FD Thioesterolytic
Fluorescent Assay and a MAC deposition assay.
FD
IC (ꢀM)
MAC
1
Entry
Cpd
23
27
26
28
2
Y
CF
CF
N
R
P1ꢀmotif
e
a,c
b,c
hPPB (%)
IC (ꢀM)
5
0
50
1
2
3
4
5
7
OCF3
OCF3
Br
A
B
A
C
D
E
0.012
0.37
0.25
nd
d
nd
nd
96.4
86.9
92
0.006
0.066
0.006
0.015
0.009
1.86
N
Br
N
Br
0.05
29
N
Br
0.035
nd
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 77
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
6
30
CF
Br
D
0.010
0.15
nd
aꢀe
For details, see Table 2.
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
We determined the Xꢀray crystal structure of 2 in complex with FD at high 1.33 Å resolution
3
4
(Figure 4). The overall binding mode of the proline scaffold and the indazole moiety to the S1
and S1’ pockets, respectively, is very similar to the binding mode observed for 13. The network
of Hꢀbonds in the S1 pocket and the Hꢀbond between the amide spacer and the Leu backbone
41
carbonyl are conserved for 2. In contrast, the S2’ binding modes are strikingly distinct for the
two compounds with the 6ꢀBrꢀpyridine motif (27) located in the lower part of the pocket, thereby
positioning the bromine atom into the region that is occupied by the CF moiety of the
3
trifluoromethoxyꢀphenyl residue present in 13 (Figure 4B). Notably, the bromine atom forms a
27ꢀ30,33
σꢀhole halogen bond interaction with the backbone carbonyl of Trp .
The C ꢀBr…O
pyr
1
41
angle is 164.6º, thus fulfilling the typical angular preference with a measured bromineꢀoxygen
distance (3.09 Å) significantly lower than the sum of the van der Waals radii (RvdWBrꢀO:
2
7,29
3
.37Å).
The pyridyl nitrogen in 2 further increases the size of the σꢀhole of the bromine
atom, thereby leading to an increase in the strength of the halogen bond with the carbonyl of
29
Trp . Similarly, the presence of an orthoꢀF atom in the 3ꢀBrꢀanilide P2’ moiety, as in 30, is
1
41
2
9,32
expected to strengthen the halogen σꢀhole interaction
with the Trp141 carbonyl (Table 3).
ACS Paragon Plus Environment

Page 21 of 77
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
Figure 4. A. Crystal structures of human FD in complex with 2 (the protein is shown in grey
color and the ligand in cyan color) (PDB code: 5NB7) overlaid with the crystal structure of the
complex with compound 13 (only the ligand is shown, magenta color). The network of Hꢀbonds
(black dashed lines) in the S1 pocket, to the backbone of Gly193 and the Hꢀbond between the
amide spacer and Leu backbone carbonyl are conserved. B. Close up view of the S2’ region,
4
1
the halogen bond (green dashed lines) between the bromine atom in compound 2 and the Trp141
backbone carbonyl is highlighted.
To this point of our work, we had been quite successful in improving the in vitro potency by
>2,000ꢀfold, starting from our initial hit 3 and by iteratively optimizing the interactions of the
inhibitors to each single key hot spot of the FD S1–S1’–S2’ pockets, without any significant
increase in ligand molecular size. This accomplishment is reflected in a remarkable increase in
the BEI and LLE, respectively, for 2 (MW: 482; BEI: 17.1; LLE: 6.9, cLogP: 1.3), one of the
most potent FD inhibitors discovered in this series, as compared to 3 (MW: 465; BEI: 10.4; LLE:
1.4; cLogP: 3.5). The ability to effectively manage MW and cLogP proved to be very important
3
5
as increase in those parameters has been associated with deterioration of the ADMET profile.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 77
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 initial lead compound 7 was shown to have modest aqueous solubility, moderate Cacoꢀ2
cell permeability, lowꢀmicromolar CYP2C9 inhibitory activity (Table 4). The presence of a fused
4,5ꢀcyclopropyl ring (13), while slightly improving solubility, appeared to further increase
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
CYP2C9 inhibition. Introduction of a polar OH or NH substituent to the proline ring
2
ameliorated CYP2C9 affinity (17 and 18), and in case of 18 improved solubility, albeit at the
cost of permeability. Notably, both 14 and 17 demonstrated strong efflux in the Cacoꢀ2
permeability assay.
