Discovery of Tyrosine Kinase 2 (TYK2) Inhibitor (PF-06826647) for the Treatment of Autoimmune Diseases

Article abstract of DOI:10.1021/acs.jmedchem.0c00948

Tyrosine kinase 2 (TYK2) is a member of the JAK kinase family that regulates signal transduction downstream of receptors for the IL-23/IL-12 pathways and type I interferon family, where it pairs with JAK2 or JAK1, respectively. On the basis of human genetic and emerging clinical data, a selective TYK2 inhibitor provides an opportunity to treat autoimmune diseases delivering a potentially differentiated clinical profile compared to currently approved JAK inhibitors. The discovery of an ATP-competitive pyrazolopyrazinyl series of TYK2 inhibitors was accomplished through computational and structurally enabled design starting from a known kinase hinge binding motif. With understanding of PK/PD relationships, a target profile balancing TYK2 potency and selectivity over off-target JAK2 was established. Lead optimization involved modulating potency, selectivity, and ADME properties which led to the identification of the clinical candidate PF-06826647 (22).

Full text of DOI:10.1021/acs.jmedchem.0c00948

Subscriber access provided by CARLETON UNIVERSITY
Drug Annotation
Discovery of Tyrosine Kinase 2 (TYK2) Inhibitor
PF-06826647) for the Treatment of Autoimmune Diseases
(
Brian S Gerstenberger, Catherine M. Ambler, Eric P Arnold, Mary Ellen Banker, Matthew F. Brown,
James D Clark, Alpay Dermenci, Martin E. Dowty, Andrew Fensome, Susan Fish, Matthew M. Hayward,
Martin Hegen, Brett D Holingshead, John D. Knafels, David W Lin, Tsung H. Lin, Dafydd R Owen, Eddine
Saiah, Raman Sharma, Felix F. Vajdos, Li Xing, Xiaojing Helen Yang, Xin Yang, and Stephen W Wright
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.0c00948 • Publication Date (Web): 05 Aug 2020
Downloaded from pubs.acs.org on August 5, 2020
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Discovery of Tyrosine Kinase 2 (TYK2) Inhibitor (PF-06826647) for the Treatment of
Autoimmune Diseases
Brian S. Gerstenberger,* Catherine Ambler, Eric P. Arnold, Mary-Ellen Banker, Matthew F.
Brown, James D. Clark, Alpay Dermenci, Martin E. Dowty, Andrew Fensome, Susan Fish,
Matthew M. Hayward, Martin Hegen, Brett D. Hollingshead, John D. Knafels, David W. Lin,
Tsung H. Lin, Dafydd R. Owen, Eddine Saiah, Raman Sharma, Felix F. Vajdos, Li Xing, Xiaojing
Yang, Xin Yang, and Stephen W. Wright*
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
Abstract:
Tyrosine kinase 2 (TYK2) is a member of the JAK kinase family that regulates signal
transduction downstream of receptors for the IL-23/IL-12 pathways and Type I interferon family,
where it pairs with JAK2 or JAK1, respectively. Based on human genetic and emerging clinical
data, a selective TYK2 inhibitor provides an opportunity to treat autoimmune diseases delivering
a potentially differentiated clinical profile compared to currently approved JAK inhibitors. The
discovery of an ATP-competitive pyrazolopyrazinyl series of TYK2 inhibitors was accomplished
through computational and structurally-enabled design starting from a known kinase hinge
binding motif. By understanding PK/PD relationships, a target profile balancing TYK2 potency
and selectivity over off-target JAK2 was established. Lead optimization involved modulating
potency, selectivity, and ADME properties which led to the identification of the clinical candidate
PF-06826647 (22).
