The Identification and Pharmacological Characterization of 6-(tert-Butylsulfonyl)-N-(5-fluoro-1H-indazol-3-yl)quinolin-4-amine (GSK583), a Highly Potent and Selective Inhibitor of RIP2 Kinase

Article abstract of DOI:10.1021/acs.jmedchem.6b00211

RIP2 kinase is a central component of the innate immune system and enables downstream signaling following activation of the pattern recognition receptors NOD1 and NOD2, leading to the production of inflammatory cytokines. Recently, several inhibitors of R

Full text of DOI:10.1021/acs.jmedchem.6b00211

Subscriber access provided by SUNY DOWNSTATE
Article
The identification and pharmacological characterization of 6-
tert-butylsulfonyl)-N-(5-fluoro-1H-indazol-3-yl)quinolin-4-amine
GSK’583), a highly potent and selective inhibitor of RIP2 Kinase
(
(
Pamela A Haile, Bartholomew J. Votta, Robert W Marquis, Michael J. Bury, John F. Mehlmann, Robert R.
Singhaus, Adam K. Charnley, Ami S. Lakdawala, Maire A. Convery, David B. Lipshutz, Biva M. Desai, Barbara
Swift, Carol A. Capriotti, Scott B. Berger, Mukesh K. Mahajan, Michael A. Reilly, Elizabeth J. Rivera, Helen H.
Sun, Rakesh Nagilla, Allison M. Beal, Joshua N. Finger, Michael N. Cook, Bryan W. King, Michael T. Ouellette,
Rachel D Totoritis, Maria Pierdomenico, Anna Negroni, Laura Stronati, Salvatore Cucchiara, Bartlomiej
Ziolkowski, Anna Vossenkamper, Thomas T. MacDonald, Peter J. Gough, John J. Bertin, and Linda N. Casillas
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.6b00211 • Publication Date (Web): 25 Apr 2016
Downloaded from http://pubs.acs.org on April 26, 2016
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
MacDonald, Thomas; Blizard Institute, Queen Mary University of London
Gough, Peter; GlaxoSmithKline
Bertin, John; GlaxoSmithKline
Casillas, Linda; GSK, ImmunoInflammation
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 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
The identification and pharmacological
characterization of 6-(tert-butylsulfonyl)-N-(5-
fluoro-1H-indazol-3-yl)quinolin-4-amine
GSK’583), a highly potent and selective inhibitor
of RIP2 Kinase
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
‡
1‡
1
1†
Pamela A. Haile, Bartholomew J. Votta, Robert W. Marquis, Michael J. Bury, John F.
1
1
1
2
2
Mehlmann, Robert Singhaus, Jr., Adam K. Charnley, Ami S. Lakdawala, Maire A. Convery,
1
1
1
1
1§
David B. Lipshutz, Biva M. Desai, Barbara Swift, Carol A. Capriotti, Scott B. Berger,
1
1
1
1
1
2
Mukesh K. Majahan, Michael A. Reilly, Elizabeth J. Rivera, Helen H. Sun, Rakesh Nagilla,
1
1
1
2
Allison M. Beal, Joshua N. Finger, Michael N. Cook, Bryan W. King, Michael T. Ouellette,
2
3
3
4
Rachel D. Totoritis, Maria Pierdomenico, Anna Negroni, Laura Stronati, Salvatore
5
6
7
7
Cucchiara, Bartłomiej Ziółkowski, Anna Vossenkämper, Thomas T. MacDonald, Peter J.
1
§
1
1*
Gough, John Bertin, Linda N. Casillas
1
Pattern Recognition Receptor Discovery Performance Unit, Immuno-inflammation Therapy
2
Area Unit; Platform Technology and Science, GlaxoSmithKline, Collegeville, PA, USA &
ACS Paragon Plus Environment

Page 3 of 57
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
3
Stevenage, UK; ENEA, Italian National Agency for new Technologies, Energy and Sustainable
4
Economic Development, Rome, Italy; Department of Cellular Biotechnology and Hematology,
5
Sapienza University Hospital Umberto I, Rome, Italy; Department of Pediatrics, Pediatric
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
Gastroenterology and Liver Unit, Sapienza University Hospital Umberto I, Rome, Italy; Royal
7
London Hospital, Endoscopy Unit, Barts Health, London, UK. Centre for Immunobiology,
Blizard Institute, Barts and The London School of Medicine and Dentistry, Queen Mary
University of London, London, UK
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
Abstract
RIP2 kinase is a central component of the innate immune system and enables downstream
signaling following activation of the pattern recognition receptors NOD1 and NOD2 leading to
the production of inflammatory cytokines. Recently, several inhibitors of RIP2 kinase have been
disclosed that have contributed to the fundamental understanding of the role of RIP2 in this
pathway. However, since they lack either broad kinase selectivity or strong affinity for RIP2,
these tools have only limited utility to assess the role of RIP2 in complex environments. We
present, herein, the discovery and pharmacological characterization of GSK‘583, a next
generation RIP2 inhibitor possessing exquisite selectivity and potency. Having demonstrated the
pharmacological precision of this tool compound, we report its use in elucidating the role of
RIP2 kinase in a variety of in vitro, in vivo, and ex vivo experiments, further clarifying our
understanding of the role of RIP2 in NOD1 and NOD2 mediated disease pathogenesis.
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
The NOD1 and NOD2 (nucleotide-binding oligomerization domain-containing proteins 1 and 2)
pathways have generated much interest in recent years. Through human and murine genetic
evidence, as well as varied in vitro and in vivo investigations, this pathway has been implicated
1,2
in a wide variety of inflammatory disorders, including inflammatory bowel disease (IBD),
3
,4
5,6
7
8,9,10
Blau Syndrome, rheumatoid arthritis, sarcoidosis, and asthma.
NOD1 and NOD2 are
well characterized cytoplasmic NLR’s (nucleotide-binding domain leucine-rich repeat
1
1,12
family)
that have been shown to play essential roles in innate immunity by recognizing
1
3,14,15,16
peptidoglycan fragments from both gram negative and gram positive bacteria.
Activation
of either NOD1 by diaminopimelic acid (meso-DAP) or NOD2 by muramyldipeptide (MDP)
initiates a signaling cascade culminating in synthesis of pro-inflammatory cytokines and
ACS Paragon Plus Environment