Table 4. High Throughput Equilibrium Solubility, Permeability and CYP2C9 Inhibition for (S)ꢀ
Prolineꢀbased FD Inhibitors
Cpd
HTꢀeq solubility,
pH 6.8 (mM)
Cacoꢀ2 Papp(AB)
(BA)x 10 (cm/s) /
ratio(BA)/(AB)
CYP2C9
ꢀ6
a
b
/
IC (ꢀM)
5
0
7
0.017
0.077
0.007
0.023
0.28
6.1/ 6.3/ 1
3.4
1
1
3
4
4.5 / 5.7 / 1.3
0.75 / 8.0 / 10
0.77 / 17 / 22
2.1 / 5.1 / 2.5
5.1 / 6.5 / 1.3
10 / 7.0 / 0.7
0.8
2.5
17
>10
>10
1
2
8
3
0.01
0.25
0.15
4.3
24
0.046
0.11
2
2
5
6
11.3 / 14.2 / 1.3
9.6 / 10.3 / 1.1
15.3 / 18.1 / 1.2
0.057
0.17
4.0
2
>20
a
b
Equilibrium highꢀthroughput solubility in aqueous buffer (pH 6.8); Diclofenac was used as
CYP2C9 substrate.
Introduction of an orthoꢀF atom to the P2’ aniline moiety, substituted by either OCF or CF (23,
3
3
24), reduced solubility and led to subꢀmicromolar affinities for CYP2C9. In contrast, switching
ACS Paragon Plus Environment

Page 23 of 77
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
from an aniline to a 2ꢀaminopyridyl P2’ motif led to a superior CYP2C9 profile by increasing the
IC value ~30ꢀfold for 25 and 26, as compared to 24, accompanied by a marginal improvement
50
in solubility. Introduction of the 3ꢀcarboxamide indazole S1 motif (2) resulted in a slight
improvement in aqueous solubility, cell permeability without efflux potential in Cacoꢀ2 cells and
in the P450 inhibition profile (IC >20 and 17 ꢀM for CYP2D6 and 3A4, respectively). The
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
5
0
excellent overall selectivity of inhibitor 2 in a broad panel of proteases, kinases, receptors and
7
ionꢀchannels has been previously reported. Furthermore, 2 did not inhibit the hERG channel
(
IC >30 ꢀM; dofetilide binding and Qpatch assays) and showed no evidence for mutagenic,
5
0
clastogenic or aneugenic potential when assessed in a miniaturized version of the Amesꢀtest and
2
1,22
in the standard in vitro micronucleus test, respectively.
Chemistry
Efficient synthetic routes to access Nꢀurea (compounds 5, 7–26) and Nꢀacylated (2, 27–30)
substituted prolineꢀcarboxamides are outlined in Scheme 1. Briefly, NꢀBoc protected (S)ꢀ, (R)ꢀ
proline, as well as monoꢀ and disubstituted proline analogs, all available as single stereoisomers
from commercial sources, were coupled with substituted anilines or aminoꢀpyridines,
respectively, to afford the corresponding amide intermediates A. Removal of the NꢀBoc
protecting group to intermediates B and subsequent reaction with 1ꢀmethylꢀ1Hꢀindolꢀ3ꢀylꢀ
isocyanate (31) or the 1ꢀcarbamoyl analog 32 furnished the Nꢀurea linker derivatives 5, 7–26.
Similarly, the HBTUꢀmediated coupling reaction of intermediates B with carboxylic acids 33–36
afforded target compounds 2, 27–30 containing an amide linker. For the preparation of the 4(S)ꢀ
aminoꢀproline analog 18 from the commercially available (4S)ꢀ4ꢀNꢀFmocꢀaminoꢀ1ꢀBocꢀLꢀ
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 77
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
proline the deprotection of the Fmoc protecting group was performed using piperidine 20% in
DMF (Scheme 1, step d).