Introduction:
Intercellular cytokine signaling plays a critical role in immunological regulation and host
immune response in health and disease. Excessive cytokine signaling is a significant contributor
to the pathogenesis of autoimmune diseases. Targeting cytokine signaling pathways through
pharmacological inhibition or modulation has been shown to be a successful strategy for
treatment of several autoimmune and inflammatory diseases.1 Effective targeting of cytokine
signaling pathways via inhibition of Janus kinases (JAK) has been demonstrated as clinically useful
in rheumatoid arthritis (RA), psoriatic arthritis, and ulcerative colitis as well as for
myeloproliferative disorders and graft-versus-host disease (GvHD).2, 3 Tyrosine kinase 2 (TYK2) is
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 54
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
one of the four members of the JAK non-receptor tyrosine kinase family which also includes JAK1,
JAK2, and JAK3. The JAK kinases are important in both the innate and adaptive immune systems
through the modulation of cytokine signaling and their influence on the Signal Transducer and
Activation of Transcription (STAT) pathways.4, 5 TYK2 is associated with the receptors for
interleukin-12 (IL-12), IL-23 and type I interferons (IFNs) and regulates the downstream signal
transduction pathways.6,7 These cytokines are critical for the functions of T helper 1 (TH1), TH17,
B cells, and myeloid cells that play a key role in a number of autoimmune and chronic
inflammatory diseases. JAK kinases function as pairs of homo- or heterodimers in the JAK-STAT
pathways. TYK2 pairs with JAK1 to mediate multiple cytokine pathways including type I IFNs, IL-
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 and IL-22. TYK2 also pairs with JAK2 to transmit the signals of IL-12 and IL-23.
Genome-wide association studies have identified the TYK2 loci associated with several
autoimmune diseases including rheumatoid arthritis, psoriasis, systemic lupus erythematosus,
7
Crohn’s disease and ulcerative colitis. A TYK2 variant encoding a proline to alanine substitution
at amino acid 1104 is expressed but is inactive and has been shown to provide protection from
several autoimmune diseases.7, 8 This coding variant has diminished IL-12, IL-23 and type I
interferon (IFN) signaling, suggesting that function of TYK2 is critical to the pathogenesis of above
mentioned autoimmune diseases. Notably, humans homozygous for this inactive variant do not
have an increase in malignancy, bacterial, mycobacterial, fungal or viral infections suggesting that
inhibition should be tolerated. Targeting inhibition of IL-12 and IL-23 signaling has been
demonstrated in the clinic by ustekinumab (STELARA®), a monoclonal antibody against the p40
subunit common to IL-23 and IL-12, which is approved for the treatment of psoriasis (PsO),
psoriatic arthritis, Crohn’s disease and ulcerative colitis.9-11 Additionally, antibodies targeting the
IL-23-specific p19 subunit guselkumub (TREMFYA®), tidrakizumab (ILUMYA®) and risankizumab
(SKYRIZI®) have been approved for the treatment of adults with moderate to severe plaque
psoriasis.12, 13 The pharmacological profile of a small molecule selective TYK2 kinase inhibitor
provides an opportunity to treat several autoimmune diseases with an oral therapy.14-16
Currently there are four oral FDA approved JAK kinase inhibitors, tofacitinib (XELJANZ®)
(1), baricitinib (OLUMIANT®) (2), ruxolitinib (JAKAFI®) (3), and upadacitinib (RINVOQ®) (4) along
ACS Paragon Plus Environment

Page 3 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
with peficitinib (SMYRAF®) (5) approved in Japan.17 All approved JAK kinase inhibitors
demonstrate activity against JAK1 without high specificity for TYK2.18 These compounds bind to
the ATP site of the catalytically active Janus Homology 1 (JH1) domain. BMS-986165 (6)19, 20 is a
TYK2 inhibitor in phase 3 clinical trials for psoriasis that, in contrast to most other JAK inhibitors,
binds to the pseudokinase Janus Homology 2 regulatory domain (JH2). We have previously
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
disclosed the discovery of a JH1 TYK2/JAK1 dual inhibitor brepocitinib (PF-06700841) (7) which is
currently under clinical investigation for a number of autoimmune indications.21, 22
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 54
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
Figure 1: Approved JAK kinase inhibitors and disclosed clinical stage TYK2 kinase inhibitors
F
F