Page 5 of 57
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
chemokines. Under healthy physiological conditions, this pathway functions as a sentinel by
recognizing and responding to invasive bacteria. Dysregulation of this signaling pathway,
1
7
however, can lead to pathological levels of inflammatory cytokines, and ultimately disease.
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
RIP2 kinase (also known as RIPK2, RICK, CARDIAK, and CARD3) is a multi-domain, dual
specificity kinase that is expressed ubiquitously, and is the convergence point for NOD1 and
1
8,19,20,21,22
NOD2 signaling.
The C-terminal CARD domain (Caspase Activation and Recruitment
Domain) of RIP2 is required for association with both NOD1 and NOD2, and functional activity
of the N-terminal kinase domain is required for downstream signal transduction. The kinase
domain of RIP2 is known to undergo auto-phosphorylation and subsequent ubiquitinylation, thus
enabling downstream signaling through activation of nuclear factor kappa-light-chain-enhancer
of activated B cells (NF-kB) and mitogen-activated protein (MAP) kinases, resulting in the
23,24,25,26
transcription of multiple inflammatory cytokine genes.
In Blau syndrome patients, a
gain-of-function mutation in NOD2 results in over-activation of this pathway, evidenced by a
3
high level of phospho-RIP2.
Genetically modified mice lacking the RIP2 protein are
unresponsive to either NOD1 or NOD2-activating ligands, highlighting the essential role of RIP2
2
7
kinase in NOD1 and NOD2 signaling. Additionally, it has been shown that the abundance of
activated, phosphorylated RIP2 kinase is increased in intestinal biopsies obtained from pediatric
2
8,29,30
Crohn’s and ulcerative colitis patients.
Thus, pharmacologically targeting the NOD/RIP2
signaling pathway (or “nodosome”) via direct inhibition of RIP2 is expected to have broad
therapeutic utility for auto-inflammatory diseases.
In this paper we detail our efforts to identify and characterize a potent and highly selective
inhibitor of RIP2 kinase, with the intent to further elucidate the biological effects of aberrant
NOD1 and NOD2 signaling. Ultimately, we aim to discover a safe and effective inhibitor of
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
RIP2 kinase for the treatment of chronic inflammatory diseases. Both endeavors require a
rigorous commitment to achieving broad kinase selectivity. Until recently, kinase inhibitors
have not been utilized for chronic non-oncology indications due to off-target toxicities associated
with a lack of selectivity of many of these compounds. Several investigational and research
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
3
1
compounds have been reported in the literature to have activity against RIP2 kinase; however,
those with the highest affinity for RIP2 (e.g. Ponatinib; 4-chloro-7,10-dioxa-13,17,18,21-
2
,6 17,20
tetraazatetracyclo[12.5.2.1 .0
]docosa-1(20),2(22),3,5,14(21),15,18-heptaene (OD36); and,
3
2,33,34
SB-203580, 2)
suffer from poor overall kinome selectivities that could readily confound
interpretation of their pharmacological effects in experimental systems. Other inhibitors such as
Gefitinib and N-(2-(4-amino-3-(p-tolyl)-1H-pyrazolo[3,4-d]pyrimidin-1-yl)-2-
methylpropyl)isonicotinamide (WEHI-345), although more selective across the kinome, have
somewhat reduced activity in complex systems such as bone marrow-derived macrophage
3
0,35
(BMDM) cultures and in vivo.
To attain the highest quality molecule, we pursued a screening
and triage strategy which allowed us to rapidly identify a broadly selective hit chemotype, that
quickly evolved into the discovery of a precision tool molecule. In this report, we detail the
identification and characterization of a highly differentiated RIP2 inhibitor tool compound
distinguished from currently available tools by its combination of exquisite selectivity, robust
potency, and demonstrated efficacy in both rodent and human models of inflammatory disease.
Screening and Hit Identification
We employed a broad screening strategy attempting to identify chemical diversity in a rapid
fashion. This included screening the GSK kinase inhibitor set against the full RIP2 protein
utilizing a fluorescence polarization (FP) based binding assay, as well as screening the kinase
domain truncate protein (residues 1-310) in an affinity screen against our DNA Encoded Library
ACS Paragon Plus Environment

