Discovery of 3-({5-Chloro-1-[3-(methylsulfonyl)propyl]-1 H-indol-2-yl}methyl)-1-(2,2,2-trifluoroethyl)-1,3-dihydro-2 H-imidazo[4,5- c]pyridin-2-one (JNJ-53718678), a Potent and Orally Bioavailable Fusion Inhibitor of Respiratory Syncytial Virus

Article abstract of DOI:10.1021/acs.jmedchem.0c00226

Respiratory syncytial virus (RSV) is a seasonal virus that infects the lungs and airways of 64 million children and adults every year. It is a major cause of acute lower respiratory tract infection and is associated with significant morbidity and mortality. Despite the large medical and economic burden, treatment options for RSV-associated bronchiolitis and pneumonia are limited and mainly consist of supportive care. This publication covers the medicinal chemistry efforts resulting in the identification of JNJ-53718678, an orally bioavailable RSV inhibitor that was shown to be efficacious in a phase 2a challenge study in healthy adult subjects and that is currently being evaluated in hospitalized infants and adults. Cocrystal structures of several new derivatives helped in rationalizing some of the structure-activity relationship (SAR) trends observed.

Full text of DOI:10.1021/acs.jmedchem.0c00226

Subscriber access provided by Uppsala universitetsbibliotek
Article
Discovery of 3 ({5 Chloro-1 [3 (methylsulfonyl)propyl]-1H
indol-2 yl}methyl)-1 (2,2,2 trifluoroethyl)-1,3 dihydro-2H
imidazo[4,5-c]pyridin-2 one (JNJ-53718678), a potent and orally
bioavailable fusion inhibitor of respiratory syncytial virus
Sandrine Vendeville, Abdellah Tahri, Lili Hu, Samuël Demin, Ludwig
Cooymans, Ann Vos, Leen Kwanten, Joke Van den Berg, Mike Battles, Jason
McLellan, Anil Koul, Pierre Raboisson, Dirk Roymans, and Tim H.M. Jonckers
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.0c00226 • Publication Date (Web): 14 May 2020
Downloaded from pubs.acs.org on May 15, 2020
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 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 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.
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 34
Journal of Medicinal Chemistry
1
2
3
4
5
6
Discovery of 3-({5-Chloro-
1-[3-(methylsulfonyl)propyl]-1H-indol-
2-yl}methyl)-1-(2,2,2-trifluoroethyl)-1,3-dihydro-
2H-imidazo[4,5-c]pyridin-2-one (JNJ-53718678), a
potent and orally bioavailable fusion inhibitor of
respiratory syncytial virus
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
Sandrine Vendeville,#† Abdellah Tahri,# Lili Hu,# Samuel Demin,# Ludwig Cooymans,# Ann Vos,#
Leen Kwanten,# Joke Van den Berg,# Michael B. Battles,‡ Jason S. McLellan,$ Anil Koul,# Pierre
Raboisson#†, Dirk Roymans,# Tim H. M. Jonckers#*
#Janssen Pharmaceutica NV, Janssen Pharmaceutical Companies of Johnson & Johnson,
Turnhoutseweg 30, 2340 Beerse, Belgium
‡Department of Biochemistry and Cell Biology, Geisel School of Medicine at Dartmouth,
Hanover, NH 03755, USA
$Department of Molecular Biosciences, The University of Texas at Austin, Austin, TX 78712,
USA
Abstract
Respiratory syncytial virus (RSV) is a seasonal virus that infects the lungs and airways of 64
million children and adults every year. It is a major cause of acute lower respiratory tract infection
and is associated with significant morbidity and mortality. Despite the large medical and economic
burden, treatment options for RSV-associated bronchiolitis and pneumonia are limited and mainly
consist of supportive care. This publication covers the medicinal chemistry efforts resulting in the
identification of JNJ-53718678, an orally bioavailable RSV inhibitor which showed to be
efficacious in a phase 2a challenge study in healthy adult subjects and which is currently being
evaluated in hospitalized infants and adults. Co-crystal structures of several new derivatives helped
rationalizing some of the structure activity relationship (SAR) trends observed.
1
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 34
1
2
3
4
Introduction
5
6
7
8
Respiratory syncytial virus (RSV) is a negative-stranded ribonucleic acid (RNA) virus belonging
to the Pneumoviridae family. In most patients, RSV usually results in upper respiratory tract
infections eliciting “common cold”-like symptoms, which might last up to 2 weeks and are usually
self-limiting. Nevertheless, RSV is considered the most important virus causing acute lower
respiratory tract infection (LRTI) and is a major cause of hospital admissions and death in young
children worldwide.1,2 Infants that are born prematurely or close to the RSV season and/or
suffering from bronchopulmonary dysplasia or congenital heart disease are at the highest risk of
developing severe RSV-related acute LRTI.3 In addition, RSV infection results in substantial
illness and morbidity in the elderly and adults with underlying chronic illnesses, underlying
disorders of cellular immunity, or suppressed immune systems.4 Despite the large medical and
economic burden, treatment options for RSV-associated bronchiolitis and pneumonia are limited:
there is no vaccine and beside prophylactic passive immunization with the humanized monoclonal
antibody palivizumab, or treatment with the non-specific and poorly effective antiviral agent
ribavirin, supportive care remains the major treatment option.5-8 There is therefore an unmet
medical need, prophylactically (pre- and post-exposure) and therapeutically, for both children and
adults.
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
Initial studies in the adult and pediatric population demonstrated that RSV viral load and severity
of disease symptoms are closely correlated.9-11 Since then, two viral targets, the RSV fusion protein
(F) and the RNA-dependent RNA polymerase (L), have been clinically validated, with
respectively, the small-molecule inhibitors GS-5806/presatovir (1)12 and ALS-8176/lumicitabine
(2),13 having demonstrated proof-of-concept in a phase 2 human challenge study upon oral dosing.
2
ACS Paragon Plus Environment