O
O
N
S
N
O
N
N
N
O
N
N
NC
O
N
N
NC
N
N
N
HN
F3C
HO
N
NH
N
O
NC
N
N
O
HN
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
N
N
N
O
NH2
D3C
N
N
N
N
N
N
O
H
N
N
N
N
H
N
N
N
H
H
N
N
N
N
H
H
N
H
H
5
peficitinib
6 BMS-986165
Phase III
7 PF-06700841
Phase II
1 tofacitinib
FDA Approved
2 baricitinib
FDA Approved
3 ruxolitinib
FDA Approved
4 upadacitinib
FDA Approved
Approved in Japn
An inhibitor of TYK2 with enhanced selectivity over JAK1, JAK2, and JAK3 could be
beneficial based on an orthogonal cytokine inhibition profile driving IL-23, IL-12 and IFN
signaling, while simultaneously sparing suppression of other JAK-dependent cytokines, thus
potentially avoiding broader immunosuppression. Sufficient selectivity over JAK2 would be
desirable to minimize inhibition of erythropoietin (EPO) and thrombopoietin (TPO) signaling,
which is mediated via a JAK2 homodimer. Decreased inhibition of EPO and TPO signaling should
minimize effects on the production of red blood cells and platelets that may contribute to anemia
and thrombocytopenia seen with some JAK inhibitors.23 In addition, selectivity over JAK1 and
JAK3 would limit inhibition of IL-6 and common gamma-chain family cytokines, including anti-
inflammatory cytokines IL-2, IL-15, and IL-21. The overall profile of a TYK2 inhibitor, particularly
with clear selectivity over JAK1, is hypothesized to offer the potential for differentiation over the
currently approved JAK family-based therapies, while blocking numerous proinflammatory
pathways implicated in psoriasis, inflammatory bowel disease (IBD) and other autoimmune
diseases.
This manuscript describes the discovery and pre-clinical profile of PF-06826647 (22), a
potent TYK2 inhibitor binding to the JH1 domain, that is under clinical development for the
treatment of psoriasis (PsO), ulcerative colitis (UC), and hidradenitis suppurativa (HS).24,25,26
Program Objectives:
The objectives for our TYK2 inhibitor program were based on the clinical and preclinical
PK:PD relationships previously observed and discussed for tofacitinib (1). Biochemical kinase
ACS Paragon Plus Environment

Page 5 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
assays for TYK2, JAK1, JAK2, and JAK3 and human whole blood cytokine assays, to measure
pathway inhibition, were utilized as the primary screens for the SAR campaign. Previous analysis
supports that efficacy would be driven by average (C ) IC coverage of IL-12 as the primary
av
80
cytokine target of combined TYK2 and JAK2 inhibition.21, 27 Similar coverage analysis for JAK
inhibitors approved for the treatment of RA has been reported. JAK inhibitors for the treatment
of RA demonstrate similar profiles of in vitro cytokine receptor inhibition.28 These drugs inhibit
JAK1 with limited selectivity against other JAK family members and show about IC coverage of
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
8
0
+
IFN signaling (mediated by JAK1 and TYK2). JAK2-mediated EPO inhibition (CD34 cells spiked
into human whole blood) would be limited to Cav ≤ IC30 to minimize anemia and
thrombocytopenia. To enable the analog advancement, a predicted human PK model would be
established to calculate target cytokine inhibition levels. For the expedience of regular SAR
interpretation, a calculated C inhibition or IC * was used (IC * = 100*((IC IL-12)/ (IC IL-12 +
av
xx
xx
80
80
IC EPO) to assess functional selectivity.
5
0
Results and Discussion
Identification of a selective ATP site inhibitor of TYK2 is challenging due to the high
homology across the four JAK isoforms.29 Rational design of a TYK2 inhibitor requires a hinge
binding motif that allows for the optimal scaffolding of vectors to further drive potency and
selectivity. In examining the JH1 ATP site of the four JAK kinases, there is a common methionine
gatekeeper, however the back pocket is defined by Ile960 in TYK2 which is a valine in the other
three JAK family members. This unique amino-acid difference sets the stage for an opportunity
to rationally design the hinge group moiety. Overlaying of the pyrrolopyridimidine hinge binding
group used in 1, 2 and 3 shows a 2-point hydrogen bond donor and acceptor interaction with
Glu979 and Val981 amino acids (Figure 2A). This binding mode places a steric clash against the
larger sidechain of Ile960 in TYK2 and forces the rotation of the lipophilic side chain towards the
polar NH of the pyrrolopyridimidine hinge. This can, in part, rationalize why compounds 1, 2, and
3 have ~30x, 20x, and 5x decreased potency for TYK2 compared to JAK1, respectively.