Page 7 of 57
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
36
(
ELT) collection. Not unexpectedly, a variety of potent, ATP-competitive, Type I and Type II
kinase inhibitor chemotypes were identified from this screening effort. Compounds 1, 2 and 3
illustrate screening hits from 3 different chemotypes and 2 different binding modes (Figure 1).
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
37
Compound 1, is an example of a type II inhibitor that targets the DFG-out inactive
conformation of the kinase. It had a modest binding affinity (pIC = 6.5) but relatively poor
5
0
binding and ligand efficiencies typical of this class of inhibitor (BEI = 13.9, LipE = 1.2). The
3
4
type I inhibitor 2, which targets the active DFG-in conformation, had a lower MW and clogP
3
8,39
and improved binding affinity compared to 1, which led to superior BEI and LipE values.
Compound 3, also a Type I inhibitor, had the best balance of binding affinity (pIC = 8.5) and
5
0
properties, with BEI, and LipE far exceeding the values for the other hits. By extension, the
physicochemical properties of 3, namely, the low molecular weight (MW = 328) and lipophilicity
(
clogP = 3.1), presented an excellent starting point for further optimization. Interestingly,
although screening hits 2 and 3 were identified specifically from the FP binding screen, the ELT
affinity screen also identified compounds related to these chemotypes.
Figure 1. Representative hits from the high throughput screen with binding affinity and
efficiency measures. BEI = (pIC /MW)x1000; LipE = pIC – clogP.
50
50
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
To drive toward selection of a highly selective biological tool, a strategy to rapidly triage the
screening hits by a cell-based RIP2-dependent selectivity assay was devised. To do this we
evaluated inflammatory chemokine IL-8 secretion from HEK cells by two distinct but parallel
cellular pathways. In a direct measure of RIP2-dependent cellular function, IL-8 production was
measured in HEK cells over expressing NOD2 following stimulation with the NOD2-specific
ligand MDP. In contrast, stimulation by the TLR2-specific ligand, Pam CSK , in HEK cells over
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
4
expressing TLR2 resulted in IL-8 production by a RIP2-independent pathway. Evaluation of
multiple compounds from each chemotype represented by compounds 1, 2 and 3 in the 2
orthogonal cell assays resulted in clear differentiation in their cell-based selectivities.
Specifically, compound 3 exhibited an IC of 20 nM in the MDP stimulated assay and an IC of
5
0
50
>
5000 nM in the Pam CSK -stimulated TLR2-dependent assay, translating to >500-fold
2 4
difference in activity between RIP2-dependent and RIP2-independent cellular pathways (Table
). In fact, the 4-aminoquinoline chemotype emerged as the most compelling series by virtue of
1
its superior selectivity in these cellular assays in comparison to all other chemotypes identified in
the screen (Figure 2).
Table 1. Activity of representative hits across biochemical and cell assays
Cmpd
RIP2 FP
MDP/NOD2 IL-8 Pam CSK / TLR2 IL-8 Fold difference
2
4
IC (nM)
IC (nM)
IC (nM)
in cell assays
5
0
50
50
1
2
3
790
250
3
320
250
20
2000
6310
6.2x
25x
>10000
>500x
ACS Paragon Plus Environment

Page 9 of 57
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
8
7.5
7
Non 4AQ
AQ
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
6
5
4
.5
6
.5
5
.5
4
4
6
8
10
NOD2(HEK) MDP-IL8 pIC50
Figure 2. Correlation of inhibition of IL-8 production (pIC ) in NOD2(HEK) cells stimulated
5
0
by MDP (x-axis) and in TRL2(HEK) cells stimulated by Pam CSK (y-axis) for the 4-
2
4
aminoquinoline (4AQ) series compared to other inhibitor series (Non 4AQ).
Encouraged by the excellent cell-based selectivity demonstrated by 3, the compound was further
evaluated against a panel of 337 kinases in a kinase activity assay at Reaction Biology
4
0
Corporation (RBC). At 1 micromolar concentration, compound 3 inhibited 38 kinases by ≥90%
and 17 kinases by 70 – 90% (Supporting Information, Table S2). We presumed that the lack of
broader kinase selectivity of 3 was driven by the promiscuous binding nature of the 3-amino-4-
methyl-phenol back-pocket moiety, and as such, this emerged as the first area of exploration for
optimization of this template.
Chemistry
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
The synthesis of 4-aminoquinoline analogs including hit compound 3 was conducted as
4
1
described in the literature (Scheme 1). Condensation of 4-(methylthio)aniline with diethyl 2-
(ethoxymethylene)malonate by cyclization in inert solvent at high temperature provided the
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
quinoline 5. Decarboxylation in two steps afforded quinoline 6. Oxidation of the sulfide to the
sulfone utilizing potassium peroxymonosulfate followed by chlorination at the 4-position of the
quinoline with thionyl chloride provided the intermediate required for diversification with
multiple aromatic back-pocket groups. Heating of chloroquinoline 7 under microwave
conditions in the presence of the appropriate amines afforded compounds 3 and 8 - 10.
Scheme 1. Synthesis of analogs containing varied back-pocket moieties.
Reagents and conditions: (a) diethyl 2-(ethoxymethylene)malonate, 160 °C, 83%; (b) Ph O,
2
2
50 °C, 100%; (c) NaOH, EtOH/H O, 130 °C, 97%; (d) Ph O, 250 °C, 60%; (e) potassium
2 2
peroxymonosulfate, MeOH/ H O; (f) SOCl , DMF, 100 °C, 91% over two steps; (g) EtOH, 150
2
2
°C, microwave
with 3-amino-4-methylphenol, 1H-benzo[d][1,2,3]triazol-5-amine,
ACS Paragon Plus Environment