Page 3 of 34
Journal of Medicinal Chemistry
1
2
3
4
5
6
Both compounds decreased viral load and disease symptoms when administered early after
infection.14,15
7
8
9
NH2
O
O
O
HO
N
N
N
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
N
N
O
N
O
N
N
N
N
N
N
N
O
N
H2N
Cl
O
N
N
N
O
F
OH
N
HN
S
Cl
O
O
HO
O
1
2
3
4
GS-5806
ALS-8176
TMC353121
BMS-433771
Figure 1. Structures of compounds 1-4
Moreover, the window between the onset of clinical symptoms (day 1 and day 3, for influenza and
RSV, respectively) and peak viral load (day 2 and day 6, for influenza and RSV, respectively) was
suggested to be considerably longer for RSV than for influenza16, increasing the likelihood that
RSV treatment with direct antivirals can present a viable therapeutic strategy by achieving high
and sustained antiviral responses and thus improve disease outcome. A recent study of presatovir
in hematopoietic stem cell transplant recipients, however, has indicated that antiviral treatment
may more likely result in clinical benefit in patients at high risk for poor RSV-related outcomes
from LRT complications presenting early in the disease course.17
RSV F protein plays a critical role in establishing and propagating RSV infection by catalysing
membrane fusion and virus entry as well as cell-cell fusion resulting in the formation of
syncytia.18,19 Different chemotypes have been shown to inhibit the RSV F protein by binding into
a threefold symmetrical central cavity of the pre-fusion metastable state of F, thereby stabilizing
it and preventing it from refolding to the post-fusion state which is required for membrane fusion.20
In addition to GS-5806, other chemical series have been discovered21-26 but only a few molecules
have been developed until early development and clinical trials due to non-optimal efficacy or PK
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 34
1
2
3
4
5
6
7
8
9
profile, limited safety, or due to strategic decision. For example, TMC353121 (3), an extremely
potent RSV fusion inhibitor, showed a dose-dependent antiviral activity in a non-human primate
model upon continuous infusion27 but side effects at the site of injection prevented this compound
to be developed further. Another example is the benzimidazole-based inhibitor BMS-433771 (4),
the first orally active RSV fusion inhibitor to show efficacy in the mouse and cotton rat models,
as well as primate model.28,29 Although BMS-433771 was advanced to early development stage
based on a promising in vitro and in vivo profile, our internal assessment showed that it suffered
from limitations such as: (i) moderate in vitro potency, (ii) sub-optimal exposure in preclinical
species characterized by poor absorption, high clearance and short half-life, (iii) limited lung
distribution30, and (iv) the potential for time-dependent inhibition (TDI) of cytochrome P450 due
to the presence of the amino cyclopropyl moiety.31
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
In this article, we describe the optimization effort that resulted in the selection of JNJ-53718678,
a sub-nanomolar indole-based fusion inhibitor, having demonstrated evidence of clinical activity
in a Phase 2a challenge study conducted in healthy adult subjects experimentally infected with
RSV, and currently being evaluated in both infants and adults.32 The synthesis, structure-activity
relationships, ADME properties, and PK parameters of the most interesting analogues as well as
their efficacy in the cotton rat model after oral dosing are discussed. The co-crystal structures
obtained with JNJ-53718678 and other analogues are also presented and help rationalize the
obtained SAR.
Results and Discussion
4
ACS Paragon Plus Environment

Page 5 of 34
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
The identification of JNJ-53718678 is the result of an iterative optimization process encompassing
several key learnings. Scheme 1 outlines some of the important steps leading to the selection of
the final candidate. The detailed SAR analysis and optimization strategy is discussed thereunder.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 1: SAR summary leading to JNJ-53718678.
N
N
N
O
O
O
Step 2
Step 1
R2
R3
Xa
Xb
Xc
X
N
N
N
N
N
N
N
Xd
N
N
N
R1
R1
(4)
OH
Step 3
R2
N
CF3
N
N
O
Step 4
O
O
Cl
Cl
Cl
N
N
N
N
N
N
N
N
N
SO2CH3
SO2CH3
SO2CH3
JNJ-53718678
SAR and PK properties leading to the selection of the lead series
Based on the structure of 4, azabenzimidazole, azaindole and indole analogues bearing an
hydroxybutyl side-chain or the more chemically inert isoamyl side-chain, were prepared as initial
exploratory analogues. Adding one extra nitrogen to the bicycle gave 4 different azabenzimidazole
analogues (compounds 5-10). This modification was very well tolerated when at the 4-position of
the bicycle, giving compounds 5 and 6 with a similar activity compared to 4. In the 5-
azabenzimidazole system, a potency drop of about 1 log10 for both the polar hydroxybutyl and
isoamyl side-chain derivative was observed (see compounds 7 and 8). The two other isosteres, the
5
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 34
1
2
3
4
6- and 7-azabenzimidazoles 9 and 10, showed an even stronger loss in potency of about 1.1 and
5
6
2.6 log10 reduction in pEC50 value respectively.
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
Table 1 shows the in vitro potencies of a first set of isosteres. Antiviral activity was measured in
HeLa cells infected with rgRSV224, using a fluorescence read-out.33 In parallel, cytotoxicity was
measured in HeLa cells after a three-day incubation time and was used to calculate selectivity
indices (SI).
Table 1. Antiviral activities of new bicyclic series in RSV-infected HeLa cells.
N
O
Xa
N
X
N
Xb
Xc
N
Xd
R1
a
N°
Bicycle
Xb
CH
CH
CH
N
R1
pEC50
SIb
Series
X
N
N
N
N
N
N
N
CH
CH
CH
CH
CH
CH
Xa
CH
N
Xc
CH
CH
CH
CH
CH
N
CH
CH
CH
N
Xd
>20.103
>10.103
>12.103
>18.103
>3.103
>987
4
5
6
7
8
CH
CH
CH
CH
CH
CH
N
CH
CH
CH
N
(CH2)4OH
(CH2)4OH
(CH2)2CH(CH3)2
(CH2)4OH
(CH2)2CH(CH3)2
(CH2)2CH(CH3)2
(CH2)2CH(CH3)2
(CH2)4OH
(CH2)4OH
(CH2)4OH
(CH2)2CH(CH3)2
(CH2)3SO2Me
(CH2)2CH(CH3)2
8.1
8.0
8.1
7.3
7.5
7.0
5.9
6.8
8.8
6.9
<4
benzimidazole
4-azabenzimidazole
4-azabenzimidazole
5-azabenzimidazole
5-azabenzimidazole
6-azabenzimidazole
7-azabenzimidazole
4-azaindole
N
CH
CH
CH
CH
N
CH
CH
CH
CH
CH
N
9
CH
CH
CH
N
CH
CH
CH
CH
10
11
12
13
14
15
16
>80
>613
>57.103
>76
5-azaindole
6-azaindole
7-azaindole
CH
CH
CH
nd
>10.103
181
CH
CH
8.0
6.4
indole
indole
aRSV assay in HeLa cells. Data generated from two or more determinations. bselectivity index.
In the azaindole subseries, the 6- and 7-aza indoles 13 and 14 also led to a poor antiviral activity
(respectively, pEC50 of 6.9 and < 4.0). With the 4-aza indole derivative 11, a tenfold drop in
6
ACS Paragon Plus Environment

