Identification and Optimization of Pyrrolo[3,2-d]pyrimidine Toll-like Receptor 7 (TLR7) Selective Agonists for the Treatment of Hepatitis B

Article abstract of DOI:10.1021/acs.jmedchem.7b00365

Pyrrolo[3,2-d]pyrimidines were identified as a new series of potent and selective TLR7 agonists. Compounds were optimized for their activity and selectivity over TLR8. This presents an advantage over recently described scaffolds that have residual TLR8 ac

Full text of DOI:10.1021/acs.jmedchem.7b00365

Subscriber access provided by NEW YORK UNIV
Article
Identification and Optimization of Pyrrolo[3,2-d]pyrimidine Toll-like
Receptor 7 (TLR7) Selective Agonists for the Treatment of Hepatitis B
David Craig McGowan, Florence Herschke, Frederik Pauwels, Bart Stoops, Ilham Smyej, Stefaan Last, Serge
Pieters, Werner Embrechts, Mourad Daoubi Khamlichi, Tine Thoné, Bertrand Van Schoubroeck, Wendy
Mostmans, Debbie Wuyts, Dorien Verstappen, Annick Scholliers, Dorien De Pooter, Deborah Dhuyvetter,
Herman Borghys, Marianne Tuefferd, Eric Arnoult, Jin Hong, Gregory Fanning, Jacques Bollekens, Vijay
Urmaliya, Ard Teisman, Helen Horton, Tim H.M. Jonckers, and Pierre Jean Marie Bernard Raboisson
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.7b00365 • Publication Date (Web): 03 Jul 2017
Downloaded from http://pubs.acs.org on July 4, 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 57
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Identification and Optimization of Pyrrolo[3,2ꢀ
d]pyrimidine Tollꢀlike Receptor 7 (TLR7) Selective
Agonists for the Treatment of Hepatitis B
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
,
†
,†
†
†
†
David C. McGowan,* Florence Herschke,* Frederik Pauwels, Bart Stoops, Ilham Smyej,
†
†
†
‡
†
Stefaan Last, Serge Pieters, Werner Embrechts, Mourad Daoubi Khamlichi, Tine Thoné,
†
†
†
†
Bertrand Van Schoubroeck, Wendy Mostmans, Debbie Wuyts, Dorien Verstappen, Annick
†
†
†
†
†
Scholliers, Dorien De Pooter, Deborah Dhuyvetter, Herman Borghys, Marianne Tuefferd,
§
|
†
†
†
Eric Arnoult, Jin Hong, Gregory Fanning, Jacques Bollekens, Vijay Urmaliya, Ard
†
†
†
†
Teisman, Helen Horton, Tim H.M. Jonckers, Pierre Raboisson
†
Janssen Pharmaceutica, N.V. Turnhoutseweg 30, 2340 Beerse, Belgium
§
Janssen Research & Development L.L.C., 1400 McKean Rd, Spring House, PA 19454
‡
Villapharma Research S.L., Parque Tecnológico de Fuente Álamo. Ctra. El EstrechoꢀLobosillo,
Km. 2.5ꢀ Av. Azul 30320 Fuente Álamo de Murcia, Murcia, Spain
±
Alios Biopharma, Inc., 260 East Grand Ave., South San Francisco, CA 94080
KEYWORDS
TLR, TollꢀLike Receptor; Interferon, IFNα; pyrrolo[3,2ꢀd]pyrimidine, HBV
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 57
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
ABSTRACT
Pyrrolo[3,2ꢀd]pyrimidines were identified as a new series of potent and selective TLR7 agonists.
Compounds were optimized for their activity and selectivity over TLR8. This presents an
advantage over recently described scaffolds that have residual TLR8 activity, which may be
detrimental to the tolerability of the candidate drug. Oral administration of the lead compound 54
effectively induced a transient interferon stimulated gene (ISG) response in mice and
cynomolgus monkeys. We aimed for a high first pass effect, limiting cytokine induction
systemically, and demonstrated the potential for the immunotherapy of viral hepatitis.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
INTRODUCTION.
Chronic hepatitis B affects over 350 million people worldwide, as it may eventually lead to
1
,2
cirrhosis, or hepatocellular carcinoma (HCC), responsible for one million deaths every year.
Under 1 in 10 patients achieve functional cure, defined as HBsAg loss and sustained HBVꢀDNA
suppression with or without the appearance of HBsAb (seroconversion), with the approved
3,4,5
treatments, nucleosides/nucleotides and interferonꢀα (IFNα).
Low cure rates are due to high
viral antigen loads, inability of the host immune system to control HBV and the
6
,7
immunosuppressive environment of the liver.
IFNα therapy, unlike nucleosides/nucleotides, has the potential to control all three issues when
directly exposed to hepatocytes and liver nonparenchymal cells. IFNα treatment of HBVꢀ
8
infected hepatocytes, which induces the expression of interferonꢀstimulated genes (ISG),
9
decreases both HBsAg and HBeAg levels. IFNα also induces the maturation of myeloid
10
dendritic cells towards a Th1 immuneꢀresponse, which is key to control HBV. The major
ACS Paragon Plus Environment

Page 3 of 57
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
reason for the limited clinical use is the intramuscular route of administration, leading to local
and systemic exposure of exogenous recombinant IFNα responsible for both injection site and
10
systemic side effects.
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
Conversely, orally administered agents inducing endogenous IFNα, like TLR7 agonists, could be
1
1
more potent and better tolerated. Toll Like Receptors (TLRs) play an important role in human
innate immunity by recognizing pathogen associated molecular patterns (PAMPs) present in a
1
2,13,14
variety of viruses and microorganisms.
TLR7 recognizes foreign genetic material in the
form of singleꢀstranded viral RNA. The receptor is localized in the endosome of plasmacytoid
15,16
dendritic cells (pDC), and B lymphocytes.
Detection of foreign nucleic acid in pDCs leads to
an antiviral immune response, resulting in an abundant production of endogenous IFNα, and
8,15,16
leading to the transcription of interferon stimulated genes (ISGs).
In B cells, activation of
TLR7 induces B cell maturation and secretion of antibodies which could promote HBsAg
1
7,18
seroconversion.
TLR8 shares a similar protein structure, is also involved in the detection of
viral RNA, but differs in that it is found to be expressed on myeloid dendritic cells and
1
9
monocytes rather than on pDC and B cells.
TLR8 agonism may have the advantages of a
strong Th1 response, but may suffer drawbacks such as the induction of proꢀinflammatory
2
0
cytokines. This manuscript focuses on the effort towards an oral, smallꢀmolecule TLR7
selective agonist with a high first pass effect would limit IFNα levels in the plasma, resulting in
a preꢀsystemic immune modifying response that could be beneficial toward the treatment of
chronic HBV.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 57
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
NH2
NH2
N
NH2
NH2
H
N
O
N
N
N
N
N
N
N
N
N
OH
N
O
N
N
N
H
H
N
N
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
SM-276001 (2)21
Imiquimod (1)
GS-9620 (3)
N
9
Figure 1. Described TLR 7 agonists and novel pyrrolo[3,2ꢀd]pyrimidine (9).
There are several TLR7 agonist series described to date, some examples are shown in Figure 1.
2
2
Imiquimod (1), is a marketed medication to treat actinic keratosis and basal cell carcinoma. A
second example TLR7 selective agonist from the literature is 2, of the 8ꢀhydroxyꢀadenine series,
21
described as a potent IFN inducer in both mice and monkeys. The pteridone series of agonists,
1
1
represented by GSꢀ9620 (3), was examined in clinical trials to treat HBV. A cynomolgus
monkey study performed with 3 has demonstrated that an oral dose of 2 mg/kg resulted in a Cmax
that was 2ꢀfold lower than that achieved following i.v. infusion of 0.1 mg/kg 3. However, mean
IFNα plasma levels induced following oral dosing were greater than 60ꢀfold higher than that
23
produced by i.v. infusion. Drug candidates having a rapid elimination after oral administration
would potentially improve the overall tolerability and limit the risk for tachyphylaxis, this was
indeed one objective in this program. Another objective was to maximize the selectivity of the
TLR7 agonist series over TLR8, to limit the induction of proꢀinflammatory cytokines (e.g.
2
4
TNFα). Indeed, 3 has been reported to be 30ꢀfold selective for human TLR7 vs TLR8.
A structural observation of known TLR 7 agonists from Figure 1 reveals several common
aspects. They all contain, at minimum, a bicyclic core scaffold, with an aryl group pendant, or
fused as in Imiquimod. Contained in the core structure is a minimal pattern of a nitrogen
hydrogen bond acceptor, and an adjacent, unsubstituted NH hydrogen bond donor. A third
2
ACS Paragon Plus Environment