Page 11 of 57
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
benzo[d]thiazol-5-amine or 5-fluoro-1H-indazol-3-amine [3 (39%), 8 (64%), 9 (67%), 10
(14%)].
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
4
-Chloro-quinolines with sulfone alternatives at the C6 position such as hydrogen,
trifluoromethyl, and methoxy were purchased from commercial sources (Scheme 2). Addition of
3
6
-amino-4-methylphenol under microwave conditions provided analogs 15, 19 and 20. From the
-iodoquinoine 16, the meta-pyridyl quinoline 17 was formed via a Suzuki reaction while the
tetrahydropyranylmethylamide 18 was installed by a one pot carbonylation–amidation
42
sequence.
Scheme 2. Synthesis of analogs containing sulfone alternatives.
Reagents and conditions: (a) 3-amino-4-methylphenol, EtOH, 150 °C, microwave [15 (55%),
1
6 (92%), 19, 20 (15%)]; (b) 3-(1,3,2-dioxaborinan-2-yl)pyridine, Pd(dppf[Cl ]), K CO , 150
2 2 3
°
C, microwave, 41%; (c) (tetrahydro-2H-pyran-4-yl)methanamine; DBU, Mo(CO) , t-Bu PBF ,
6 3 4
trans-di(µ-acetato)bis[0-(di-o-tolyl-phosphino)benzyl]dipalladium(II), THF, 150 °C, microwave,
1%.
1
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
Diversity on the sulfone substituent was facilitated by utilizing a C6 halogenated quinoline core
as in Scheme 3. Condensation of 4-iodoaniline with Meldrum’s acid afforded the malonic acid
ester 22. Cyclization to the 4-hydroxy-6-iodoquinoline under heated conditions and subsequent
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
treatment with POCl provided the chlorinated quinoline core 12 setting the stage for eventual
3
introduction of the aniline moiety.
Alternatively, commercially available 6-bromo-4-
chloroquinoline was used as in 25. Palladium catalyzed coupling at C6 of either iodide 12 or
bromide 25 was used to install the desired thioether. Oxidation to sulfones 23 and 26 was
followed by introduction of the back-pocket moiety as in Scheme 1 leading to 24 and 27.
Scheme 3. Synthesis of analogs containing varied alkyl sulfone moieties.
Reagents and conditions: (a) triethoxymethane, Meldrum’s acid, 90 °C to 70 °C, 75%; (b)
Ph O, 240 °C, 82%; (c) POCl , 92%; (d) i-PrSH, Pd(PPh ) , KOt-Bu, toluene, 50 °C, 45%; (e)
2
3
3 4
potassium peroxymonosulfate, THF/ H O, 80%; (f) EtOH with benzo[d]thiazol-5-amine or 5-
2
fluoro-1H-indazol-3-amine, 47%; (g) i-PrSH, Pd(PPh ) , Et N, acetonitrile, 90° C, 91%; (h)
3
4
3
ACS Paragon Plus Environment

Page 13 of 57
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
potassium peroxymonosulfate, THF/ H O, 41%; (i) HCl, EtOH, 5-fluoro-1H-indazol-3-amine,
2
9
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
Results and Discussion
Following identification of 4-aminoquinoline 3, we established unambiguous biophysical
evidence that compound 3 binds within the ATP binding site of the kinase domain of RIP2. We
recently published the first X-ray crystal structure of apo-RIP2 at 2.4Å resolution and the co-
crystal structures of several RIP2/inhibitor complexes with different binding motifs within the
43
ATP pocket of RIP2.
Utilizing this in-house RIP2 crystallization methodology, we
subsequently solved the co-crystal structure of aminoquinoline 3 bound to the kinase domain of
RIP2 (Figure 3 and Supporting Information Table S1). The co-crystal structure confirms that 3
binds within the ATP binding pocket of RIP2 in a traditional Type I DFG-in inhibitor binding
motif with a single hinge interaction between the backbone N-H of Met98 and the quinoline
nitrogen of 3. The glycine-rich loop of RIP2 possesses a polar Ser25 residue, which is rare
among protein kinases with only three other kinases having a serine at this position.
Interestingly, the crystal structure revealed a 2.7Å hydrogen bonding interaction between the
sulfone oxygen of 3 and Ser25. The methyl group of the C6 methyl sulfone projects into a small
pocket on the floor of the active site, and is also solvent accessible suggesting an attractive point
for inhibitor optimization. The quinoline core of 3 occupies the adenine binding pocket and
projects the C4 aniline around the Thr95 gatekeeper and into the hydrophobic back-pocket of
RIP2. The phenol group forms hydrogen bonds with 2 of the 3 conserved catalytic residues,
Glu66 and Asp164, and the methyl group is oriented toward a hydrophobic patch in the ceiling of
the back-pocket of RIP2. Unlike the Ser25 interaction, these interactions within the back-pocket
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
are not predicted to be unique to RIP2 and are likely the basis for the broad kinase inhibition
seen with 3.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 3. Co-Crystal structure of 4-anilinoquinoline 3 bound within the ATP binding pocket of
RIP2 kinase (2.7Å).
By leveraging our diverse collection of co-crystal structures of ligand-bounded RIP2, we
postulated that the large and flexible hydrophobic back-pocket provided an opportunity to drive
optimization of inhibitor selectivity. Interpretation of the ELT affinity screening results from a 4-
substituted quinoline containing library of >400 anilines led to the identification of two amines,
5
-aminobenzothiazole and 5-aminobenzotriazole, that presented unique structural features when
compared to the SAR from the FP screen. Drawing on the observation of complementary SAR
between hits identified by both ELT and FP based screens, we incorporated these amines as part
of an array of >200 aniline backpocket replacements on the methylsulfone quinoline core using
the method described in Scheme 1. This combined approach led to the identification of several
ACS Paragon Plus Environment