Page 7 of 34
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
potency was observed, whereas the 5-azaindole derivative 12 unexpectedly led to a compound 4
times more potent than 4, with a single digit nanomolar potency (pEC50 = 8.8). Finally, the more
lipophilic indole core was also evaluated. Combination of the indole with the isoamyl side-chain
led to a poorly active compound (compound 16, pEC50 of 6.4). The combination of the indole core
with the isoamyl side-chain resulted in a highly lipophilic (cLogP of 5.1) and protein bound (0.6%
free in human serum) compound which might hamper proper engagement with the target. In order
to counterbalance the high lipophilicity, compound 15 bearing the more polar methylsulfone
propyl side-chain was prepared. This combination resulted in an inhibitor with a greatly improved
potency (pEC50 of 8.0), similar to 4. Of note, none of these inhibitors reached a 50% maximal
cytotoxic concentration (CC50) value, the highest concentration tested being 100 µM, and all
compounds exhibited high SI, with inhibitor 16 being a slight exception. This first initial
exploration, in which nine different bicycles were evaluated, resulted in three of them (the 4-
azabenzimidazole, the 5-azaindole and the indole) being retained for further optimization as they
demonstrated potencies equal to or below 10 nM.
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 next step, we investigated substitution on the indole ring itself (see compounds 17-25 in table
2). Introducing a halogen at the 5-position had a positive impact on potency. The 5-fluoro
derivatives 17 and 18 had a pEC50 value of 7.8 and 8.4, respectively. A much more pronounced
effect was seen with the 5-chloro and 5-bromo derivatives 19-22. Compared to their fluoro
analogues, a tenfold improvement in potency was observed (e.g. compare 18 with 20 and 22),
leading to sub-nanomolar potent analogues. In contrast, when a chloro substituent was introduced
at the 6-position of the indole, a poorly active compound was obtained (compound 23, pEC50 =
6.9). The latter result is in agreement with what was seen in another series, where a 7-chloro-
imidazo[1,2a]pyridine proved to be more potent than the non-substituted analogue, which itself
7
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 34
1
2
3
4
5
6
7
8
9
was more potent than the 6-chloro-imidazo[1,2a]pyridine derivative.24,34 The electron-
withdrawing CF3 group was also introduced at position 5 of the indole to give compound 24.
Although a single digit nanomolar inhibitor was obtained, the potency dropped compared to the 5-
chloro derivative. Finally, introducing the electron-donating methoxy group was found to be
detrimental for activity (compound 25, pEC50 = 7.4).
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
Table 2. Effect of indole substitutions on antiviral potency
N
O
R2
R3
N
N
N
R1
a
N°
17
18
19
20
21
22
23
24
25
R1
(CH2)4OH
(CH2)3SO2Me
(CH2)4OH
(CH2)3SO2Me
(CH2)4OH
(CH2)3SO2Me
(CH2)4OH
(CH2)3SO2Me
(CH2)4OH
R2
F
F
Cl
Cl
Br
Br
H
CF3
OMe
R3
H
H
H
H
H
H
Cl
H
H
pEC50
7.8
8.4
8.7
9.7
8.9
9.4
6.9
8.1
7.4
a RSV assay in HeLa cells. All compounds demonstrated CC50 values > 50 μM.
Having identified the 5-chloro substituent as beneficial, we next studied the effect of the lateral
side-chain on potency and metabolic stability. As already reported,35 this side-chain is positioned
towards the solvent and both apolar moieties as well as alkyl chains with a terminal polar group
are tolerated.27 We introduced side-chains of varying degrees of lipophilicity, in order to modulate
both potency and ADME properties by variation of the overall physicochemical properties of the
obtained compounds. Table 3 summarizes the antiviral potency, metabolic stability, and cLogP of
these new derivatives 26-33.
Table 3. Appending side-chain optimization in the 5-chloroindole sub-series
8
ACS Paragon Plus Environment

Page 9 of 34
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
N
O
Cl
N
N
8
9
N
R1
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
N°
R1
pEC50
HLMb
% metab.
49%
18%
56%
17%
37%
23%
-
clogP
19
20
26
27
28
29
30
31
32
33
(CH2)4OH
(CH2)3SO2Me
(CH2)3Cme2OH
(CH2)3CN
8.7
9.7
8.8
9.7
9.1
9.6
6.7
7.3
8.6
8.2
3.5
2.9
4.2
3.8
4.8
5
3.9
3.1
2.4
4.3
(CH2)4F
(CH2)3CF3
(CH2)3COOH
(CH2)3CONH2
(CH2)3C=NHNHOH
(CH2)3N-morpholine
-
37%
99%
aRSV assay in HeLa cells. Data generated from two or more determinations. All compounds demonstrated CC50 values > 50 μM.
bHuman liver microsome turnover, expressed as the percentage of compound metabolized after 15 minutes incubation.
The carboxylic acid moiety was clearly not tolerated on the side-chain and led to a dramatic loss
of potency compared to the hydroxyl derivative (compare 30 and 19). The simple amide derivative
31 also suffered from a very poor potency, whereas the basic hydroxyamidine analogue 32 led to
a potency similar to 19. The morpholine derivative 33 led to a pEC50 of 8.2 but was found to be
highly metabolically unstable, with 99% of the compound metabolized after 15 minutes in human
liver microsomes (HLM). In an attempt to tackle the metabolic lability of the hydroxybutyl
derivative 19, we synthesized compound 26, in which two methyl groups are positioned alpha to
the hydroxyl moiety. This steric hindrance was well tolerated in terms of potency but did not
improve metabolic stability in HLM. Replacing the hydroxyl group by a fluoro atom led to
compound 28 with an improved potency below 1 nM, but a sub-optimal metabolic stability (37%
metabolized after 15 min in HLM). Finally compounds 20, 27 and 29, possessing either a terminal
methylsulfone, cyano, or trifluoromethyl group at the end of the propyl side-chain, all led to
extremely potent inhibitors, with potencies below 250 pM. These three compounds also had
9
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 34
1
2
3
4
5
6
7
8
9
acceptable metabolic stability in HLM. Based on its superior ADME profile, especially the
intrinsic permeability and solubility, the cyano derivative 27 was chosen as a representative
molecule to evaluate the PK properties of the indole series in vivo in rat (see in vivo PK section
below).
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 transfer of SAR of the indole series to the 4-azabenzimidazole and 5-azaindole series led to
the synthesis of derivatives 34-44 (see table 4). The following observations were made. First, in
the 4-azabenzimidazole series, a decreased potency was observed compared to the indole-based
series, even in the presence of the potency-enhancing 5-Cl substitution. pEC50 values ranged from
8.0 for the cyano derivative 37 to 9.0 for the trifluorobutyl and isoamyl analogues 39 and 40,
respectively. Second, metabolic stability was improved and all derivatives of this 4-
azabenzimidazole series, except 36, showed minor turnover in HLM. Finally, a substantial
permeability was measured for the most lipophilic derivatives 38-40.
Table 4. Appending side-chain optimization in the 4-azabenzimidazole and 5-azaindole series
N
N
O
O
Cl
N
N
N
N
N
R1
N
N
N
N
R1
Series A: 4-azabenzimidazole Series B: 5-azaindole
a
N°
R1
Series
pEC50
HLMb
(% met.)
Pabc
clogP
(× 10−6
cm/s)
2
34
35
36
37
38
39
(CH2)4OH
(CH2)3SO2Me
(CH2)3CMe2OH
(CH2)3CN
A
A
A
A
A
A
8.3
8.4
8.8
8.0
8.7
9.0
0%
0%
41%
0%
0%
0%
1.4
0.8
2.1
1.6
2.7
2.9
0.7
ND
4
18
21
(CH2)4F
(CH2)3CF3
10
ACS Paragon Plus Environment