Page 5 of 57
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
common aspect is the hydrophobic tail, typically four carbon atoms, connected by an oxygen or
nitrogen atom, and positioned three to four bonds away from the aryl group. If indeed there is a
recognition element of the TLR7 protein toward RNA bases then perhaps new scaffolds would
best imitate the shape and electrostatics of adenine and guanine. The design of pyrrolo[3,2ꢀ
d]pyrimidine (9) started from the concept to use the core pyrimidine scaffold of previously
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
2
5,26
described TLR agonists,
keeping the hydrogen bond donor and acceptor pattern of the
pyrimidine and adding a fused ring so that the pendant benzyl group was placed proximal to the
alkyl chain. Among the many permutations attempted in our laboratory, we would like to focus
this manuscript on the success of the TLR7 selective pyrrolo[3,2ꢀd]pyrimidine series starting
from the initial compound synthesized (9). Indeed, this novel structure holds all three of the
desired design attributes. The butyl chain and pendant benzyl group project ipsilaterally from the
bicyclic aromatic scaffold containing a similar donor acceptor pattern as the described agonists
in Figure 1.
We employed two in vitro assays to assess the activity of compounds during the SAR
exploration. In the first in vitro assay, the ability of compounds to agonize human TLR7 and
TLR8 was assessed in a cellꢀbased reporter assay using HEK293 cells transiently transfected
with a wildꢀtype TLR7 or TLR8 expression vector and a NFκBꢀluc reporter construct. To
confirm the activity compounds were tested in the second in vitro assay, using IFNꢀmediated
antiꢀHCV activity in the replicon system upon incubation with conditioned media from human
peripheral blood mononuclear cells (hPBMCs). In this assay, activation of human TLR7 on
pDCs results in the production of IFN. Promising derivatives were evaluated in rat
pharmacokinetic studies and a mouse pharmacokineticꢀpharmacodynamic model. The selected
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 57
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
lead candidate was tested in the mouse hydrodynamic injection (HDI) model and finally in
cynomolgus monkey.
MEDICINAL CHEMISTRY STRATEGY
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
Our strategy was to systematically investigate the structure activity relationship of this novel
series of pyrrolo[3,2ꢀd]pyrimidines, which was divided into three subꢀseries. The initial subꢀ
series, represented by analogues 9 to 27 (Table 1), where the nꢀbutyl amine was conserved and
the phenyl ring is replaced by aliphatic and heterocyclic substituents, was designed to optimize
activity and selectivity toward TLR7. The second subꢀseries (compounds 28 to 45) allowed the
exploration of alternatives to the aliphatic amine. Finally, in the third subꢀseries (compounds 46
to 59), to afford selective TLR7 agonists with limited offꢀtarget activity, the nitrogen atom of 42
was replaced with an oxygen that afforded the lead compound 54 for further evaluation. In the
context of this project, and in contrast to a traditional medicinal chemistry program, rapidly
metabolized compounds were desired, thus limiting compound exposure and systemic cytokine
activation. Furthermore, it has been demonstrated that a robust antiꢀviral response can be
27
achieved in nonꢀhuman primates with less than daily dosing of a TLR7 agonist. Despite the
expected low compound plasma concentrations, an effort was made to remove offꢀtarget activity
including, for example, the potential to inhibit the hERG potassium ion channel.
CHEMISTRY
The synthesis of this series of pyrrolo[3,2ꢀd]pyrimidines is depicted in Scheme 1 and began with
2,4ꢀdichloropyrrolo[3, 2ꢀd]pyrimidine (4). Alkylation with benzyl bromide in DMF afforded Nꢀ
benzyl analogue 5 in 50% yield. Displacement of the first chlorine atom with nꢀbutylamine in
1,4ꢀdioxane afforded 6 in a modest 67% yield. On the side, the 2ꢀchlorine was converted to a
ACS Paragon Plus Environment

Page 7 of 57
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
hydroxyl group by heating in aqueous acetic acid to afford 7. Finally, nucleophilic aromatic
substitution at the 2ꢀchlorine with sodium azide in a mixture of NMP/water afforded the 2ꢀazido
intermediate 8, followed by a Staudinger reduction provided the initial agonist 9 after
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
28
purification via silica gel column chromatography.
For later analogues, the synthesis was ameliorated to a more scalable route, avoiding the
2
9
hazardous use of azides and high temperature microwave reaction. Scheme 2 depicts an
example of the synthetic pathway starting with the condensation between ethyl 3ꢀaminoꢀ1Hꢀ
pyrroleꢀ2ꢀcarboxylate and 1,3ꢀbis(methoxycarbonyl)ꢀ2ꢀmethylꢀ2ꢀthiopseudourea, analogous to
3
0
methods described in the literature. The resulting 4ꢀhydroxyꢀpyrrolo[3,2ꢀd]pyrimidyl
carbamate (10) was converted to the 4ꢀchloroꢀpyrrolo[3,2ꢀd]pyrimidinyl carbamate (11) via
phosphorus oxychloride in 46% yield. Alkylation of the pyrrole nitrogen under standard
Mitsunobu conditions proceeded in moderate yield (65%) to afford 12. This was followed by
base mediated deprotection of the 2ꢀamino carbamate (13) and finally displacement of the
chlorine by an amine nucleophile in a polar solvent provided 37, or by an oxygen nucleophile, as
in the third subꢀseries, facilitated by NaH (e.g. 55).
a
Scheme 1. Representative Synthetic Route of Pyrrolo[3,2-d]pyrimidines.
a
Reagents and conditions: (a) BnBr, Cs CO , DMF, 70ºC, 50% isolated yield; (b) nꢀbutylamine, 1,4ꢀdioxane, 100ºC,
67% yield; (c) HOAc, aq. HCl, 100ºC, 24h, 40%; (d) NaN , NMP/H O: 9/1, microwave 170ºC, 5h, 86%; (e) PPh ,
2
3
3
2
3
1,4ꢀdioxane:H O 9:1, 120ºC, 16h, 76%.
2
a
Scheme 2. Representative Alternate Synthesis of Pyrrolo[3,2-d]pyrimidines.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 57
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
a
Reagents and conditions: (a) 1,3ꢀbis(methoxycarbonyl)ꢀ2ꢀmethylꢀ2ꢀthiopseudourea, HOAc, NaOCH , CH OH, rt,
3
3
1
6h, 87%; (b) POCl , DIPEA, ACN, 70ºC, 1h, 46%; (c) 2ꢀpyridylmethanol, PPh , DIAD, THF, rt, 30 min., 65%; (d)
3
3
NaOH, 1,4ꢀdioxane, 60ºC, 5h, 75%; (e) 5ꢀmethylꢀ2ꢀoxazolemethanamine, DMSO, 110ºC, 16h, 15%; (f) (5ꢀ
methyloxazolꢀ2ꢀyl)methanol), NaH, THF, 0 to 50ºC, 30 min., 22%. See WO2014056953 and WO2014207082.
RESULTS AND DISCUSSION
StructureꢀActivity Relationship. Changes in the 2ꢀamino group led to inactivity (e.g. 7, 8 were >
2
5 µM [hTLR7, 8]) and so further modifications were not pursued. The first example, having an
unsubstituted phenyl ring and an nꢀbutyl carbon chain (9) showed exceptional TLR7 activity, but
unsatisfactory selectivity over the hERG receptor (IC = 2.1 ꢀM). When the phenyl group was
5
0
more distal to the scaffold (e.g. 14, 15) TLR7 potency resulted in the same range as 9, however
undesired TLR8 activity was observed. Replacement of the phenyl group of 9 with an amide led
to lower activity (16), but improved with the isobutene replacement (17). The thiophene
analogue 18 had TLR7 selectivity superior to 9 in contrast to the 2ꢀthiazole congener 19 with a
selectivity ratio of only 12. The imidazo[1,2ꢀa]pyridine fragment (20) and the three pyridine
analogs 21 to 23 demonstrated that TLR7 activity was gained when the nitrogen of the
heterocycle was adjacent to the methylene group. This substitution pattern showed that the
ACS Paragon Plus Environment