Page 15 of 57
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
analogs with dramatically improved kinase selectivity over the screening lead 3. Exemplars 8
(5-aminobenzotriazole), 9 (5-aminobenzothiazole) and 10 (3-amino-5-fluoroindazole) stood out
in terms of activity and selectivity as shown in Table 2. Kinase targets p38α and VEGFR2 were
chosen as surrogate kinases because they represent different branches of the kinome and cross
reactivity between RIP2 with p38α and VEGFR2 has been observed with earlier inhibitors.
Inhibition data for these kinases provided surrogate kinase selectivity data in a high through-put
fashion, thus providing a filter preceding the more extensive kinome profiling applied to
advanced analogs. Clearly, selectivity increased dramatically for compounds 8, 9, and 10
relative to p38α and VEGFR2 kinases in comparison to 3. While it is unclear precisely what is
driving the selectivity for RIP2 over VEGFR2 and p38α, we have surmised that subtle
electrostatic potential differences in the back-pocket region of each protein may play a role, in
conjunction with the higher degree of flexibility in the back-pocket of RIP2. This hypothesis is
based upon SAR observations of selectivity differences between back-pocket groups and
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
4
4,45
correlative electrostatic potential calculations using ab initio DFT/B3LYP/6-31G* methods.
While all three backpocket groups exhibited desirable selectivity profiles combined with potent
RIP2 inhibition and favorable LipE, benzotriazole 8 was not progressed further due to poor oral
rat PK (data not shown). Representative inhibitors containing both benzothiazole 9 and indazole
1
0 backpocket groups were evaluated in a kinase panel screen against >280 kinases at Reaction
Biology Corporation (RBC). Selectivity of the 3-aminoidazole was generally superior to that of
the benzothiazole. Despite the slightly lower potency and ligand efficiency, the 3-aminoindazole
was chosen as the preferred back-pocket substitution based on the kinase selectivity profile.
Table 2. Back-pocket analogs
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
RIP2 FP
p38α
VEGFR2
Cmpd
R
LipE
5.5
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
IC (nM)
IC (nM) IC (nM)
5
0
50
50
3
3
502
186
8
9
50
4.7
4.7
>5000
>5000
>5000
2385
6.3
10
50
4.0
>5000
>5000
Next we set out to determine whether sterics or electostatic interactions were driving other
aspects of the protein/ligand interaction for this chemical series. To assess the importance of the
sulfone / Ser25 interaction observed in the co-crystal structure, a set of C6-substituted
compounds with varying abilities to hydrogen bond with serine were tested against RIP2 as well
as p38α and VEGFR2 (Table 3). Because we expected many of these substitutions to have
varying electronic effects on the quinoline ring system, thus affecting the N1 interaction with the
NH of Met98 at the hinge of the kinase, analogs were also designed to determine the effect of the
electron donating and electron withdrawing groups on RIP2 activity as a measure of hinge
binding affinity. As a baseline comparison, the C6 unsubstituted quinoline 15 was prepared,
which should have no interaction with Ser25 and a stronger N1 – Met98(NH) interaction relative
to sulfone 3. A >200 fold shift in potency was observed between the hit sulfone 3 and the
ACS Paragon Plus Environment

Page 17 of 57
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
unsubstituted quinoline 15 suggesting that the Ser25 electrostatic interaction was indeed
important for binding affinity. The pyridine-containing quinoline 17 and amide 18 both retained
activities of 10 and 20 nM, respectively at RIP2, as expected, based on the predicted interactions
of these substituents with Ser25 from modelling exercises. Methoxyquinoline 19 and
trifluoromethylquinoline 20 were designed as analogs that would have negligible interactions
with Ser25, and so would allow an isolated evaluation of the electron withdrawing and donating
substituent effect upon hinge binding. It is generally accepted that increasing the H-bond
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
4
6
strength at the hinge binding interaction should lead to higher affinity inhibitors. Indeed, ab
47
initio calculations showed that the quinoline N1 hydrogen bonding strength was predicted to
increase with increasing electron donating abilities of the C6 substituents [OMe (19) > H (15) >
CF (20) > SO Me (3)]. Thus, it was anticipated that 19 would retain RIP2 activity despite no
3
2
Ser25 interaction due to the increased hinge binding strength, but that analog 20 would lose
activity. It was therefore surprising that 20 was very active at RIP2 (IC = 8 nM) despite the
50
predicted weaker hinge binding affinity. Interestingly, while fluorine generally forms weaker
hydrogen bonds than oxygen (2.4 kcal/mol vs. 5-10 kcal/mol), they have been reported in the
4
8,49
literature.
Modelling of trifluoromethylquinoline 20 in the RIP2 active site revealed a very
close interaction between the fluorine and the Ser25 hydroxyl of RIP2 with an F-H distance of
1
.7Å. There are very few documented examples in which a fluorine and hydroxyl are in such
5
0,51
close proximity suggesting that this fluorine hydrogen bond is particularly strong.
We
believe that this putative H-F bonding interaction is compensating for the reduced hinge binding
affinity for this compound. In addition, most of these C6 substitutions had a beneficial impact on
kinase selectivity compared to the methyl sulfone as indicated by the activity at the surrogate
kinases p38α and VEGFR2. In the end, the methyl sulfone was chosen as the preferred C6
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
substitution because it provided the best balance of potency, physicochemical properties,
permeability, and lipophilic ligand efficiency (LipE).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Table 3. Changes at C6 to Assess the Importance of the Ser25 Interaction
RIP2 FP
p38α
VEGFR2
Permeability
(nm/sec)
Compound
R
LipE
5.5
1.9
IC (nM)
IC (nM)
50
IC (nM)
50
5
0
3
3
500
>5000
3740
190
>5000
460
79
15
H
630
250
230
1
1
7
8
1
2
0
0
3.5
3.8
>5000
>5000
15
1
2
9
0
CH
3
O
250
8
>5000
1080
3000
100
230
490
2.0
3.0
CF
3
Combining the highly selective 3-aminoindazole back-pocket binder with more sterically
3
demanding C6 sulfone substituents led to compounds 24 (GSK’214) and 27 (GSK’583) (Table
ACS Paragon Plus Environment