Page 11 of 34
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
40
12
41
42
43
44
(CH2)2CH(CH3)2
(CH2)4OH
(CH2)3SO2Me
(CH2)4F
(CH2)3CF3
(CH2)2CH(CH3)2
A
B
B
B
B
B
9.0
8.8
9.0
9.3
9.6
9.6
7%
ND
32%
15%
21%
44%
20
3.7
<1
25
24
24
3.8
1.5
0.9
2.7
3.0
3.9
a RSV assay in HeLa cells. Data generated from two or more determinations. All compounds demonstrated CC50 values > 50 μM.
b Human liver microsome turnover, expressed as the percentage of compound metabolized after 15 minutes incubation.
cApparent permeability measured in LLC-PK1-MDR1 cell line, in presence of the Pgp inhibitor GF120918. ND: not determined.
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
In the 5-azaindole series, combination with the more lipophilic side chains proved to be beneficial
both for permeability and potency, with compounds 42-44 achieving sub-nanomolar EC50 values.
The fluoro and trifluoro derivatives 42 and 43 also demonstrated acceptable metabolic stability in
HLM. On the contrary, the more polar compounds 12 and 41 exhibited a slightly decreased potency
and a low permeability.
In vivo PK studies
For each series, a panel of analogues chosen based on their antiviral potency, acceptable in vitro
human metabolic stability and solubility, as well as permeability, was evaluated in vivo in fed male
Sprague-Dawley (SD) rats after IV and oral administration at 2 and 10 mg/kg respectively.
Concentrations in the lungs, the target organ, were also measured. Table 5 summarizes the PK
parameters obtained for the best representative of each series.
Table 5. PK parameters of compounds 27, 39 and 43 after single-dose administration in
Sprague-Dawley (SD) rat
IVa
POb
Compound
Cl
(L/h/kg)
2.2
Vdss
(L/kg)
1.5
0.82
3.3
Cmax
(ng/mL)
1230
1690
532
AUC
F(%)
Lung/Pc
(ng.h/mL)
27
39
43
3300
3370
545
73
46
28
1.6
0.4
11
1.4
4.5
aDose and vehicle: 2 mg/kg in 20% HP-B-CD. bDose and vehicle: 10 mg/kg in 20% HP-B-CD. clung/plasma ratio calculated
based on AUC0-last
.
Table 5 illustrates the direct impact of the type of heterocycle on PK parameters. Indeed indole 27,
azabenzimidazole 39 and azaindole 43 have a similar metabolic stability in RLM (around 80% of
11
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 34
1
2
3
4
5
6
7
8
9
compound metabolized after 15 min) and a similar permeability. A major difference resides in
their plasma protein binding properties, however, reflecting their differential lipophilicity, with the
indole 27 being rather highly protein bound in presence of human plasma (3.6% free fraction),
whereas the azabenzimidazole 39 and azaindole 43 are much less protein bound (30% and 47.2%
free fraction, respectively). Indole 27 demonstrated a promising PK profile, characterized by a
moderate clearance after IV dosing (lower than predicted by rat liver microsome turnover), a high
Cmax and AUC after oral dosing leading to an absolute bioavailability of 73%, and a favorable
distribution to the target organ (ratio lung/plasma of 1.6). Azabenzimidazole 39 also had an
excellent plasma exposure (bioavailability (F) = 46%) associated with a moderate clearance after
IV dosing. However, a low distribution to the lung was measured (ratio lung/plasma of 0.4). This
latter property was found to be common to all compounds tested in this sub-series, which could
not be explained only by a lower lipophilicity or the nature of the side-chain. Azaindole 43 had a
moderate exposure (absolute bioavailability (Fabs) of 28%) characterized by a high clearance
(more than 100% of rat liver blood flow). However, the compound distributed very well to the
lung, probably due to the presence of the basic center on the azaindole. Indeed, it is well known
that basic lipophilic compounds can be sequestered in the acidic lysosomes of the lungs, therefore
considerably increasing lung concentration.36,37 However, this property may not necessarily be a
desirable attribute, since it can lead to an overly-long half-life due to tissue retention and
lipodystrophy.25,38,39
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
Based on its overall better in vivo profile, the indole compound 27 was selected as lead molecule
for further optimization of the substitution on the azabenzimidazolone bicycle.
Lead optimization of the indole series
12
ACS Paragon Plus Environment

Page 13 of 34
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Keeping the 5-chloro indole ring fixed, the replacement of the cyclopropyl group in combination
with either the cyanopropyl or the methylsulfone propyl side-chains was explored. A major goal
was to avoid the potential for time-dependent inhibition (TDI) of cytochrome P450 3A4 (CYP3A4)
that could be conferred by the presence of the cyclopropyl moiety.31,40 Indeed, SAR analysis of
compounds from the different sub-series having this fragment in common, showed that TDI
liability was attributable to this moiety. As a typical example, compound 20 demonstrated potential
for TDI of CYP3A4-mediated metabolism of testosterone with an 18-fold difference of IC50
measured in the shift assay.41
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
We hypothesized that reactive metabolites could be formed via the removal of the single H-atom
(C1-H) of the cyclopropyl moiety and/or via an oxidative metabolite on the pyridine ring, and
therefore evaluated three main strategies: i) close (cyclo) alkyl-based analogues lacking this
removable C1-H-atom, ii) a diversity exercise involving the replacement of the cyclopropyl by
aryls and heterocycles, and iii) introduction of small substituents on the pyridine ring in order to
modulate steric hindrance and electronic properties.
Analogues of the lead molecule 20 having an alkyl moiety in place of the cyclopropyl group are
depicted in Table 6.
Table 6. Variation of the cyclopropyl moiety
R2
N
O
Cl
N
N
N
S
O
O
13
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 34
1
2
3
4
5
6
a
N°
R2
pEC50
HLMb
(% met.)
18
20
9.7
7
8
9
45
46
-H
8.1
8.8
9
97
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
47
9.3
29
48
49
8.6
9.4
28
98
50
51
8.6
7.5
52
14
O
NH
52
53
54
8.8
9.3
8.3
15
1
SO2Me
N
CF3
2
CF3
55
-CH2CN
8.7
66
aRSV assay in HeLa cells. All compounds demonstrated CC50 values > 50 μM. bHuman liver microsome turnover, expressed as
the percentage of compound metabolized after 15 minutes incubation.
Analogue 45, lacking any substitution on the azabenzimidazolone moiety, illustrates the
importance of this group to retain potency, since a loss in potency of 1.6 log10 was observed
compared to the cyclopropyl lead compound 20. Notably, this compound didn’t exhibit any TDI
of CYP3A4. Next, analogues 46-49 were synthesized with the goal to obtain TDI-free inhibitors
while retaining good potency and trying to improve the metabolic stability of compound 20.
Replacing a cyclopropyl by its isopropyl analogue is a known strategy to decrease the risk of TDI.29
In our case, the isopropyl derivative 46 led to almost 1 log10 drop in potency, while eliciting no
potential for TDI. However, a strong decrease in metabolic stability was observed. Similar findings
14
ACS Paragon Plus Environment