Page 9 of 57
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
pyridine ring could be modified (24-26) and still retain TLR7 selectivity with the exception for
the 2ꢀpyridylꢀ6ꢀmethoxy analogue (27) that exhibited outstanding potency on TLR7, but carried
undesired TLR8 activity. Indeed, 21 displayed remarkable activity and selectivity, and lower
CYP450 inhibition than the phenyl (9) or thiophene (18) congeners (see supporting information)
and would be held constant for exploration in the following subseries. In addition, extensive
possibilities exist to access substituted pyridines, facilitating future SAR exploration.
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
Table 1. The Structure Activity Relationship of Benzyl Group Replacements.
Human TLR7 Human TLR8 Selectivity hPBMC
Entry
R
b
b
(
LEC ꢀM)
(LEC ꢀM)
Ratio
(ꢀM)
9
0.23
0.07
0.16
>25
1.16
3.78
>109
0.1
14
5
0.06
0.06
1
1
5
6
24
6.95
>25
>5
1.0
1
1
7
8
0.4
>25
>25
>63
0.22
0.03
0.13
>192
1
2
9
0
0.19
0.15
2.27
>25
12
0.13
0.04
>167
2
1
0.15
>25
>167
0.06
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 57
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
2
2
1.36
1.53
>25
>25
>18
>16
0.26
0.75
23
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
2
2
4
5
N
0.08
0.3
>25
>25
>313
>83
<0.01
0.28
26
0.18
>25
>139
74
0.14
2
7
0.008
0.59
<0.01
a
b
All compounds had a CC50 > 24 ꢀM in the HEK 293 cytotoxicity assay (see supporting information). LEC; least
effective concentration values were averaged when determined in two or more independent experiments. See
supporting information.
The subsequent subseries of analogues would focus on systematic changes of the amine group.
Variation of the amine while holding the unsubstituted pyridyl group constant is described in
Table 2. Differences in activity were seen when the butyl chain was replaced by other aliphatic
or heteroaromatic groups, all were selective for TLR7 over TLR8. An ether variation on the
butyl chain of 21 displayed a loss of activity (28). The same held true for alkyl ethers 29 and 30.
A modest decline in activity for aminoalcohol 31 was observed when compared to 21, but
highlighted that a polar group was well tolerated. Heterocyclic substituents presented a range of
activity depending on the ring size and position of the heteroatoms. The bisꢀpyridyl analogues
ACS Paragon Plus Environment

Page 11 of 57
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
(
32, 33) did not show an increase in potency over the butyl chain, and furthermore the racemic
alphaꢀmethyl group of 33 did not display a favorable increase in activity and so further
modifications were not pursued. In contrast, thiazole and 1,3,4ꢀoxadiazole (34, 35) ring
modifications were deleterious to activity, suggesting that the electrostatic properties of the ring
are important in addition to the regiochemisty of the heterocycle. The oxazoles displayed higher
potency when the methyl group was placed next to the oxygen (37), as opposed to neighboring
the nitrogen atom (38). Both oxazoles showed binding to the serotonin receptors (e.g. IC50
5HT2a = 0.65 and 2.3 µM, respectively). The methyl isoxazole 42 was an equipotent TLR7
agonist and had less inhibition of the 5HT2a receptor (IC = 6.2 µM). The cyclopropyl congener
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
(43) was also an effective agonist, but displayed more potent hERG inhibition than 42 (hERG
IC = 0.2 vs. 1.2 µM respectively). Finally, if the NꢀO bond in the isoxazole was reversed (45)
5
0
then potency was lost.
a
Table 2. Structure Activity Relationship of Amine Substitution.
hTLR7
hPBMC
(ꢀM)
2.2
Entry
R
b
(
LEC ꢀM)
1.81
2
2
8
9
4
1
.45
.24
0.5
0.5
3
0
0
.43
0.05
0.07
31
0.4
3
2
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 57
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
.26
0.04
3
3
S
6.13
2.2
3
3
4
5
N
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
.24
0.09
R1
H
R2
H
X
O
O
O
S
3
3
38
39
6
7
0.52
0.06
1.16
3.04
0.06
0.13
0.006
0.15
0.56
0.009
CH
H
H
3
3
3
CH
H
3
H
4
0
CH
H
S
4
4
4
4
1
2
3
4
H
CH
ꢀ
0.9
0.42
0.006
0.008
0.02
0.04
0.05
0.34
iPr
4
5
0.49
0.13
a
All compounds were > 25 µM LEC [hTLR8], and CC50 > 24 ꢀM in the HEK 293 cytotoxicity assay described in
the supporting information. LEC; least effective concentration values were averaged when determined in two or
more independent experiments. See supporting information.
b
Compound 42 had achieved the potency and selectivity that was desired and was identified to
have no appreciable receptor kinase activity, and no CYP inhibition across several isozymes (see
supporting information). However, 42 remained a strong hERG binder (IC = 1.2 µM).
50
The strategy to identify the best lead candidate was twoꢀfold: First, the preferred 5ꢀ
methylisoxazole moiety was held constant allowing the examination of the effect of substituted
3
1
pyridines, that could influence hERG receptor binding as suggested in the literature. Secondly,
it has been described that introduction of an oxygen atom into a aliphatic chain can reduce the
32
propensity of hERG binding. Thus, we explored pyridine ring variations, and converted the
amine connection of the heterocycle (Table 3, R ) to an ether (Table 3, X = O).
5
ACS Paragon Plus Environment