Identification of a hinge binding group that does not interact with the carbonyl of Glu969
could resolve the clash with TYK2 Ile960. A [5,6]-fused ring hinge binder with a hydrogen bond
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 54
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
acceptor and donor both to Val981 was proposed to allow for improved TYK2 potency. A [5,6]-
fused aromatic system would provide 1) a position off the six-membered ring to build into the P-
loop region and 2) a second opportunity to building towards the solvent front to manipulate
compound physicochemical properties (Figure 2B). The combination of the steric bulk of the
solvent exposed group and the unsubstituted 5-membered ring would prevent the hinge group
from flipping the binding mode to use the Glu979 and NH-Val981 positions. Targeting of the
Tyr980 and Val981 hinge binding mode was previously exploited in identifying TYK2 active
compounds, a strategy which ultimately led to the identification of the TYK2/JAK1 dual inhibitor
7.
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
ACS Paragon Plus Environment

Page 7 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Figure 2: a) Two different TYK2 hinge binding modes: modeled 2 (yellow) superimposed with 7
TYK2 protein X-ray structures (Green, PDB:6DBM) with Ile960 highlighted in purple; b) General
scaffolding strategy for a [5,6]-fused ring hinge binder.
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 [5,6]-fused aromatic hinge binding group would require a hydrogen bond acceptor
nitrogen atom on the five-membered ring, along with a hydrogen bond donor to project from the
6
-membered ring as an aryl C-H (Figure 2B). Analysis and selection of known kinase hinge binding
scaffolds as a starting point was used.30 Polarization of the aryl C-H to increase the hydrogen
bonding potential was believed to be necessary to bring about sufficient TYK2 activity and two
point binding. To that end several [5,6]-fused aromatic systems were designed and synthesized
based off previously known hinge motifs. Initial SAR employed a methylpyrazole moiety as a
solvent front group in a similar manner to that used in PF-06700841 (7). A substituted
trifluoroethylazetidinyl pyrazole group similar to baricitinib to engage the P-loop region was also
used. The imidazopyridine 8 was found to have modest potency against TYK2 (enzyme IC 290
5
0
nM) and showed low electrostatic potential on the C-H, which was utilized as a relative measure
of H-bond donor ability (Figure 3, Table 1).31 Addition of a nitrogen into either position of the
five-membered ring was calculated to modestly increase the electrostatic C-H potential and thus
both the [1,2,4]triazole[1,5]pyridine 9 and [1,2,4]triazole[4,3]pyridine 10 analogs were
synthesized. However, compound 9 was found to be only a moderate TYK2 inhibitor (476 nM)
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 54
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
and compound 10 was inactive against TYK2 with an IC50 >10 uM. In both cases there was
increased electro-positive potential (~ 50 kJ/mol) not only on the C7-H but also on the 2- or 3-
positions of the 5-membered ring. These positions face towards the lipophilic gatekeeper the
added polarity in this region likely contributed to the reduced TYK2 potency. In addition, the 2-
position nitrogen of 10 could also be clashing with the carbonyl oxygen of Glu979 leading to
complete loss of TYK2 activity.
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
With these results in hand, we turned to modification of the six-membered ring system
to improve potency. A pyrazole[1,5]pyrazine core32-38 was of particular interest due to the
calculated increased polarization of the C-H for hydrogen bond donor properties (122 kJ/mol). In
contrast to 9 and 10, the most electronegative nitrogen atom was now placed towards the
solvent front where it could be better accommodated by bulk water and not clash with the
protein (Figure 3). Gratifyingly the pyrazole[1,5]pyrazine analog 11 was found to be a potent
TYK2 enzyme inhibitor at 10 nM with a lipophilic efficiency (LipE) of 5.5. With the identification
of an appropriately potent and efficient hinge binding core, compound 11 was chosen for further
optimization to improve physicochemical properties and JAK selectivity.