Page 15 of 34
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
were made with the tert-butyl derivative 48 and oxetane analogue 50. Replacing the C1-H
hydrogen by a methyl (compound 47) or introducing a methylene spacer between the bicycle and
cyclopropyl moieties (compound 49) was well tolerated in terms of potency but led to compounds
characterized by higher turnovers in HLM, although devoid of TDI. Azetidine analogue 51 was
not tolerated in terms of potency (pEC50 = 7.5), but this could be partially reversed by substituting
the azetidine nitrogen like in the sulfonamide analogue 52. Replacing one of the methyl groups of
isopropyl derivative 46 with a CF3 led to a very metabolically stable compound, although to some
detriment of potency (compound 54, pEC50 = 8.3). Finally, the simple trifluoroethyl derivative 53
was found to combine both excellent potency (pEC50 = 9.3) and metabolic stability (only 1%
metabolized after 15 min in HLM), whereas the cyano analogue 55 displayed a loss of potency
and metabolic stability. Of note, analogue 53 did not exhibit any TDI of CYP3A4. When combined
with other appending chains, the resulting trifluoroethyl analogues displayed a potency similar to
53 but higher turnovers in HLM (not shown). As a result, inhibitor 53 was selected as the most
optimal compound from this exercise and was further characterized (see in vivo efficacy section).
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
In a second strategy, we studied the replacement of the cyclopropyl moiety by substituted aryls
and heteroaryls (see compounds 56-71 in table 7). Anticipating that aqueous solubility could
become insufficient with this more lipophilic sub-series, the cyanopropyl moiety was used instead
of the methylsulfone propyl appending chain.
Table 7. SAR in the top aryl and heteroaryl sub-series
R2
N
O
Cl
N
N
N
N
15
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 34
1
2
3
4
5
6
7
8
9
a
N°
R2
pEC50
HLMb
(% met.)
56
57
58
59
60
p-Me-Ph
p-COOH-Ph
p-CONH-cPr-Ph
p-SO2Me-Ph
p-OCF3-Ph
8.9
9.3
9.3
9.7
8.9
36
9
24
9
0
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
p-OMe-Ph
2-F-4-OMe-Ph
3-Me-4-OMe-Ph
3,4-di-OMe-Ph
2-Me-4-OMe-Ph
N
9.5
9.2
9.6
10.0
9.0
20
17
33
19
36
60
8.0
N
67
8.2
34
N
N
N
68
69
8.8
8.8
45
49
O
N
N
N
70
71
8.6
9.5
23
27
N
F
aRSV assay in HeLa cells. All compounds demonstrated CC50 values > 50 μM. bHuman liver microsome turnover, expressed as
the percentage of compound metabolized after 15 minutes incubation.
Compared to the initial cyclopropyl derivative 27, inhibitors with similar or even superior potency
together with similar metabolic stability could be obtained (see compounds 59, 61, 63 and 64).
The nature of the phenyl substitution was found to have a substantial impact on potency; the least-
tolerated substituents were the methyl and trifluoromethoxy on the 4 position of the phenyl (56
and 60, respectively; both pEC50 = 8.9). The acid analogue 57 and the amide derivative 58
exhibited a slightly lower potency compared to the cyclopropyl analogue 27 (pEC50 of 9.3 versus
9.7 for 57 and 58 versus 27, respectively). Compound 59, whose phenyl ring is substituted with a
16
ACS Paragon Plus Environment

Page 17 of 34
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
methylsulfonyl moiety in para position, was as potent as 27, with a metabolic stability in HLM,
however suffering from a very low aqueous solubility. The potency of the 4-OMe derivative 61
was found equal to 27, and could further be modulated by the introduction of a second substituent,
preferably on the position 3 of the phenyl, with the 3,4-dimethoxyphenyl derivative 64 being the
most potent analogue of this sub-series (pEC50 = 10), while the presence of a 2-fluoro or 2-methyl
decreased potency (see 62 and 65, respectively).
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
Encouraged by the potency of the phenyl derivatives, and in order to maximize solubility, attention
was then turned to the synthesis of the heterocyclic derivatives 66-71, with an emphasis on
nitrogen-containing heterocycles. In this sub-series, and contrary to the phenyl sub-series
described above, a small substitution at the position 2 of the heterocycle conferred a much-
improved potency (compare 70 and 71). However, the most potent compound of this sub-series 71
exhibited a sub-optimal metabolic stability. Disappointingly, all the other heterocycles tested 66-
70 demonstrated much lower potency compared to the cyclopropyl reference compound 27, as
well as a higher turnover in HLM.
Finally, the third strategy, based on the introduction of small substituents on the pyridine ring in
order to modulate steric hindrance and electronic properties, was not successful, resulting in much
less potent inhibitors (ie a chloro or methyl in meta to the pyridine nitrogen; data not shown).
Crystallography
The biological profile, including selectivity over other RNA viruses as well as the co-crystal
structure of compound 53 bound to prefusion RSV F has been previously reported.35 It was shown
that the inhibitor binds within the central cavity along the three-fold trimeric axis occupying two
of the three equivalent lobes of the binding pocket (Figure 2). The binding mode involves (i) pi-
17
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 34
1
2
3
4
5
6
7
8
9
pi stacking between Phe488 and Phe140 and the 2 aromatic bicycles of the inhibitor (ii) C-H
(indole) / pi and C-H (azabenzimidazolone) / pi interactions with Phe140 (iii) halogen bond with
backbone carbonyl oxygen of Thr397 and (iv) water-mediated hydrogen bond between one oxygen
of the sulfone moiety and side chain of Arg339. In addition, the trifluoroethyl group optimally fits
into a hydrophobic clamp formed by Phe140, Phe488 and Phe137.
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 2. Co-crystal structure of JNJ-53718678 bound to prefusion RSV F (PDB ID 5KWW). (a) View of
the central cavity looking down toward the viral membrane. (b) Side view of the central cavity view parallel
to the viral membrane. The three RSV F protomers are coloured green, pink, and tan. Water molecules are
represented as red spheres. JNJ-53718678 (compound 53) is shown in ball-and-stick representation.
New co-crystal structures of compounds 39, 41 and 64 bound to prefusion RSV F revealed an
essentially similar binding mode, but also new features. As previously described, the chloro atom
of compounds 39 and 64 is engaged in a halogen bond interaction with the backbone carbonyl of
Thr397 (Figures 3a and 3c). This interaction also explains the increase in potency from a fluoro
to a chloro or bromo (10-fold potency increase).43 On the other hand, the N-atom located at the 5
position of azaindole derivative 41 is too remote from the backbone carbonyl of Thr397 (distance
is 4.7 Å) to form a hydrogen bond after protonation. This nitrogen has a measured pKa value of
18
ACS Paragon Plus Environment

Page 19 of 34
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
7.4 indicating that at physiological pH about 50% of the compound is in its protonated form. This
may enhance compound binding by engaging in an electrostatic interaction with Asp489 (Figure
3b). The small alkyl chains with CF3 (compound 39), sulphone (compound 41) or cyano
(compound 64) substitutions are pointing towards solvated regions and are probably replacing
water molecules with unfavorable binding energies because of the hydrophobic nature of the
binding pocket. Furthermore, the top cyclopropyl moieties of compounds 39 and 41 display a very
similar binding mode compared to the trifluoroethyl group of compound 53. For compound 64,
the bigger 3,4-dimethoxyphenyl substituent results in the strengthening of the interaction with the
aromatic side chains of Phe137 and Phe140 and explains the superior antiviral potency of
compound 64. It also explains why a large diversity of moieties is tolerated in this part of the
molecule.
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 3. Co-crystal structure of compounds 39 (PDB ID 6VKC), 41 (PDB ID 6VKD) and 64 (PDB ID
6VKE) bound to prefusion RSV F. Side view of the central cavity, view parallel to the viral membrane.
The three RSV F protomers are coloured green, pink, and tan. Compounds are shown in ball-and-stick
representation.
In vivo PK and efficacy studies leading to the selection of JNJ-53718678
19
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 34
1
2
3
4
5
6
7
8
9
In order to differentiate the best compounds selected after the different optimization phases, the
PK profile of compounds 53, 61 and 64 was assessed after IV and oral dosing in fed male SD rat
and dog. The resulting PK parameters are summarized in table 8.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 8. PK parameters after single administration of compounds 53, 61 and 64 in rat and dog
Ratb
Dogc
Compound R/D/HLM a (%
met)
Cld
(L/h/kg)
3.9
0.97
1.9
Lung/Pe
F(%)f
Cld
F(%)f
(L/h/kg)
53
61
64
62/28/1
40/96/20
74/56/19
3.5
1.3
2.5
42
86
32
0.51
1.49
0.86
89
9
26
aRat/dog/human liver microsome turnover, expressed as the percentage of compound metabolized after 15 minutes incubation.
bDose and vehicle IV: 2.5 mg/kg in PEG400/saline 7/3, PO: 10 mg/kg in PEG400-HCl. cDose and vehicle IV: 1 mg/kg in HP-B-
d
e
f
CD 20%, PO: 5 mg/kg in PEG400. plasma clearance after IV dosing. lung/plasma ratio calculated based on AUC0-last. Oral
bioavailability.
In dog, plasma clearance after IV dosing correlated well with the measured in vitro metabolic
stability in dog liver microsomes. Compound 53 was the most stable and exhibited a moderate
plasma clearance and an excellent bioavailability of 89% after oral dosing. The 2 substituted aryl
analogues 61 and 64 displayed a higher clearance, 90% and 50% of the dog liver blood flow,
respectively, in agreement with their higher turnover in dog liver microsomes. Their bioavailability
was consequently also reduced 9% and 26%, respectively. In rat, in vivo clearance was found lower
than expected based on rat microsomal stability for compounds 61 and 64. The lowest clearance
was obtained with analogue 61, which displayed an excellent bioavailability of 86%. For analogue
53, despite a rather high clearance, a moderate bioavailability of 42% was obtained. In addition,
this compound also exhibited the highest distribution to the lung with a lung/plasma ratio of 3.5.
Those 3 compounds were further evaluated in an efficacy study in cotton-rats. In this well-
established semi-permissive model of RSV infection, animals were infected with an intranasal
dose of 105 plaque forming units (PFU) Long strain RSV, establishing infection of their lower
respiratory tract.42 The inhibitor was administered prophylactically 1h prior to the infection at
20
ACS Paragon Plus Environment