Page 13 of 57
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
a
Table 3. Modifications to the Heterocyclic Rings.
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
Human
hERG
IC50
(ꢀM)
hPBMC
(ꢀM)
Entry
R1
R2
R3
R4
X
R5
TLR7
cLogP
b
(
LEC ꢀM)
46
H
H
H
F
H
H
H
H
H
CH
3
O
CH
CH
3
3
O
N
N
N
N
N
N
N
N
O
O
O
O
O
O
0.003
0.006
0.070
0.086
0.050
0.050
0.050
0.050
0.4
0.0006
0.001
0.009
0.009
0.03
0.7
2.0
1.9
1.3
1.3
2.1
2.1
1.9
1.9
1.1
0.8
1.7
0.5
1.3
1.6
4
7
H
O
2.3
48
F
H
CF
H
H
H
H
H
H
H
H
H
H
2.0
4
5
5
5
5
5
5
5
9
0
1
2
3
4
5
6
H
H
H
H
H
H
H
H
H
H
H
2.1
H
H
Cl
H
H
H
H
H
0.5
3
CF
H
0.02
n.d.
0.9
3
0.04
Cl
H
0.02
0.5
0.061
0.10
>50
>50
44.7
>50
2.2
H
0.6
H
0.8
0.099
1.96
57
H
6
5
8
CH
CH
3
3
O
H
0.07
0.03
0.006
0.006
59
O
H
1.6
a
All compounds were > 25 µM LEC [hTLR8], and showed CC > 24 ꢀM in the HEK293 cytotoxicity assay
described in the supporting information. LEC values were averaged when determined in two or more independent
experiments. n.d., not done.
50
b
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 57
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
Variations in the pyridine substitution (compounds 46 to 53) afforded several potent agonists,
especially when substituted by electron rich moieties (46, 47). It is described in the literature that
within a given series of compounds there can be a correlation between hERG binding and a
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
33
measurement of lipophilicity such as cLogP. However, within this chemical series, there was a
weak correlation between cLogP and hERG IC . On the contrary, the best progress in avoiding
5
0
hERG inhibition came with the discrete structural modification of 42 to the corresponding ether
4, despite the two compounds having nearly the same cLogP of 1.1. This modification led to an
5
improved cardiovascular safety profile with a minor loss in TLR7 potency. Further heterocyclic
ether variants (oxazole 55, thiazole 56, oxadiazole 57) also bore relatively no hERG inhibition
and had intriguing differences in potency that may be rationalized by their difference in charge
distribution.
The electrostatic potential alignment of analogs 54 to 57 showed that the most potent compounds
share a similar electrostatic profile (Figure 2, see supporting information). There are two
positions on the ring on either side of the methylene group that are key. One alpha position
(South side, Figure 2) can be electronegatively charged, while the other (North side) must be
electropositively charged (thiazole) or neutral as the potency decreases when this side of the
molecule becomes strongly electronegative, as in oxadiazole derivative 57. The electrostatic
profile of the best substituent could be used as a signature for further exploration of this series.
Finally, substituted pyridines 58 and 59 once again displayed that potency could be regained by
the addition of electron donating methoxy groups on the pyridine ring, but unfortunately afforded
decreased selectivity over the hERG antitarget.
ACS Paragon Plus Environment

Page 15 of 57
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 2. Left: Electrostatic potential comparison, based on AM1ꢀBCC charges, between isoxazole (54), oxazole
(
55), thiazole (56), and oxadiazole (57). Right: Predicted binding mode by docking (Glide XP) for compound 54 in a
monkeyTLR7 structure (one TLR7 monomer in cyan and the other in green). See supporting information.
The binding mode of 54, obtained by docking into monkey TLR7 (see supporting information),
illustrated in Figure 2, corresponds to the expected binding mode and is comparable to the
34
binding mode described for Resiquimod.
Strong, direct dual hydrogen bonds between the 2ꢀamino group of 54, the acidic side chain of
Asp555, and the backbone of Thr586 (Carbonyl) were observed. An additional Hꢀbond between
the 3ꢀposition of the pyrimidine and the backbone of Thr586 (Amino) contributes to the
stabilization of the compound in the binding site. The scaffold is further stabilized by a πꢀ
stacking interaction with Phe408 on one side, and by CHꢀπ interactions with Leu557 on the other
side.
The methylisoxazole is deeply buried in the hydrophobic pocket formed by residues Phe349,
Phe351, Val381 and Ile585, stabilized by a πꢀstacking interaction with Phe351, and has good
surface complementarities. In addition, the pyridine ring of 54 has a CHꢀπ interaction with the
Cα of Gln354.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 57
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
In comparison, the binding mode of 54, is less favorable in hTLR8 for two main reasons
(illustrated in figure S1): 1) Leu557 is replaced by Asp545 in hTLR8. The scaffold no longer has
a CHꢀπ interaction. Instead, Asp545 is making nonꢀfavorable electrostatic interactions with the
compound. 2) Thr586 is replaced by Thr574 in hTLR8. The side chain of Thr574 adopts a
different orientation, providing less room to accommodate 54. The Thr574 side chain is making
only one Hꢀbond with the 2ꢀamino group instead of two with its backbone in monkeyTLR7. In
addition, the Phe351 (TLR7) is replaced by Tyr358 in TLR8 and does not allot enough room to
accommodate the methylisoxazole in the hydrophobic pocket. The combination of both a less
stable interaction of the scaffold and the methylisoxazole, predicts that 54 would be less potent
on hTLR8.
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
Agonist 54 was found to be selective for TLR7 over TLR8 in the same cellular reporter assay
using HEK293 cells transiently transfected with mouse, rat, or cynomolgus TLR7 or TLR8
expression vector and NFκBꢀluciferase reporter construct (see supporting information). In
addition, no direct activation of NFκB was observed. After repeating at higher concentration, 54
was found to be very selective over TLR8 (>100 µM LEC human TLR8).
a
Table 4. Activity of 54 Across Species in vitro.
Transient transfection
of HEK293 cells with:
Avg. LEC
(µM)
Species
Human
Mouse
Rat
Cynomolgus monkey
Human
0.41
0.34
0.33
0.80
>100
>25
TLR7 + NFκBꢀluc
Mouse
TLR8 + NFκBꢀluc
NFκBꢀluc
Rat
>25
>25
Cynomolgus monkey
>25
a
see supporting information for n, and confidence interval.
ACS Paragon Plus Environment