ACS Paragon Plus Environment

Page 9 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Figure 3: [5,6] Fused aromatic TYK2 lead structures 8-11 with electrostatic potential maps
calculated in Spartan ‘18 with Hartree-Fock 6-31G*
imidazopyridine
[1,2,4]triazole[1,5]pyridine
[1,2,4]triazole[4,3]pyridine
pyrazole[1,5]pyrazine
F
F
F
F
F
F
F
F
F
F
F
F
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
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
3
4
N
N
5
N
N
N
N
N
2
N
N
N
N
6
N
N
N
N
N
N
1
7
H
N
N
H
H
H
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 54
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 1: TYK2 biochemical IC , sfLogD, and TYK2 LipE of compounds [5,6] Fused aromatic TYK2
5
0
lead structures 8-11.
Compound
#
8
9
10
11
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
TYK2a
2
90
476
2.3
4.0
>10,000
1.2
10
2.5
5.5
IC (nM)
5
0
sfLogDb
2.1
4.5
TYK2
LipEc
<3.8
a
Enzyme assay. Compounds were assayed at least twice, and the IC reported as the geometric
5
0
b
mean. ATP concentration = 1 mM. Shake-flask LogD (sfLogD) measured between octanol/water
c
phosphate buffered to pH 7.4. LipE = pIC – LogD;
5
0
An X-ray crystal structure of 11 bound to TYK2 confirmed the expected binding mode with
N1 accepting a hydrogen bond from the amide NH of Val981 in the TYK2 hinge region (Figure 4a).
The proposed polarized C7-H was observed to be 2.4 Å from the carbonyl of Val981, within van
der Waals radii. The positioning of the alkyl cyano group was shown to be towards the C-terminal
lobe in the active site’s lower lipophilic pocket. This mimicked the positioning of the 4R-
methylpiperidine positioning observed with tofacitinib.39 The trifluoromethyl azetidine group
provided the desired TYK2 kinase potency through engagement of the P-loop, but the moderate
LipE of 5.5 reflected the high overall lipophilicity of the molecule. Based on the physicochemical
-6
properties and high permeability (MDCK-LE = 18 x10 cm/sec) the primary clearance mechanism
was predicted to be oxidative metabolism.40 Optimization of the azetidine substituent was
guided by LipE through increasing potency/lowering LogD as well as driving towards improving
the modest metabolic stability profile of 11 as judged by human liver microsome turnover (HLM)
of 18 L/min/mg protein. In addition, replacement of the trifluoroethyl amino group was
preferred due to association with testicular toxicity liability for potential metabolites.41 The des-
fluoro ethyl compound 12 lost approximately 5-fold in TYK2 potency relative to 11 with no change
to the overall JAK selectivity profile (Table 2). Interestingly, both compounds showed moderate
ACS Paragon Plus Environment

Page 11 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
metabolic profiles (HLM = 18 and 21 L/min/mg protein). Attempts to recapitulate the trifluoro-
group impact using a cyclopropane analog 13 did not improve potency (TYK2 IC = 40 nM) and
5
0
was reflected in a loss of LipE (4.8) and increased HLM clearance (30 L/min/mg protein).
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
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 54
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
Figure 4: a) Protein X-Ray structure of 11 bound to TYK2 (Green, PDB: 6X8F) at 2.15 Å
with crystallographic waters; b) Overlay of P-loop region of protein X-Ray structure of 11 bound
to TYK2 (green) at 2.15Å with Connolly surface and JAK2 (Blue, PDB: 6X8E,) at 1.75Å.