Page 21 of 34
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
different single oral doses (1, 4, 10, 40, or 100 mg/kg). Four days after viral challenge (at peak
viral replication), the RSV viral titer was determined in the broncho alveolar lavage fluid (BALF)
and in the lavaged-lung tissue of the animals by plaque assay. On the basis of these results, a
pharmacokinetic/pharmacodynamic (PK/PD) model according to maximum effect (Emax)-fitting
was built to estimate the in vivo EC50, based on compound exposure measured in plasma at the
time of challenge. Results of this PK/PD model are presented in table 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
Table 9. In vivo efficacy of compounds 53, 61 and 64 in the cotton-rat model
EC50 (ng/mL)
EC50 (ng/mL)
BALF
Lavaged-lung
53
61
64
75
499
180
81
121
157
A dose-dependent reduction of the viral titers in BALF and lavaged-lung was observed for the 3
compounds, as well as a decrease of viral RNA production in the lungs. Despite their higher in
vitro potency, the 2 aryl derivatives 61 and 64 were found to be less efficacious in vivo compared
to compound 53. This latter inhibitor exhibited a more pronounced effect in vivo with comparable
EC50s for BALF and lavaged-lung (75 ng/mL and 81 ng/mL respectively). We speculate that this
higher efficacy may result from a combination of more favorable parameters such as ADME
properties (especially lung distribution) and free fraction (6%, 0.5% and 0.8% free fraction in
human plasma measured for 53, 61 and 64, respectively), although more experiments would be
required to extrapolate on the parameters conferring superior efficacy.
The in vivo efficacy of 53 was further evaluated in the neonatal lamb model which is a fully-
replicative RSV infection model displaying very similar disease symptoms as observed in
infants.44-45 Our experiments demonstrated that once-daily oral administration of JNJ-53718678
21
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 34
1
2
3
4
5
6
completely eliminated an established lower respiratory tract infection in these animals, along with
a suppression of virus-induced lung inflammation.35
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
Based on its superior in vivo efficacy, its exposure in preclinical species supporting a low predicted
efficacious QD dose in patients, as well as its favorable safety profile and low potential for drug-
drug interaction, compound 53 (JNJ-53718678) was selected as the clinical candidate. Currently
it is being evaluated in Phase II studies in both infants and adults. Recently, data from a randomized
phase 1b study evaluating the pharmacokinetics, safety and antiviral effects of 53 in hospitalized
infants was reported46.
Chemistry
Compounds 5-71 were synthetized following the general pathway shown in scheme 2. In general,
a Mitsunobu reaction between the different bicycles I and azabenzimidazolones II afforded
intermediates IIIa-e and final compounds 6, 8-10, 13-16, 18, 20, 22, 24, 27-29, 33, 35, 37-44, 46-
50, 53, 54, 56 and 59-71.47 While straightforward and generally high yielding, this coupling
method also yielded limited (typically less than 10%) amounts of the undesired O-alkylated side-
product IV.48 When the appending side-chain also contained an alcohol moiety, its protection as a
benzylether like in IIIa or silyl ether like in IIIb was necessary to perform the Mitsunobu reaction.
Protecting group removal then afforded final compounds 5, 7, 11, 12, 17, 19, 21, 23, 25, 26, 34,
36. Compounds 30-32 were obtained from the cyano precursor 27 using the conditions depicted in
scheme 2. The non-substituted azabenzimidazolone derivative 45 was obtained after acidic
treatment of the isopropylidene precursor IIIc. Further reaction of 45 with chloroacetonitrile in
presence of cesium carbonate in DMF gave compound 55. Compound 51 was obtained after TFA
treatment of the Boc-azetidin-3-yl precursor IIId and was further reacted with mesylchloride to
22
ACS Paragon Plus Environment