Page 17 of 57
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
The permeability of 54 was measured in a MDCKꢀMDR1 cellular assay at 5 ꢀM. 54 displayed
6
ꢀ1
decent permeability with a Papp (A→B, in the presence of GF120918) of 41 x 10 cm s , and an
efflux ratio of 18, indicating that the compound is a Pꢀgp substrate. 54 demonstrated low
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
inhibition across five CYP450 isozymes (IC > 10 ꢀM, see supporting information) and was
50
also not a time dependent inhibitor of CYP450 3A4. Table 5 lists the in vitro metabolic stability
of 54 in liver microsomes and wholeꢀcell hepatocytes across species. Intrinsic clearance (CLint)
in liver microsomes was moderate in mouse, dog and human and high in rat and monkey. The
intrinsic clearance in hepatocytes evaluated the potential for nonꢀCYP mediated metabolism and
was low to moderate in mouse, dog and human and high in monkey and rat. Plasma protein
binding across species remained relatively low and ranged from 54 to 63% bound. 54 had
3
limited inhibition of the hERG potassium ion channel Hꢀdofetilide binding in vitro (IC > 50
5
0
TM
ꢀ
M), which correlated with a relatively low functional hERG inhibition in the QPatch planar
35
patch clamp assay (32% inhibition at 3 ꢀM). To evaluate whether this effect, albeit low, or
other undetected compound characteristics could lead to undesired electrophysiology effects in
vivo, 54 was subjected to a cardiovascular safety pharmacology study in anesthetized female
guinea pigs. In this model, electrophysiological changes were measured after administration of
54 by i.v. infusion in pentobarbital anesthetized animals (cumulatively dosed up to 4.9 mg/kg
i.v.). From this study, no major cardioꢀelectrophysiological effects were observed up to maximal
total plasma exposures of 9700 ng/mL (at the end of the infusion). There was no tendency for the
formation of micronuclei in vitro in the presence and absence of rat S9 fraction, and no potential
for mutagenicity revealed by the AMES II test. In further support of the lead compound,
cytokine induction was measured after compound stimulation of whole blood of healthy
volunteers ex vivo and included IFNα, IFNγ, ILꢀ12p40, and ILꢀ12p70.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 57
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
a
Table 5. In vitro Metabolic Stability and Plasma Protein Binding of 54.
CLint
microsomes
ꢀL/min/mg) (ꢀL/min/10 cells)
CLint
Plasma
protein binding
(% bound)
Species
hepatocytes
6
(
Mouse
Rat
Dog
55
>346
44
87
25
19
>277
7
80
6
63
61
62
52
54
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
Cynomolgus monkey
Human
a
Values were rounded to the nearest whole number.
Pharmacokinetics. Among preclinical species, the in vitro metabolism in rat hepatocytes was
exceptionally high, and the in vivo clearance was predicted to be above rat liver blood flow. A
pharmacokinetic study in healthy male SpragueꢀDawley rats investigated the exposure of 54 in
systemic circulation following intravenous administration at 2.5 mg/kg and in the portal vein,
liver and systemic circulation following oral administration of 10 mg/kg. The i.v. clearance
correlated well with the in vitro clearance, and exceeded rat liver blood flow, resulting in a very
short terminal halfꢀlife. The maximum concentration in the portal vein was equal to 314 ng/mL
and occurred one hour postꢀdose. The concentration measurements in the liver and systemic
circulation were below quantifiable limits. Oral administration in dogs (1 mg/kg) resulted in fast
absorption (Tmax at 0.5 h, first sampled time point). The Cmax was 125 ng/mL and the systemic
concentration dropped below 10 ng/mL, 4 hours post administration. The systemic concentration
in dog was higher than in rat which is in line with the in vitro stability in microsomes and
hepatocytes.
Pharmacokinetic and pharmacodynamic studies in mice and cynomolgus monkeys. A
mouse in vivo model was used to demonstrate the initial proof of concept to induce endogenous
IFNα. Single oral administration of 0.3, 1, 3 and 10 mg/kg doses of 54 were given to healthy,
female, fasted C57Bl/6 mice. Concentration of compound and mouseꢀIFN via ELISA were
ACS Paragon Plus Environment

Page 19 of 57
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
measured from the plasma and compared to vehicle. The results are shown in Table 6. 54 was
found to be rapidly cleared, in conjunction with our target profile. Both Cmax and AUC increased
less than dose proportionally between 0.3 and 3 mg/kg and more than doseꢀproportionally
between 3 and 10 mg/kg. Tmax occurred between 0.5 and 1 h, and a liver to plasma ratio of 2 was
measured. Levels of IFNα peaked between one and two hours post administration in both
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
plasma and in the liver. Increase in IFNα was doseꢀproportional in plasma, but sigmoidal in the
liver, suggesting a possible saturation of endogenous IFNα production especially in the liver
(Table 6). ISG (interferonꢀstimulated gene) expression was measured in parallel by RTꢀqPCR.
Interestingly, all three Isg15, Oas1b and Mx1 were upregulated to a higher level in the liver than
in the plasma, while IFNα, Cxcl10 or the target Tlr7 were upregulated similarly in blood and
liver (Table 6). These data reveal that 54 can induce an antiviral ISG response without inducing
an IFNα response at a low dose. One potential drawback of the TLR7 selective agonist 2, are the
2
2
significant plasma concentrations observed at 8h in mice after an oral dose of 1mg/kg, in
contrast to 54 that was below quantifiable limits at the same dose and timepoint. Therefore,
compound 54 has rapid clearance after oral administration in mice, limiting compound exposure
and systemic cytokine activation, resulting in an improved potential tolerability profile.
Table 6. Mean Maximum Levels in the Plasma and Liver and ISG upregulation, induced
a
by 54 after oral dosing in C57Bl/6 mice.
Dose (mg/kg)
0.3
1
3
10
Cmax plasma (ng/mL)
Cmax liver (ng/g)
3.7
6.8
105
BQL
1.85
16
7.1
6.3
205
55
2.1
14
2.8
23.0
560
470
2.45
11
72.9
152.3
1365
580
2.2
14
IFNαmax plasma (pg/mL)
IFNαmax liver (pg/g)
b
Oas1bmax plasma
b
Oas1bmax liver
b
Isg15max plasma
1.35
10
12
60
3.6
2
60
18
315
3.7
1.95
75
20
745
4.05
1.75
100
19
775
4.15
b
Isg15max liver
b
Mx1max plasma
b
Mx1max liver
b
Tlr7max plasma
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 57
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
b
Tlr7 liver
3.15
2.45
2.9
3.7
max
a
Groups of 9 mice were dosed with 0.3, 1, 3 or 10 mg/kg of 54 At 0.5 hours, blood was drawn to quantify
compound. At 1, 2, or 4 hours, three mice of each group were sacrificed and blood was drawn to measure plasma
and liver compound levels (by LCꢀMS), IFNα (by ELISA) and ISG expression levels (by RTꢀqPCR). Values were
b
rounded to the nearest whole number. BQL, below quantifiable limit; foldꢀchange relative to vehicle.
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
The potential of 54 to inhibit HBV replication in vivo was tested in the hydrodynamic injection
(
HDI)ꢀHBV mouse model. This model allows an initial assessment of the antiꢀHBV activity, as
3
6
mice develop an acute infection which is cleared within two weeks. 54 was administered orally
to C57Bl/6 male mice one day after HDI, for five consecutive days at 0.3, 1, 2.5, and 5 mg/kg
QD and was found to be well tolerated (no clinical observations were noted). The dose was
chosen based on the robust ISG upregulation observed in healthy C57Bl/6 mice (Table 6). 54
induced a 2.7 log decrease in serum HBV viral load from 0.3 mg/kg, and a maximum 3.1 log
decrease was observed for doses between 1 and 5 mg/kg (Figure 3). Also, one day after the last
dose, mice were sacrificed to evaluate liver HBV viral load (Figure 3). A small but statistically
significant (Oneꢀway ANOVA with Dunnett’s multiple comparison) decrease was observed at all
doses, and was maximal for the 5 mg/kg dose (0.63 log decrease). In contrast, a similar decrease
in viral load, in either the serum or the liver, required a higher dose of 3 (20 mg/kg).
Figure 3. Efficacy of 54 in the HDIꢀHBV mouse model. Left: Serum HBV viral load over time (mean +/ꢀ SD).
Right: Liver HBV viral load at end of dosing (qPCR, mean +/ꢀ SD). Seven groups of eight male C57Bl/6 mice, 4ꢀ6
weeks old, were administered 20 µg of pAAV2ꢀHBV1.3mer genotype D plasmid by hydrodynamic injection. 1 day
later, groups 1 to 7 received compound 54 orally at 0, 0.3, 1, 2.5, and 5 mg/kg QD, respectively, for 5 days.
HPbCD20% at acidic pH was used as vehicle. At different time points, blood was sampled to measure serum HBV
ACS Paragon Plus Environment