Page 19 of 57
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
4
). Although these changes did not provide any additional benefits in terms of kinase selectivity
or rat oral pharmacokinetics, increasing the bulk and lipophilicity at the sulfone functionality
translated to moderate improvements in the binding affinity while maintaining reasonable
lipophilic ligand efficiency (LipE). In addition to robust selectivity against p38α and VEGFR2,
27 exhibited excellent selectivity in a panel of 300 kinases at 1 µM compound concentration
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(Figure 4, and Supporting Information Table S3). In this panel, only RIP2 was inhibited >50%
by 27, and only two additional kinases were inhibited by approximately 30% (BRK, Aurora A).
This level of broad kinome selectivity is a significant improvement over most published kinase
21,31-34
inhibitors.
Inhibitor 27 possesses comparable binding affinity for RIP3 kinase (not
included in RBC panel) as demonstrated by an in-house FP binding assay configured similarly to
the RIP2 FP assay (RIP2 FP IC50 = 5 nM; RIP3 FP IC50 = 16 nM). Despite this potent
biochemical activity against RIP3 kinase, 27 showed little or no inhibition of RIP3-dependent
necroptotic cell death in a cellular assay up to 10 µM concentration (Supporting Information
Figure S3). The reason for this shift in potency between biochemical and cell-based formats is
not understood completely. However, we have experimentally determined that the KM,ATP for
RIP3 is 5 µM, a relatively low KM,ATP among kinases. By comparison, the KM,ATP for RIP2 is
1
00 µM (data not shown). Given the relationship between KM,ATP and IC from the Cheng-
50
5
2,53
Prusoff equation,
and the expectation that the intracellular concentration of ATP is in the 1 –
10 mM range, it is not surprising that for RIP3 in particular, there would be a large shift in
54
cellular potency.
Table 4. Small Modifications of the Methylsulfone
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
Rat Oral PK a
Cmax (ng/mL) /
AUC (ng*h/mL)
p38α
RIP2 FP
VEGFR2
Compound
Structure
LipE
IC (nM)
IC (nM)
50
IC (nM)
50
5
0
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
0
50
4.0
>5000
>5000
500 / 1700
2
2
4
7
1
0
3.9
3.8
>5000
5119
>5000
>5000
370 / 1100
260 / 980
5
a
Cassette data, 2 mg/kg oral dose; female Sprague-Dawley rat
Figure 4. Kinome plot for 27, 1 µM compound, RBC panel of 300 kinases.
Green = <50% inhibition (299 out of 300 kinases); RIP2 (yellow) inhibited by 60% at 1 µM
ACS Paragon Plus Environment

Page 21 of 57
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
The co-crystal structure of 27 bound to the human RIP2 catalytic domain (res 1-310) was solved
at 2.5 Å resolution (Figure 5 and Supporting Information Table S1). The observed binding mode
is consistent with ATP competitive inhibition of RIP2 by 27, and displays similar binding
interactions to that of the original hit 3. The C4 3-aminoindazole projects around the Thr 95
gatekeeper residue and into the back-pocket making hydrophobic interactions with the roof of
the pocket and accepting a hydrogen bond from Asp164. The C6 sulfone is making a hydrogen
bond contact with the hydroxyl side chain of Ser25. The co-crystal structure, however, provides
no additional rationale to explain the tremendous kinase selectivity of this compound. A small
molecule X-ray provided further confirmation of the structure (Supporting Information Figure
S4).
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 5. Co-Crystal structure of 27 bound within the ATP binding pocket of RIP2 kinase
(2.5Å).
Pharmacological Characterization
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
The potency of 27 was also evaluated in primary human monocytes isolated from the blood of
normal healthy volunteers. In this context, 27 potently and dose dependently inhibited MDP-
stimulated tumor necrosis factor-alpha (TNFα) production with an IC = 8 nM. Importantly, as
5
0
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
shown in Table 5, 27 demonstrated only a modest reduction in potency when profiled in a similar
MDP-induced TNFα production assay in human whole blood (IC = 237 nM) and rat whole
5
0
blood (IC = 133 nM), which is reasonable given the blood free fraction of 2.6% (see Table 6)
5
0
for 27.
Table 5. Assay data for 27 across multiple assays of increasing complexity
Assay IC (nM) (±) SD
n
24
4
5
0
RIP2 FP (human)
RIP2 FP (rat)
5
2
4
1
1
8
7
n.d.
n.d.
44
20
9
MDP-stimulated IL8 release (HEK/NOD2)
Pam CSK -stimulated IL8 release (HEK/TLR2)
>1000
8
2
4
Human monocytes, MDP-stimulated TNFα release*
Human whole blood, MDP-stimulated TNFα release
Rat whole blood, MDP-stimulated TNFα release
2
237
133
6
65
5
*
Individual values: 7 nM and 9 nM in two separate experiments.
The cellular specificity of 27-mediated inhibition of pro-inflammatory cytokine production in
human monocytes was further established using ligands specific to other innate immune
receptors constitutively expressed in monocytes, which are known to signal primarily via RIP2-
independent pathways. Following treatment with 27 at 1 µM, little inhibition of pro-
inflammatory signaling was observed upon activation of Toll-like receptors (TLR2, TLR4,
TLR7) or cytokine receptors (IL-1R, TNFR) (Figure 6). In contrast, complete inhibition was
ACS Paragon Plus Environment