Page 23 of 34
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
give compound 52. Carboxylic acid 57 was obtained from the methyl ester precursor IIIe after
treatment with LiOH. Reaction of 57 with cyclopropylamine and diethylcyanophosphonate as
coupling agent gave amide 58. Synthesis of intermediates I and II is described in the supporting
information.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 2. General pathway for the synthesis of final compounds 5-71.
R3
R3
N
O
R3
R2
N
X
Z
R2
O
N
R2
X
N
R1
i
X
N
N
N
+
+
O
N
N
N
N
N
N
N
N
H
R1
R1
IIIa-e and
I
IV
II
6, 8-10, 13-16, 18, 20, 22, 24, 28, 29
33, 35, 37-44, 46-50, 53, 54, 56, 59-71
ii
30-32
27
iii
iv
1
5, 7
IIIa
(R = -(CH2)4-OBn)
11, 12, 17, 19, 21
23, 25, 26, 34, 36
(R1 = -alkyl-OTBDMS)
(R2 = isopropylidene)
(R2 = Boc-azetidin-3-yl)
(R2 = p-COOMe-Phe)
IIIb
IIIc
IIId
IIIe
v
vi
45
51
55
52
vii
viii
x
ix
57
58
Reagents and conditions. (i) I, II, DIAD, Ph3P, DMF or THF, 12-97 % yield. (ii) a) for 30: 27, KOH, iPrOH, H20,
120°C, 1h, 96% yield, b) for 31: 27, NH2OH, Et3N, EtOH, 80°C, 16h, 15% yield, c) for 32: 27, NH2OH, K2CO3,
EtOH, H2O, 80°C. (iii) IIIa, H2, Pd/C, CH3COOK, MeOH, 50°C, 32-36% yield. (iv) IIIb, ammonium fluoride,
MeOH, 60°C, 32-95% yield. (v) IIIc, HCl 6M in iPrOH, H20, 70% yield. (vi) 45, ClCH2CN, Cs2CO3, DMF, 62%
yield. (vii) IIId, TFA, DCM, 77% yield. (viii) 51, ClSO2Me, Et3N, DCM, 32% yield. (ix) IIIe, LiOH, H2O/THF 1/3,
48% yield. (x) 57, cyclopropylamine, Et2CNPO3, Et3N, DMF, 68% yield.
Conclusion
A medicinal chemistry optimization program based on a formerly-disclosed sub-optimal
benzimidazole series led to novel RSV fusion inhibitors exhibiting greatly improved potencies, as
well as superior PK properties after oral dosing. The indole sub-series was chosen as lead series
for further optimization essentially based on its superior in vivo PK properties and higher
23
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 34
1
2
3
4
5
6
7
8
9
distribution to the lung, the target organ. In this context, the presence of a more polar appending
chain from the indole moiety proved to be crucial to counterbalance the higher intrinsic
lipophilicity of the indole core. Replacement of the cyclopropyl group on the azabenzimidazolone
successfully led to analogues devoid of TDI of CYP3A4. Several derivatives exhibited desired
antiviral potency and PK profiles in rats and dogs and demonstrated good in vivo efficacy in the
cotton rat model. New X-ray co-crystal structures obtained with compounds 39, 41 and 64 bound
to prefusion RSV F could rationalize the observed SAR. This lead optimization effort culminated
in the discovery of compound 53 (JNJ-53718678), a sub-nanomolar indole-based inhibitor, with a
preliminary PK profile supporting a low predicted efficacious QD oral dose in patients, as well as
a favorable safety profile and low potential for drug-drug interaction.
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
Experimental section
Animal care and experiments were performed according to the applicable European Communities
Council Directive of November 24, 1986 (86/609/EEC).
Chemistry - Analytical Methods.
Reagents and solvents were obtained from commercial sources and used as received unless stated
otherwise. All final compounds were characterized by 1H NMR and HRMS. 1H nuclear magnetic
resonance spectra were recorded on a Bruker spectrometer operating at 300 MHz, 360 MHz, and
DPX-400 MHz. The purity of all final compounds was ≥95% by LCMS. Chemical shifts are
reported in parts per million (δ) units and coupling constants in Hz. Solvents used for the NMR
experiments are reported per compound. Details of the various HPLC-MS methods used are
provided in the supporting information section. Given the large number of final compounds
reported in the manuscript, only the characterization data of compounds 39, 41, 53 and 64 is
24
ACS Paragon Plus Environment

Page 25 of 34
Journal of Medicinal Chemistry
1
2
3
4
5
6
reported below. Details for the other compounds can be found in the supporting information
section.
7
8
9
3-((5-chloro-1-(4,4,4-trifluorobutyl)-1H-imidazo[4,5-b]pyridin-2-yl)methyl)-1-cyclopropyl-
1,3-dihydro-2H-imidazo[4,5-c]pyridin-2-one (39)
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
To
a
stirred solution of (5-chloro-1-(4,4,4-trifluorobutyl)-1H-imidazo[4,5-b]pyridin-2-
yl)methanol (1.2 g, 4.08 mmol) triphenyl phosphine (1.18 g, 4.5 mmol), 1-cyclopropyl-1,3-
dihydro-2H-imidazo[4,5-c]pyridin-2-one (787 mg, 4.5 mmol) in dry THF (100 mL) was added
DIAD (1 mL, 4.9 mmol) dropwise at room temperature. The reaction mixture was stirred at room
temperature overnight. After completion of the reaction, the mixture was concentrated to dryness
and the residue was purified by Prep HPLC on a RP Vydac Denali C18 - 10µm column, 250g,
5cm, eluting with a mobile phase of 0.25% NH4HCO3 solution in water, MeOH to yield the target
compound (1.5 g, 81% yield) as white solid. 1H NMR (400 MHz, CHLOROFORM-d)  ppm 1.00
(s, 2 H), 1.12 - 1.23 (m, 2 H), 1.83 - 1.99 (m, 2 H), 2.12 - 2.31 (m, 2 H), 2.91 (spt, J=3.50 Hz, 1
H), 4.38 - 4.54 (m, 2 H), 5.38 (s, 2 H), 7.13 (dd, J=5.27, 0.50 Hz, 1 H), 7.27 (d, J=8.28 Hz, 1 H),
7.61 (d, J=8.53 Hz, 1 H), 8.36 (d, J=5.27 Hz, 1 H), 8.77 (s, 1 H). HPLC method J, Rt = 0.86, purity
99%, m/z = 451 (M+H)+.
1-cyclopropyl-3-((1-(3-(methylsulfonyl)propyl)-1H-pyrrolo[3,2-c]pyridin-2-yl)methyl)-1,3-
dihydro-2H-imidazo[4,5-c]pyridin-2-one (41)
A solution of (3-bromo-1-(3-(methylsulfonyl)propyl)-1H-pyrrolo[3,2-c]pyridin-2-yl)methanol
(200 mg, 0.57 mmol) and 1-cyclopropyl-3H-imidazo[4,5-c]pyridin-2-one (101 mg, 0.57 mmol) in
THF (10 mL) was stirred at room temperature. Triphenylphosphine (151 mg, 0.57 mmol) was
added followed with DIAD (113 µL, 0.57 mmol) dropwise under a nitrogen atmosphere. The
25
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 34
1
2
3
4
5
6
7
8
9
resulting mixture was stirred overnight. The mixture was evaporated and diluted with
diisopropylether. The precipitate was collected by filtration (83 mg, 27% yield). This was
dissolved in methanol (15 mL) and stirred in the presence of a catalytic amount of palladium on
carbon under hydrogen atmosphere to get the title compound after column chromatography using
dichloromethane and methanol 9/1 as eluent (26.6 mg, 40% yield). 1H NMR (400 MHz, DMSO-
d6) δ ppm 0.90 - 0.97 (m, 2 H) 1.03 - 1.11 (m, 2 H) 1.97 (m, J=15.06, 7.78, 7.78 Hz, 2 H) 2.94 -
3.05 (m, 4 H) 3.16 (t, J=8.00 Hz, 2 H) 4.43 (t, J=7.65 Hz, 2 H) 5.36 (s, 2 H) 6.73 (s, 1 H) 7.29 (d,
J=5.27 Hz, 1 H) 7.62 (d, J=5.52 Hz, 1 H) 8.25 (d, J=5.27 Hz, 1 H) 8.26 (d, J=6.02 Hz, 1 H) 8.42
(s, 1 H) 8.83 (s, 1 H). HPLC method A, Rt = 1.15, purity 96%, m/z = 426 (M+H)+.
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
3-((5-chloro-1-(3-(methylsulfonyl)propyl)-1H-indol-2-yl)methyl)-1-(2,2,2-trifluoroethyl)-
1,3-dihydro-2H-imidazo[4,5-c]pyridin-2-one (53)
To a stirred solution of (5-chloro-1-(3-(methylsulfonyl)propyl)-1H-indol-2-yl)methanol (40 g,
132.5 mmol), 1-(2,2,2-trifluoroethyl)-1,3-dihydro-2H-imidazo[4,5-c]pyridin-2-one (28.8 g, 132.5
mmol) and tributylphosphine (49.7 mL, 199 mmol) in dry THF (400 mL) was added DIAD (38.7
mL, 199 mmol) dropwise at room temperature. The resulting mixture was stirred at room
temperature for one hour. The mixture was concentrated to dryness and the residue was crystallized
in ethanol/acetonitrile mixture (10/1) (1200 mL) and stirred overnight. The white crystals were
collected by filtration (36.5 g, 55% yield). 1H NMR (400 MHz, DMSO-d6)  ppm 1.88 - 2.02 (m,
2 H) 2.97 (s, 3 H) 3.15 (t, J=8.00 Hz, 2 H) 4.38 (t, J=7.50 Hz, 2 H) 4.89 (q, J=9.00 Hz, 2 H) 5.40
(s, 2 H) 6.48 (s, 1 H) 7.17 (dd, J=8.78, 2.01 Hz, 1 H) 7.44 (d, J=5.27 Hz, 1 H) 7.54 (d, J=8.78 Hz,
1 H) 7.57 (d, J=2.01 Hz, 1 H) 8.31 (d, J=5.27 Hz, 1 H) 8.49 (s, 1 H). HPLC method D, Rt = 1.68,
purity 100%, m/z = 501 (M+H)+.
26
ACS Paragon Plus Environment