Page 21 of 57
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
viral load by qPCR. 1 day after the last dose, mice were euthanized to measure liver HBV viral load by qPCR.
LLOQ, lower limit of quantification.
Healthy cynomolgus monkeys were chosen as the next preclinical pharmacokineticꢀ
3
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
pharmacodynamic model, as they are the closest available model to human immunology. Oral
doses of 1, 3 and 10 mg/kg of 54, were well tolerated (no clinical observations and no relevant
change in body weight, body temperature observed). Maximum mean plasma concentration
occurred at 0.25 to 2 hours post dose and increased doseꢀproportionally, despite the large
variability observed. The low oral bioavailability (<1% F) can potentially be attributed to a high
first pass effect, assuming that absorption was similar to that generated in the rat hepatic portal
vein pharmacokinetic study. According to the mode of action, a transient and doseꢀdependent
3
8ꢀ40
decrease in lymphocyte counts and increase in neutrophil counts was observed from 1 mg/kg.
Cytokines were measured in plasma by Luminex or ELISA. IFNα production was detected in all
animals starting from the 3 mg/kg dose, and showed a more than doseꢀproportional profile of
maximum response, despite high interꢀindividual variability. Interestingly, an ISG response was
induced starting from 1 mg/kg: Isg15, Oas1b, Oasl, Mx1, Cxcl11, Ifit44l where upregulation
(measured by microarray) was detected from 1 mg/kg and peaked at four to eight hours postꢀ
dose. A second wave of gene induction was observed at 72 hours postꢀdose, including ISG
Cxcl10, as well as monocyteꢀinduced Ccl18. Monocyte activation was also suggested by the
production of MCP1 (from 1 mg/kg) and ILꢀ1Ra (from 3 mg/kg). General proinflammatory
cytokines ILꢀ6 and GꢀCSF were measurable from 10 mg/kg, while antiꢀinflammatory cytokine
ILꢀ10 was induced from 3 mg/kg. This is encouraging in regards to the safety of compound 54.
Overall, these results demonstrate that 54 is more than simply an interferon inducer and warrant
further evaluation for HBV indications.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 57
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
Table 7. Mean Plasma Pharmacokinetic and Pharmacodynamic Parameters of 54 in
Cynomolgus Monkey Following Administration by Oral Gavage.
a
Dose (mg/kg)
Cmax(ng/mL)
Tmax(h)
1
3
11.2
10
23.4
4.12
1ꢀ2
0.5ꢀ2
26.7
0.5ꢀ2
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
AUC0ꢀlast(ng.h/mL)
NA
BQL
57.5
a
IFNα max
120 (3/5)
2980 (5/5)
a
MCP1 max
325 (2/5) 1560 5/5) 8295 (5/5)
75 (5/5) 1820 (5/5) 10575 (5/5)
20 (4/5) 10 (3/5)
90 (3/5) 75 (2/5)
20 (3/5) 45 (3/5)
a
ILꢀ1Ra max
ILꢀ6 max
a
525 (5/5)
120 (4/5)
125 (5/5)
a
GꢀCSF max
a
ILꢀ10 max
a
pg/mL, frequency; BQL, below quantification limit; NA, not applicable. Five cynomolgus monkeys were
administered 1, 3 or 10 mg/kg of 54 by oral gavage. At 0 (preꢀdose), 0.25, 0.5, 1, 2, 4, 8, 24, 48 and 72 hours postꢀ
dose, blood was sampled to measure compound concentrations in plasma, cytokine levels (by ELISA or Luminex)
and gene expression (by microarray). Doses were administered in different animals, but staggered by 3 days for
acute toxicity evaluation.
CONCLUSION
In this manuscript, the research on a novel class of TLR7 selective agonists for immunotherapy
of hepatitis B was described. This new series of pyrrolo[3,2ꢀd]pyrimidines was designed by
taking key elements of chemical series previously described in the literature and combining them
into a new structure. We found 54 to be potent and selective over other TLRs in preclinical
species. Oral administration of 54 was well tolerated in vivo, displayed a high first pass effect,
and induced an interferon stimulated gene response in both mice and cynomolgus monkeys. This
ISG response, was of higher magnitude in the liver, the site of viral infection, than in the
periphery in mice. In the HDIꢀHBV mouse model, 54 reduced viral load in the serum and in the
liver. In cynomolgus monkey, 54 did not induce IFNα systemically after a low (1 mg/kg) oral
dose, but induced an upregulation of ISGs. In summary, we have identified a series of novel and
potent TLR7 agonists and presented data which indicate the potential for the immunotherapy of
viral hepatitis.
ACS Paragon Plus Environment

Page 23 of 57
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
EXPERIMENTAL SECTION
Intrinsic clearance (CL ) of 54. CL was determined in mouse, rat, dog, cynomolgus
int
int
monkey, and human liver microsomes. Incubations were performed at 37°C, at a compound
concentration of 1 ꢁM, and a microsomal protein concentration of 0.5 mg/mL. Samples were
removed at intervals of up to 60 min and analyzed for the concentration of compound to
determine its intrinsic clearance rate. 54 was incubated in mouse, rat, dog, cynomolgus monkey,
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
6
and human hepatocytes (10 cells/mL) at 1 µM for 0, 10, 20, 40, 60 and 120 min. Samples were
removed at intervals of up to 120 min and analyzed for the concentration of compound to
determine its intrinsic clearance rate.
Plasma Protein Binding. The free fraction in mouse, rat, dog, cyno and human plasma was
determined by rapid equilibrium dialysis (RED device, Thermo Fisher Scientific, Geel,
Belgium). The RED device consists of a 48 well plate containing disposable inserts bisected by a
semipermeable membrane creating two chambers. A 300 ꢁL aliquot of plasma containing test
compound at 5 ꢁM was placed one side and 500 ꢁL of phosphate buffered saline (PBS) on the
other. The plate was sealed and incubated at 37 °C for 4.5 h. Samples were removed from both
the plasma and buffer compartment and analyzed for test compound using a specific LCꢀMS/MS
method to estimate free and bound concentrations.
Permeability and Efflux. The in vitro permeability and potential to be transported by Pꢀ
glycoprotein (Pꢀgp) was determined using an MDCK cell line transfected with human MDR1 (Pꢀ
glycoprotein). 5 ꢁM 54 was added to the apical (A) or basolateral (B) side of a confluent
monolayer of MDCKꢀMDR1 cells. Apparent permeability in the A→B direction in absence (Papp
ꢀ6
ꢀ6
A→B, 5.5 x 10 cm/s) and presence (P A→B, 41 x 10 cm/s) of GF120918 (Pꢀgp inhibitor)
app
ꢀ6
and in the B→A direction in absence of GF120918 (Papp A→B, 23 x 10 cm/s) was measured by
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 57
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
monitoring the appearance of the test compound on the opposite side of the membrane using a
specific LCꢀMS/MS method. The efflux ratios (B→A ꢀGF120918/A→B –GF120918) were
calculated to determine whether the test compound was a Pꢀgp substrate.
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
All in vivo studies were performed in AAALACꢀaccredited sites, and ethical approval by the
corresponding ethical committee was obtained.
Chemistry. Reagents and solvents were purchased from commercial sources and used without
purification. All final products were > 95% pure. NMR spectra were recorded on a Bruker
1
13
Avance 400 spectrometer, operating at 400 MHz for H NMR and 101 MHz for C NMR or a
1
13
Bruker 360 spectrometer, operating at 360 MHz for H NMR and 91 MHz for C NMR.
Chemical shifts and multiplicity data are given according to the ACS NMR guidelines. High
®
®
resolution mass spectrometry was performed on a Waters Acquity IClass UPLC ꢀDAD and
Xevo G2ꢀS QTOF. The samples were run on a Waters BEH C18 (1.7 µm, 2.1 x 50 mm, at 50ºC)
column, with a flow rate of 1 mL/min, using reverse phase chromatography with a gradient from
9
5% A to 5% A in 4.6 min, and held for 0.4 min (A: 95% CH COONH 6.5 mM + 5% CH CN,
3 4 3
B: CH CN).
3
5-benzyl-2,4-dichloro-5H-pyrrolo[3,2-d]pyrimidine (5). Into a 50 mL vial was placed 2,4ꢀ
dichloroꢀ5Hꢀpyrrolo[3,2ꢀd]pyrimidine (1 g, 5.32 mmol), DMF (10 mL), DIPEA (2.75 mL, 16
mmol) and benzylbromide (0.7 mL, 5.85 mmol). The vial was sealed and shaken for 16 hours at
rt. The reaction mixture was filtered through silica, and the solvent of the filtrate was removed
under reduced pressure. The crude was used without further purification in the next step.
5-benzyl-N-butyl-2-chloro-5H-pyrrolo[3,2-d]pyrimidin-4-amine(6). Into a 50 mL round
bottom flask equipped with a magnetic stir bar was placed 5ꢀbenzylꢀ2,4ꢀdichloroꢀ5Hꢀ
ACS Paragon Plus Environment