Page 23 of 57
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
observed upon activation of both NOD1 and NOD2 receptors, which signal in a RIP2-dependent
manner. These data demonstrate that the potency and selectivity of 27 demonstrated in
biochemical assays translates to cellular systems.
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 6. Cellular selectivity of RIP2 kinase inhibitor 27 was evaluated in primary human
monocytes. In the presence of 1 µM 27, NOD1- or NOD2-mediated cytokine production was
completely inhibited. In contrast, only weak inhibition was observed upon stimulation with
ligands which selectively activate various TLR or cytokine receptors. The targeted receptors
were stimulated with the following ligands: NOD1 (300 µg/mL, ieDAP), NOD2 (1 µg/mL,
MDP); TLR2 (10 ng/mL Pam2Csk4), TLR4 (10 ng/mL, ultrapure LPS), TLR7 (10 µg/mL,
gardiquimod), IL-1R (10 ng/mL, IL-1β), and TNFR (100 ng/mL, TNFα). Release of pro-
inflammatory cytokines, either TNFα (NOD2, TLR2, TLR4, IL1R) or IL-8 (NOD1, TLR7,
TNFR) were measured by immunoassay. % control indicates the mean ± SD (N=3) of the
cytokine response in the absence of inhibitor.
Compound 27 was further profiled in discrete oral and i.v. crossover experiments in both rat and
mouse. The pharmacokinetic parameters are consistent across rodent species exemplifying low
clearance, moderate volumes of distribution, and moderate oral bioavailability (Table 6).
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
Unfortunately the PK/PD relationship incorporating rodent PK with the concentration-response
data from the human whole blood assay suggests 27 would not produce a human
phamacodynamic response within an acceptable dose range, thus precluding this molecule from
further development as a drug candidate. However, the oral PK in rat and mouse provides
sufficient systemic exposure for use as a preclinical in vivo tool molecule in an acute
inflammation challenge model.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Table 6. PK parameters for 27 in mouse and rat, discrete iv/po studies
Oral / IV
Dose
(mg/kg) (ng/mL)
Oral
Cmax
Cl
(mL/min
/kg)
Free
%F Fraction
(%)
Oral AUC
Vdss T ½
(L/kg) (h)
Species
Strain
(ng*h/mL)
Sprague-
Dawley
Rat
2.0 / 1.0
260
340
980
860
15
13
2.0
1.2
3.1
2.1
39
34
2.6
--
Mouse C57BL/6 2.0 / 1.0
To demonstrate the utility of 27 in this context, we employed a combination of models which
measured local and systemic inflammatory responses to intra-peritoneal injection of L18-MDP in
mice and intra-venous injection of L18-MDP in rats with oral doses of 27 of 0.1, 1, and 10
mg/kg. Compound 27 inhibited serum KC (the rodent ortholog of IL-8) levels in rats in a dose-
dependent manner with an IC derived from rat blood concentrations of 50 nM (or 20 ng/mL).
5
0
Similarly, 27 inhibited serum KC levels (data not shown) and recruitment of neutrophils into the
peritoneal cavity in mice in a dose-dependent manner with an IC of 37 nM (15 ng/mL) derived
5
0
from mouse blood concentrations (Figure 7). As expected, a single oral dose of 10 mg/kg of 27
also inhibited the increase in serum levels of KC following administration of the NOD1 ligand,
FK156 in mice, but had no effect on cytokine production in response to stimulation with the
ACS Paragon Plus Environment

Page 25 of 57
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
TLR4 ligand, LPS at the same dose (Figure 7). These data demonstrate that 27 is a potent and
selective RIP2 kinase inhibitor in vivo as well as in vitro.
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 7. 27 inhibited MDP-induced inflammatory responses in vivo. Oral administration of
2
7 (0.1, 1, 10 mg/kg) produced a dose-dependent inhibition of serum KC in rats following
intravenous challenge with L18-MDP (left panel) and inhibition of neutrophil levels from
peritoneal lavage in mice following intraperitoneal challenge with L18-MDP (center panel). The
RIP2 specificity in vivo was confirmed by demonstrating that an oral dose (10 mg/kg) of 27
inhibited the increase in serum KC levels in mice following an intravenous challenge with the
NOD1 ligand, FK156, but not with TLR4 ligand, LPS.
Recent publications have shown that inhibitors of RIP2 kinase are effective in alleviating
32,34,55
inflammation in animal models of experimental colitis.
As it is well established that these
animal models of colitis and inflammation have limited translatability, we decided to employ a
human translational model using intestinal mucosal tissue from IBD patients.
Human intestine mucosal biopsy experiments
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
Previous studies have provided evidence of increased activation of the NOD2/RIP2 signaling
pathway in intestinal biopsies taken from pediatric UC and CD patients with active disease,
1
6-18
including increased expression of the activated form of RIP2 (pS176).
We examined RIP2
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
expression and phosphorylation in biopsies from adult UC (n=14) and CD (n=14) patients
(Supporting Information Figure S5). Total levels of RIP2 protein were similar to those seen in
biopsies from normal healthy volunteers, but the proportion of phosphorylated RIP2 (anti-pS176)
was significantly increased in both CD and UC biopsies suggesting that the RIP2 pathway was
similarly activated in adult disease.
To explore the significance of RIP2 kinase pathway activation on inflammation in adult CD and
UC patient, we employed an ex vivo culture system to assess the effect of RIP2 kinase inhibition
on spontaneous pro-inflammatory cytokine release by intestinal explants. As previously
described, intestinal mucosal tissue obtained from inflamed areas of the gut spontaneously
release significantly elevated levels of inflammatory cytokines compared to non-diseased
5
6
mucosa, recapitulating the inflammatory phenotype found in vivo. In this human biopsy assay,
2
7 significantly inhibited both TNFα and IL-6 production in a concentration-dependent fashion
in both CD (n=12) and UC (n=5) samples, and this inhibition was comparable in magnitude to
that seen with the steroid prednisolone (Figure 8). The observed potency of 27 in these explant
cultures (IC ~ 200 nM) was comparable to the potency observed in MDP-stimulated human
5
0
whole blood cultures (IC = 237 nM). This agreement in apparent IC values combined with
5
0
50
the exquisite selectivity of 27 strongly supports the hypothesis that inhibition of this
inflammatory signal is primarily mediated through selective inhibition of RIP2 kinase activity.
ACS Paragon Plus Environment