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
Further attempts to introduce polarity to lower clearance while retaining TYK2 potency
provided by the trifluoro group provided mixed results. Introduction of the azetidine amide in
compound 14 lowered both sfLogD and brought HLM clearance (< 8 L/min/mg protein) into the
desired range, but also caused a significant loss in potency. This could be due to the amide sp2
character causing flattening of the azetidine which may not allow the acetyl group to directly
engage the P-loop. Appending a hydroxyl to the ethyl group provided compound 15 again showed
a loss of potency, even with the increased flexibility of the azetidine N-substituent. In contrast
to alkyl groups, 14 and 15 that bear polar groups both had lower permeability as judged by flux
-6
in an MDCK-LE cell line (4 x10 cm/sec). Unique within this series, the cyanomethyl group of
compound 16 recapitulated the potency of compound 11, but also lowered HLM clearance (<8
L/min/mg protein) and increased LipE to 6.7. Compound 16 showed good selectivity against
JAK3 and JAK1, but only a modest 5-fold selectivity against JAK2.
ACS Paragon Plus Environment

Page 13 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 2: SAR of substituted azetidines
R1
N N
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
N
N
N
N
N
MDCK-LE42
HLM
Compound
No.
TYK2a
JAK1 a
JAK2 a
JAK3 a
TYK2
sfLogDb
(x10-6
R1
L/min/mg
protein)
LipEc
IC (nM) IC (nM) IC (nM) IC (nM)
5
0
50
50
50
cm/sec)
18
F
F
F
1
1
1
1
2
3
N
N
10
358
65
>9,851
2.5
2.1
2.6
5.50
5.26
4.84
18
21
30
N
46
40
2,669
933
346
187
>10,000
>10,000
33
27
N
N
N
O
N
1
1
4
5
678
189
6,111
5,579
1,747
1,044
>10,000
>10,000
1.2
1.2
4.92
5.52
<8
<8
4
7
N
HO
N
N
N
N
16
7
250
37
6,682
1.5
6.65
<8
11
N
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 54
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
aCompounds were assayed at least twice, and the IC50 reported as the geometric mean. ATP
b
concentration = 1 mM. Shake-flask LogD (sfLogD) measured between octanol/water phosphate
c
buffered to pH 7.4. LipE = pIC – LogD.
5
0
To provide further understanding of JAK2 selectivity, a JAK2 protein X-ray crystal structure
of 11 was acquired to compare to the TYK2 structure (Figure 4b). Comparing the protein
structures of JAK2 and TYK2 superimposed, the JAK2 P-loop was slightly raised relative to the
TYK2 P-loop. Overlay of the two structures and electron density unsurprisingly showed similar
binding modes. Subtle interaction differences with the P-loop have been reported to influence
selectivity in other JAK inhibitors.43 The only minor difference seen in the overlay was the
azetidine ring configuration. In the JAK2 structure at 1.75 Å resolution, the azetidine adopted 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
‘
trans-like’ configuration placing the terminal -CF group towards the P-loop. In contrast, the
3
TYK2 azetidine configuration adopted a ’cis-like’ configuration. These observations from the X-
ray structures provided the hypothesis that fixing the four-membered ring configuration that
could provide the C2-symetric achiral cis or trans conformations could improve selectivity and
replacing the azetidine with a cyclobutane may accomplish this goal.
The methoxycyclobutane diastereomers cis-17 and trans-18 were prepared and both
retained TYK2 potency (16 and 22 nM) while the cyclobutane stereoisomer trans-18 showed
superior selectivity against JAK1 (68-fold) and JAK2 (10-fold). As expected with the replacement
of the azetidine, the methoxycyclobutane analogs showed increased sfLogD and the cis-17 had
high clearance as judged by HLM (27 L/min/mg protein). Somewhat unexpectedly, the
isolipophilic trans-18 had very low HLM clearance (< 8 L/min/mg protein) when compared to
the cis-17. The two diastereomeric cis/trans isomers demonstrated a useful divergence in
clearance in addition to differences in potency and selectivity.
The cis-19 and trans-20 isomers of 2-cyclobutylacetonitrile were prepared as homologs
of azetidine 16. The cis-19 configuration provided a near identical profile to the azetidine 16
while the trans-20 had a slight loss in potency across all JAK isoforms. The JAK2 selectivity of cis-
1
9 and trans-20 were lower than the methoxy analogs 17 and 18 with higher in vitro clearances.