Page 25 of 57
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
pyrrolo[3,2ꢀd]pyrimidine (1.4 g, 5.03 mmol), nꢀbutylamine (0.59 mL, 6.04 mmol), and 1,4ꢀ
dioxane (5 mL). The flask was equipped with a reflux condenser and allowed to heat with
stirring at 100ºC for 16 h. After cooling to rt, the solvents were removed under reduced pressure.
The crude was purified via silica gel column chromatography using a heptane to EtOAc gradient.
The best fractions were pooled and the solvents were removed under reduced pressure to afford
the titled compound.
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-benzyl-4-(butylamino)-5H-pyrrolo[3,2-d]pyrimidin-2-ol (7). A suspension of 6 (70mg, 0.22
mmol) in acetic acid (1 mL) and conc. HCl (0.2 mL) was heated in sealed tube at 100ºC for 24
hours. The solvent was removed under reduced pressure and the crude was purified via silica gel
1
column chromatography using a gradient of CH Cl to CH Cl /CH OH (95:5). H NMR (300
2
2
2
2
3
MHz, DMSOꢀd ) δ 10.33 (s, 1H), 7.16ꢀ7.35 (m, 4H), 6.98 (d, J=6.74 Hz, 2H), 6.19 (t, J=5.43
6
Hz, 1H), 5.85 (d, J=2.75 Hz, 1H), 5.52 (s, 2H), 3.22ꢀ3.30 (m, 2H), 1.23ꢀ1.37 (m, 2H), 0.91ꢀ1.06
(m, 2H), 0.73ꢀ0.80 (m, 3H).
2-azido-5-benzyl-N-butyl-5H-pyrrolo[3,2-d]pyrimidin-4-amine hydrochloride(8). Into a
glass vial equipped with a magnetic stir bar was placed 5ꢀbenzylꢀNꢀbutylꢀ2ꢀchloroꢀ5Hꢀ
pyrrolo[3,2ꢀd]pyrimidinꢀ4ꢀamine (1 g, 3.18 mmol), NaN (0.62 g, 9.53 mmol), and NMP: water
3
(9:1, 4 mL). The glass vial was sealed and the mixture was heated with stirring to 170ºC for 5 h.
After cooling to rt, the mixture was diluted with EtOAc (20 mL) and washed with water (5 x 15
mL). The organic layer was dried over magnesium sulfate, the solids were removed via filtration
and the solvents of the filtrate were removed under reduced pressure. The crude was purified via
silica gel column chromatography using a heptane to EtOAc gradient. The best fractions were
combined and the solvents were removed under reduced pressure. The product was suspended in
0.5 mL dioxane and to it was added 1 eq. 4N HCl in dioxane at 0ºC . The mixture was stirred for
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 57
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
6 hours at room temperature and then the solvent was removed under reduced pressure to afford
1
the titled compound. LCꢀMS (M+H) m/z = 322. H NMR (300 MHz, CHLOROFORMꢀd) δ 7.32ꢀ
7
.42 (m, 3H), 7.18ꢀ7.22 (m, 2H), 6.99ꢀ7.08 (m, 2H), 6.90 (d, J=3.16 Hz, 1H), 5.51 (s, 2H), 4.95
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
(br s, 1H), 3.37 (dt, J=5.22, 6.94 Hz, 2H), 1.13ꢀ1.26 (m, 2H), 0.98 (qd, J=7.31, 14.90 Hz, 2H),
0.69ꢀ0.78 (m, 3H).
4
5
-benzyl-N -butyl-5H-pyrrolo[3,2-d]pyrimidine-2,4-diamine (9). Into a glass vial equipped
with a magnetic stir bar was placed 2ꢀazidoꢀ5ꢀbenzylꢀNꢀbutylꢀ5Hꢀpyrrolo[3,2ꢀd]pyrimidinꢀ4ꢀ
amine (100 mg, 0.311 mmol), 1,4ꢀdioxane (4 mL), water (1 mL), and triphenylphosphine (245
mg, 0.93 mmol). The vial was sealed and the mixture heated with stirring to 120ºC for 16 h.
After cooling to rt, the solvents were removed under reduced pressure. The crude was purified
via silica gel column chromatography using a CH Cl to 10% CH OH in CH Cl gradient. The
2
2
3
2
2
best fractions were pooled and the solvents were removed under reduced pressure to afford the
1
titled compound. H NMR (400 MHz, DMSOꢀd ) δ 7.25ꢀ7.33 (m, 3H), 7.19ꢀ7.25 (m, 1H), 6.95ꢀ
6
7.01 (m, 2H), 5.98 (d, J=2.86 Hz, 1H), 5.73 (t, J=5.50 Hz, 1H), 5.49 (s, 2H), 5.23 (s, 2H), 3.25ꢀ
13
3
.32 (m, 2H), 1.28ꢀ1.39 (m, 2H), 1.04 (qd, J=7.39, 15.02 Hz, 2H), 0.73ꢀ0.82 (m, 3H). C NMR
(101 MHz, DMSOꢀd ) δ 158.8, 151.7, 150.8, 139.4, 133.2, 129.0, 127.8, 126.5, 110.4, 99.7, 52.0,
6
+
39.8, 31.5, 19.8, 14.2. ESIꢀHRMS (TOF) m/z: 296.1878 [M+H] (Calcd. for C H N :
1
7
21
5
2
95.1797).
Methyl
(4-hydroxy-5H-pyrrolo[3,2-d]pyrimidin-2-yl)carbamate
(10).
Aminoꢀ2ꢀ
ethoxycarbonylpyrrole hydrochloride (25.8 g, 135.7 mmol) was partitioned between CH Cl and
2
2
sat., aq. NaHCO , the organic layer was dried over MgSO , the solids were removed by
3
4
filtration, and the filtrate evaporated to dryness. The residue was dissolved in CH OH (500 mL)
3
along with 1,3ꢀbis(methoxycarbonyl)ꢀ2ꢀmethylꢀ2ꢀthiopseudourea (32.1 g, 156 mmol) and acetic
ACS Paragon Plus Environment