Page 27 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 8. Spontaneous release of pro-inflammatory cytokines by intestinal mucosal biopsies
from Crohn’s Disease (left, n = 12 patients) and Ulcerative Colitis (right, n = 5 patients) patients
is inhibited by 27. Explants were cultured overnight in the presence of medium alone (Ctrl) or
prednisolone (1 µM) or the indicated concentration of 27. Cytokine concentration was assessed
by immunoassay. Bars represent the mean ± SEM.
Given the number of different PRRs that are likely activated by penetration of commensal
luminal bacteria following breakdown of the intestinal epithelial barrier in the context of diseases
such as UC and CD, the magnitude of inhibition provided by the RIP2 inhibitor in these models
is somewhat surprising. Recent data have suggested that there is significant synergy in the
57,58,59,60,61
inflammatory response following simultaneous stimulation of multiple PRRs.
Together, these data suggest that RIP2 kinase may be a key node in the signaling network
connecting multiple PRRs.
Conclusion
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
Herein, we have reported the identification and characterization of 27, an exceptional RIP2 tool
inhibitor. The potency, selectivity, and PK/PD properties of 27 make it an ideal tool for
elucidating the pharmacology of RIP2 inhibition in a variety of in vitro and in vivo settings.
Pursuing a screening strategy designed to identify potent, efficient, and selective inhibitors of
RIP2 kinase by first intent has set the stage for rapid identification of 27. By focusing an SAR
strategy around the co-optimization of RIP2 kinase potency and broad kinome selectivity, we
were able to develop an exquisitely selective pharmacological tool compound. Although 27 has
excellent kinase selectivity, it does inhibit both the hERG channel and Cyp3A4, which precludes
it from further progression as a drug candidate. Nonetheless, by employing a chemically-driven
biology approach to specifically target inhibition of RIP2 kinase activity, we have
unambiguously demonstrated the role for RIP2 kinase activity in modulating intestinal
inflammation in both murine and human systems. These new data suggest that RIP2 kinase is a
key regulator of the inflammatory response in intestinal models of inflammation, and adds to the
growing body of evidence supporting RIP2 kinase as an attractive therapeutic target for the
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
treatment of IBD.
Further, the exquisite selectivity of 27 lends high confidence to the
interpretation of these data and the conclusions herein. The added cellular, ex vivo and in vivo
characterization of potency and selectivity, and the robust evidence of consistent pharmacology
across species and assay systems, provides a comprehensive data package supporting the use of
27 in future RIP2 pharmacological investigations.
Experimental Section
ACS Paragon Plus Environment

Page 29 of 57
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
All animal studies were conducted in accordance with the GSK Policy on the Care, Welfare and
Treatment of Laboratory Animals and were reviewed by the the Institutional Animal Care and
Use Committee either at GSK or by the ethical review process at the institution where the work
was performed.
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 human biological samples were sourced ethically and their research use was in accord with
the terms of the informed consents.
Chemistry: Without specific statement, all solvents and reagents were available from Sigma-
Aldrich, VWR, or Fisher Scientific and used as supplied or purified by standard laboratory
1
methods as required. H NMR spectra were recorded at 600, 500, 400, or 250 MHz at 25 °C.
1
3
C NMR spectra were recorded at 151 or 126 MHz at 25 °C. Deuterated solvents are noted with
the NMR data. High resolution mass spectrometry were performed by either a time of flight
mass spectrometer or Fourier transform mass spectrometer using electrospray (ES) techniques.
Melting points were obtained on a digital OptiMelt system by Stanford Research Systems.
Purifications were accomplished via flash chromatography on ISCO or Biotage systems with
normal phase silica gel column or via reverse phase chromatography on a Gilson or Waters
preparatory HPLC using a TFA and acetonitrile mixture in water. Purity of tested compounds
was assessed to be at least 95% by UV and ELSD using LCMS was carried out using a Waters
o
Acquity fitted with Waters SQD mass spectrometer (column temp 45 C, UV = 210 – 350 nm,
MS = ESI with pos/neg switching) and a Waters BEH C18 (50 mm × 2.1 mm, 1.7 μm) at a flow
rate of 1 mL/min using a solvent system of 3% water with 0.1% formic acid to 100% CH CN
3
with 0.1% formic acid over 2.0 min with a 0.4 min hold. Purity of key compound 27 was further
assessed by elemental analysis.
ACS Paragon Plus Environment