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
Scheme 1. Synthetic Routes for Inhibitors 2, 5, 7 to 30 (cf. Tables 1–3 for Structures)
a
Reagents and conditions: (a) HBTU, DIPEA, CH Cl or DMF, RT or DMF, 70°C (for X = CF
2
2
1
1
and R = OCF ); (b) 1ꢀchloroꢀN,N,2ꢀtrimethylpropenylamine, CH Cl , 0°C (for X = N and R =
3
2
2
2
CF , Br); (c) TFA, CH Cl ; (d) R NCO (31 or 32), Et N, THF; (e) piperidine 20% in DMF, RT
3
2
2
3
5
for the Fmoc deprotection leading to 18; (f) R CO H (33, 34, 35 or 36), HBTU, DIPEA, CH Cl ,
2
2
2
RT.
The 4(R)ꢀ and 4(S)ꢀmethylꢀproline analogs 9 and 10 were prepared from the commercially
available NꢀBocꢀ4ꢀmethyleneꢀproline (37) by HBTUꢀcoupling with 3ꢀOCF ꢀaniline to afford 38,
3
followed by hydrogenation reaction in the presence of Pd on charcoal to produce a mixture of
diastereoisomers, which were separated by reversedꢀphase prepꢀHPLC to afford (2S,4R)ꢀ39 and
ACS Paragon Plus Environment

Page 25 of 77
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
(
2S,4S)ꢀ40, respectively (Scheme 2). The absolute configuration was assigned based on ROESYꢀ
NMR experiments showing a nOeꢀeffect between protonꢀH2 and the 4ꢀmethyl group for the
transꢀconfigured 39, and on the other hand a nOeꢀeffect between protons H2 and H4 for cisꢀ40.
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
Scheme 2. Synthetic Routes for Intermediates 39/40
a
Reagents and conditions: (a) HBTU, DIPEA, CH Cl , RT, 100%; (b) H , Pd/C, MeOH, 39 (7%)
2
2
2
and 40 (53%).
Isocyanate 32 was prepared in four steps by starting from indoleꢀ3ꢀcarboxylic acid (41) by
formation of benzylester 42, Nꢀcarboxamidation using sodium hydride and
chlorosulfonylisocyanate to produce 43. This was followed by saponification to yield the
carboxylic acid 45, and finally by thermal Curtius rearrangement of the acyl azide intermediate
3
6
formed from 45 in the presence of DPPA and Et N in toluene (Scheme 3). It is important to
3
note that the acyl azide intermediate prepared from acid 45 precipitates in the reaction mixture
3
7
and must be purified by filtration before being added portionꢀwise to hot toluene in order to
ensure a smooth Curtius rearrangement. All attempts to perform this reaction in one pot from
acid 45 failed to give the desired isocyanate 32. The Nꢀmethyl indole isocyanate 31 was prepared
accordingly from commercially available 1ꢀmethylindoleꢀ3ꢀcarboxylic acid (44) following a
similar synthetic protocol.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 77
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
Carboxylic acid 34 is commercially available and the synthesis of 35 (Scheme 1) has been
7
reported previously. The (1ꢀcarbamoylꢀ1Hꢀindolꢀ3ꢀyl)ꢀacetic acid (33) was prepared by Nꢀ
carboxamidation of ethyl 2ꢀ(indolꢀ3ꢀyl)acetate (46) with chlorosulfonylisocyanate subsequent to
deprotonation with NaH in THF followed by saponification, and acid 36 was obtained in two
steps from 1ꢀ(1Hꢀindazolꢀ3ꢀyl)ethanꢀ1ꢀone (48) by alkylation with tertꢀbutylbromoacetate and
final esterꢀdeprotection under acidic conditions (Scheme 3).