Page 27 of 57
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
acid (39 mL, 619 mmol) and stirred at rt for 1 h. A precipitate appeared and stirring was
continued overnight. NaOCH (73.1 g, 1353 mmol) was added. An exotherm was observed and
3
the reaction mixture was stirred overnight. The reaction mixture was brought to pH 5 via
addition of acetic acid and the precipitate was isolated by filtration, triturated with water (2 x 350
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
mL), CH CN (350 mL) and diisopropylether (350 mL).
3
Methyl (4-chloro-5H-pyrrolo[3,2-d]pyrimidin-2-yl)carbamate (11). Methyl Nꢀ(4ꢀhydroxyꢀ
5
Hꢀpyrrolo[3,2ꢀd]pyrimidinꢀ2ꢀyl)carbamate (25 g, 120 mmol) was dispensed into acetonitrile
(350 mL) in a 500 mL multi neck flask at rt. POCl (22.1 mL, 238.2 mmol) was added and the
3
reaction mixture was heated to 70ºC while stirring by an overhead, mechanical stirrer at 300 rpm.
N,Nꢀdiisopropylethylamine (41.4 mL, 240.2 mmol) was added dropꢀwise by a syringe pump at a
flow rate of 0.2 mL/min. The reaction mixture was cooled to rt and poured into a stirred solution
of sodium acetate (78.8 g, 961 mmol) in water (500 mL) at 45ºC. The organic solvents were
removed under reduced pressure and the remaining liquid was stirred and cooled in an ice bath.
The formed solid was isolated by filtration, washed with CH CN and triturated with
3
1
diisopropylether to afford the titled compound as an offꢀwhite solid. H NMR (400 MHz,
DMSOꢀd ) Shift 12.27 (br s, 1H), 10.22 (s, 1H), 7.92 (t, J=2.97 Hz, 1H), 6.57 (dd, J=1.76, 3.08
6
Hz, 1H), 3.66 (s, 3H).
Methyl (4-chloro-5-(pyridin-2-ylmethyl)-5H-pyrrolo[3,2-d]pyrimidin-2-yl)carbamate (12).
To a suspension of methyl (4ꢀchloroꢀ5Hꢀpyrrolo[3,2ꢀd]pyrimidinꢀ2ꢀyl)carbamate. (5 g, 22
mmol), 2ꢀpyridinemethanol (2.6 mL, 26.5 mmol) and polystyreneꢀbound triphenylphosphine
(18.4 g, 55.2 mmol) in anhydrous THF (153 mL) was added DIAD (6.9 mL, 33 mmol) at rt and
the reaction mixture stirred for 30 minutes, then was concentrated under reduced pressure. The
product was purified via silica gel column chromatography using a gradient of CH Cl :CH OH
2
2
3
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 57
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
00:0 to 90:10. The product fractions were collected and concentrated under reduced pressure.
The product was recrystallized in CH CN, isolated by filtration and dried under vacuum to afford
3
the titled compound as a white solid.
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
4-chloro-5-(pyridin-2-ylmethyl)-5H-pyrrolo[3,2-d]pyrimidin-2-amine (13). Methyl (4ꢀ
chloroꢀ5ꢀ(pyridinꢀ2ꢀylmethyl)ꢀ5Hꢀpyrrolo[3,2ꢀd]pyrimidinꢀ2ꢀyl)carbamate (4.5 g, 14.2 mmol)
was dissolved in 1,4ꢀdioxane (68 mL) in a 100 mL round bottom flask and NaOH (1 N aq., 34
mL) was added. The mixture was heated at 60°C for 5h. The mixture was cooled and
concentrated under reduced pressure. The crude was treated with water and the precipitate was
isolated by filtration and dried to afford the titled compound. The product was used without
further purification in the next step.
ACS Paragon Plus Environment

Page 29 of 57
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
4
1
N -butyl-5-(2-phenoxyethyl)-5H-pyrrolo[3,2-d]pyrimidine-2,4-diamine (14). H NMR (360
MHz, DMSOꢀd ) δ 7.20ꢀ7.28 (m, 3H), 6.92 (t, J=7.50 Hz, 1H), 6.82 (dd, J=0.73, 8.78 Hz, 2H),
6
6
.32 (t, J=5.31 Hz, 1H), 5.94 (d, J=2.93 Hz, 1H), 5.36 (br s, 2H), 4.62 (t, J=5.31 Hz, 2H), 4.16 (t,
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
J=5.12 Hz, 2H), 3.39ꢀ3.49 (m, 2H), 1.58 (quin, J=7.32 Hz, 2H), 1.35 (qd, J=7.38, 14.82 Hz, 2H),
1
3
0.89 (t, J=7.32 Hz, 3H). C NMR (91 MHz, DMSOꢀd ) δ 156.7, 156.6, 149.3, 148.9, 131.2,
6
128.3, 119.7, 113.1, 108.9, 97.6, 66.7, 46.5, 38.4, 30.1, 18.6, 12.6. ESIꢀHRMS (TOF) m/z:
+
3
26.1980 (Calcd. for C H N O [M+H] : 326.1980).
1
8
23
5
4
1
N -butyl-5-(3-phenylpropyl)-5H-pyrrolo[3,2-d]pyrimidine-2,4-diamine (15). H NMR (400
MHz, DMSOꢀd ) δ 7.22ꢀ7.30 (m, 2H), 7.08ꢀ7.22 (m, 4H), 6.34 (br t, J=5.50 Hz, 1H), 5.94 (d,
6
J=3.08 Hz, 1H), 5.59 (br s, 2H), 4.29 (t, J=7.04 Hz, 2H), 3.38ꢀ3.51 (m, 2H), 2.39ꢀ2.55 (m, 2H),
1.82ꢀ1.99 (m, 2H), 1.49ꢀ1.64 (m, 2H), 1.34 (qd, J=7.40, 14.99 Hz, 2H), 0.90 (t, J=7.37 Hz,
1
3
3
H). C NMR (101 MHz, DMSOꢀd ) δ 157.6, 151.1, 148.8, 141.6, 132.8, 128.8, 128.5, 126.3,
6
109.9, 98.3, 48.7, 40.1, 33.8, 32.3, 31.8, 20.2, 14.3. ESIꢀHRMS (TOF) m/z: 324.2189 (Calcd. for
+
C H N [M+H] : 324.2188).
19
25
5
1
2-(2-amino-4-(butylamino)-5H-pyrrolo[3,2-d]pyrimidin-5-yl)-N-methylacetamide (16).
H
NMR (400 MHz, DMSOꢀd ) δ 8.32 (br d, J=4.40 Hz, 1H), 7.11 (d, J=3.08 Hz, 1H), 6.93 (t,
6
J=5.17 Hz, 1H), 5.94 (d, J=2.86 Hz, 1H), 5.38 (br s, 2H), 4.74 (s, 2H), 3.31ꢀ3.45 (m, 2H), 2.64
13
(
d, J=4.62 Hz, 3H), 1.49ꢀ1.64 (m, 2H), 1.29ꢀ1.44 (m, 2H), 0.92 (t, J=7.37 Hz, 3H). C NMR
(
101 MHz, DMSOꢀd ) δ 169.4, 158.5, 151.4, 150.1, 132.7, 111.4, 99.5, 51.9, 31.6, 26.1, 20.2,
6
+
14.3. ESIꢀHRMS (TOF) m/z: 277.1779 (Calcd. for C H N O [M+H] : 277.1777).
1
3
20
6
4
1
N -butyl-5-(3-methylbut-2-en-1-yl)-5H-pyrrolo[3,2-d]pyrimidine-2,4-diamine (17). H NMR
(
360 MHz, DMSOꢀd ) δ 7.14 (d, J=2.93 Hz, 1H), 5.98 (t, J=5.49 Hz, 1H), 5.89 (d, J=2.93 Hz,
6
ACS Paragon Plus Environment