Page 27 of 34
Journal of Medicinal Chemistry
1
2
3
4
5
6
4-(5-chloro-2-((1-(3,4-dimethoxyphenyl)-2-oxo-1,2-dihydro-3H-imidazo[4,5-c]pyridin-3-
yl)methyl)-1H-indol-1-yl)butanenitrile (64)
7
8
9
To a stirred solution of 4-(5-chloro-2-(hydroxymethyl)-1H-indol-1-yl)butanenitrile (3.34 g, 13.43
mmol), 1-(3,4-dimethoxyphenyl)-1,3-dihydro-2H-imidazo[4,5-c]pyridin-2-one (3.31 g, 12.21
mmol) and tributylphosphine (4.6 mL, 18.32 mmol) in dry THF (300 mL) was added DIAD (3.56
mL, 18.32 mmol) dropwise at room temperature. The resulting mixture was stirred at room
temperature for night. The mixture was concentrated to dryness and the residue was triturated in
DIPE the formed solid was filtered and then crystallized in ethanol. The white crystals were
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
collected by filtration (3.89 g, 63 % yield). H NMR (400 MHz, DMSO-d6)  ppm 1.94 (quin,
J=7.48 Hz, 2 H) 2.52 - 2.58 (m, 2 H) 3.79 (s, 3 H) 3.84 (s, 3 H) 4.32 - 4.40 (m, 2 H) 5.42 (s, 2 H)
6.56 (s, 1 H) 7.05 - 7.20 (m, 5 H) 7.55 (d, J=8.80 Hz, 1 H) 7.57 (d, J=1.98 Hz, 1 H) 8.23 (d, J=5.28
Hz, 1 H) 8.51 (s, 1 H). HPLC method J, Rt = 1.03 min, purity 100%, m/z = 502 (M+H)+.
RSV assay
Antiviral activity of the compounds against rgRSV224 was determined using a phenotypic cell-
based infection assay with fluorescence read-out15. Briefly, black 384-well clear-bottom
microtiter plates were filled in quadruplicate using a customized robot system with serial 4-fold
dilutions of compound in a final volume of 200 nL DMSO. Then, 20 µL of a HeLa cell suspension
(5 × 104 cells ml-1) in culture medium was added to each well followed by the addition of 20 µL
RSV (MOI = 0.02) in culture medium, or 20 µl culture medium for cell controls, using a multidrop
dispenser. Three days post virus exposure, viral replication was quantified by measuring eGFP
expression in the cells with a laser microscope. In parallel, compounds were incubated for 3 days
in a set of white 384-well microtiter plates and the cytotoxicity of compounds in HeLa cells (CC50)
27
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 34
1
2
3
4
5
6
was determined by measuring the adenosine triphosphate (ATP) content of the cells using the
ATPlite kit (PerkinElmer, Waltham, MA) according to the manufacturer’s instructions.
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
AUTHOR INFORMATION
Corresponding Author
*Tim H.M. Jonckers, Janssen Pharmaceutica NV, Janssen Pharmaceutical Companies of Johnson
& Johnson, Turnhoutseweg 30, 2340 Beerse, Belgium. Email: tjoncker@its.jnj.com
Present Adresses
† Current affiliation: Aligos Belgium BVBA, Gaston Geenslaan 1, bus (gebouw 3), 3001
Heverlee, Belgium.
Acknowledgments
The authors would like to thank Nigel Austin, Miek Desmidt, Brigitte Loenders, Filip Vanhoutte
and Kenny Simmen for their support in the pharmaceutical profiling of the RSV fusion
inhibitors. Results shown in this report are derived from work performed at Argonne National
Laboratory, Structural Biology Center at the Advanced Photon Source. SBC is operated by the
U.S. Department of Energy, Office of Biological and Environmental Research under contract
DE-AC02-06CH11357.
Abbreviations
RSV (Respiratory syncytial virus), SAR (Structure activity relationship), ADME (Absorption,
Distribution, Metabolism, Elimination), PK (Pharmacokinetic), RNA (ribonucleic acid), LRTI
(Lower respiratory tract infection), SI (Selectivity index), TDI (Time-dependent inhibition),
28
ACS Paragon Plus Environment

Page 29 of 34
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
HLM (human liver microsomes), SD (Sprague-Dawley), AUC (Area under curve), PFU (Plaque
forming units), PD (Pharmacodynamic); BALF (Broncho alveolar lavage fluid), LiOH (Lithium
hydroxide), DIAD (diisopropyl (E)-diazene-1,2-dicarboxylate). DIPE (diisopropylether)
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
PDB codes
Compound 39, PDB ID: 6VKC, Compound 41, PDB ID: 6VKD, Compound 64, PDB ID: 6VKE
Authors will release the atomic coordinates and experimental data upon article publication
Supporting information
The supporting information is available free of charge at
The supporting data contains 1H-NMR data, HPLC-MS details, purity determination of all final
compounds 4-74 (except 39, 41, 53 and 64); Protein crystallization and structure determination
protocols; In vivo efficacy results: dose response data; Molecular formula strings.
References
1. Nair, H.; Nokes, J. D.; Gessner, B. D.; Dherani, M.; Madhi, S. A.; Singleton, R. J.; O'Brien, K.
L.; Roca, A.; Wright, P. F.; Bruce, N.; Chandran, A.; Theodoratou, E.; Sutanto, A.; Sedyaningsih,
E. R.; Ngama, M.; Munywoki, P. K.; Kartasasmita, C.; Simões, E. A.; Rudan, I.; Weber, M.
W.; Campbell, H. Global burden of acute lower respiratory infections due to respiratory syncytial
virus in young children: a systematic review and meta-analysis. Lancet 2010, 375(9725), 1545-
1555.
2. Rudan, I.; Boschi-Pinto, C.; Biloglav, Z.; Mulholland, K.; Campbell, H. Epidemiology and
etiology of childhood pneumonia. Bull. World Health Organ. 2008, 86(5), 408-416.
3. Feltes, T. F.; Sondheimer, H. M. Palivizumab and the prevention of respiratory syncytial virus
illness in pediatric patients with congenital heart disease. Expert Opin. Biol. Ther. 2007, 7(9),
1471-1480.
4. Falsey, A. R.; Hennessey, P. A.; Formica, M. A.; Cox, C.; Walsh, E. E. Respiratory syncytial
virus infection in the elderly and high-risk adults. N. Engl. J. Med. 2005, 352(17), 1749-1759.
29
ACS Paragon Plus Environment