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
Scheme 3. Synthetic Routes for Intermediates 31-33 and 36
a
Reagents and conditions:(a) BnBr, Cs CO , DMF (83%); (b) NaH, THF, ClSO NCO (43: 29%;
2
3
2
4
7: 75%); (c) H , Pd/C 10%, THF/ DMF, (100%); (d) (i) Et N, DPPA, CH Cl , ꢀ15 °C to 15 °C,
2 3 2 2
t
(
ii) toluene, reflux (31: 89%; 32: 52%); (e) NaOH, MeOH/H O (95%); (f) BrCH CO Bu,
2 2 2
K CO , CH CN reflux; (g) TFA, CH Cl (100% for 2 steps).
2
3
3
2
2
In vivo pharmacology
ACS Paragon Plus Environment

Page 27 of 77
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Since inhibitor 2 demonstrated an excellent in vitro profile, including high potency and
selectivity, we further evaluated the potential of 2 for in vivo use, and for its ability to achieve
38
exposure in the eye, given the evidence for the role of the AP in AMD. Compound 2 displayed
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
7
an excellent oral PK profile in Sprague Dawley rats, and following an oral dose (10 mg/kg) in
Brown Norway rats demonstrated a good distribution and sustained exposure in ocular tissues
including the neural retina and the posterior eye cup (PEC), which comprises the sclera, retinal
pigmented epithelium, and the choroid. Mean exposure levels in plasma, the PEC and the retina
at 6 hours after dosing were 0.36, 0.43, and 0.09 µM, respectively (Figure 5). For context, these
values exceeded significantly the IC for 2 in the in vitro MAC deposition assay (0.05 µM,
5
0
Table 3).
Figure 5. Plasma and ocular pharmacokinetics profiles of inhibitor 2 in Brown Norway rats
after oral administration (10 mg/kg). The compound was formulated as a suspension in 0.5%
HPMC and 0.1% Tween 80 and was administered by oral gavage to male 250 g Brown Norway
rats (n = 2 per time point). Tissues from 2 animals were collected at 0.25, 0.5, 1, 3, 6 and 24
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 77
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
hours, respectively, for analysis (4 eyes/time point; each eye dissected into neural retina and
posterior eye cup). Two plasma samples per time point were collected. Points represent the
mean, and error bars represent the standard deviation.
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
In other studies we have shown that the significant difference in active site conformation of
7
murine vs. human FD precludes the binding of our compounds. To enable in vivo efficacy
assessment, we generated mice that express only the human enzyme (human FD knockꢀin mice)
7
at blood levels that are in the reported range for humans. We developed a pharmacodynamic
model of complement AP activation in these mice using lipopolysaccharide (LPS).
Intraperitoneal administration of LPS induces activation of the AP, as detected by increased
levels of AP breakdown products Ba and iC3b+C3d in the circulation and in the ocular tissues,
7
respectively. We administered 2 by oral gavage at 30 mg/kg to groups of human FD knockꢀin
mice at different time points up to 24 hours prior to the end of the experiment. LPS was
administered to all groups at 7.5 hours prior to the end of the experiment, so that all mice
experienced the same window of complement activation. At study termination, plasma and eye
tissue were collected for measurement of complement breakdown products. Compound 2 caused
a sustained inhibition of Ba generation in the plasma, with suppression to or below baseline
levels for up to 8 hours (Figure 6A). In the eye, complete suppression of complement activation
was seen for at least 12 hours (Figure 6B).
A
ACS Paragon Plus Environment

Page 29 of 77
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 6. Inhibitor 2 tested at 30 mg/kg in the human factor D knockꢀin mouse
pharmacodynamic model: (A) inhibition of Ba generation in plasma and (B) inhibition of iC3b
and C3d generation in ocular tissues. Groups of mice (n = 4, female human FD knockꢀin) were
treated either with 2 or dosing vehicle by oral gavage at 24, 16, 12, 8, 6, and 4 hours,
respectively, prior to the termination of the study. All animals were given intraperitoneal LPS to
activate complement 7.5 hours prior to study termination. Baseline complement levels were
obtained from mice that received oral dosing vehicle and intraperitoneal saline (indicated by PBS
line on graph). The positive control group received oral dosing vehicle and intraperitoneal LPS
ACS Paragon Plus Environment