To improve JAK2 selectivity by further removal of flexibility in the P-loop engagement, truncation
ACS Paragon Plus Environment

Page 15 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
of the 2-cyclobutylacetonitrile to the cyclobutanecarbonitrile group was proposed. The cis-21
cyclobutanecarbonitrile showed excellent TYK2 potency (6 nM) but lacked selectivity against
JAK2 (8 nM). In addition, even with the lower measured sfLogD (1.8), significant clearance was
still observed (HLM = 17 L/min/mg protein), in line with the previous results for cis-17. In
contrast, the trans-22 cyclobutanecarbonitrile (PF-06824467) showed good selectivity against
JAK1 and JAK2 with very low HLM clearance (<8 L/min/mg protein) consistent with trans-20.
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
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 54
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 3: Influence of cyclobutane substituent on potency, sfLogD, and metabolism of compounds
1
7-22.
R1
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
N
N
N
N
N
HLM
TYK2a
JAK1 a
JAK2 a
JAK3 a
MDCK-LE42
Compound
No.
TYK2
LipEc
R1
sfLogDb
2.2
(L/min/mg
protein)
-6
IC (nM) IC (nM) IC (nM) IC (nM)
(x10 cm/sec)
5
0
50
50
50
O
1
1
7
8
16
664
99
>10,000
5.65
5.41
27
<8
20
22
N
O
22
8
1,495
273
228
38
>9121
>6,170
2.3
N
N
N
1
9
1.8
1.9
6.32
5.74
10
11
21
20
N
2
2
0
1
24
764
159
>10,000
N
N
6
21
8
1,051
1.8
1.7
6.47
6.09
17
<8
15
17
N
N
22
15
383
74
>10,000
N
ACS Paragon Plus Environment

Page 17 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
aCompounds were assayed at least twice, and the IC50 reported as the geometric mean. ATP
b
concentration = 1 mM. Shake-flask LogD (sfLogD) measured between octanol/water phosphate
c
buffered to pH 7.4. LipE = pIC50 – LogD.
To better understand the interactions of the cyclobutane analogs with TYK2, a protein X-
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
ray crystal structure of 22 was acquired. The cis cyclobutanecarbonitrile moiety was positioned
below and towards the tip of the P-loop (Figure 5a). Simulated waters in the apo structure
showed a calculated high energy water molecule (G = 5.3 kcal/mol) in this region.44 The end of
the P-loop contains mainly hydrophobic residues and a generally satisfied hydrogen bond
network which may contribute to the calculated lability of this modeled water in the apo
structure. This predicted high energy water is displaced by the cyclobutanecarbonitrile group of
2
2. The electrostatic mapping of this pocket also showed a positive electrostatic potential which
could provide a favorable interaction with the cyano group (Figure 5a). The optimized hinge
binding group along with the rationally designed interactions with the P-loop combined to give
potent TYK2 inhibition.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 54
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
Figure 5: a) Protein X-ray structure of 22 in TYK2 (Yellow, PDB: 6X8G)) at 2.21Å with
electrostatic potential map and modeled APO waters in the P-loop (spheres); b) Top view of the
hinge and solvent exposed region of the protein X-ray structure of 22 showing 2-point hinger
interaction with Val981
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
Based on the TYK2 potency, JAK selectivity profile, and low clearance, 22 was selected to
further explore the SAR of the pyrazole group (R2) (Table 4). The SAR was also explored on the
more potent but less JAK1 selective trans-21 in order to attempt to improve TYK2 selectivity in a
parallel effort. A matrix of pyrazole structural analogs were made in an attempt to further tune
the properties. The presumed solvent exposed methyl group on the pyrazole was hypothesized
to be inefficient lipophilicity (Figure 5b). Removal of the methyl group from the 4-pyrazole in the
cis-23 and trans-24 pair showed an increase in HLM clearance despite reduction in LogD and
lowered the permeability as measured by MDCK-LE. Neither compound showed improvement
from their corresponding methyl analogs with no improvement in JAK1 selectivity from 21 to 23
and an erosion of TYK2 potency from 22 to 24. In order to improve the permeability of the NH
pyrazole, the 1H-3-pyrazole would allow a tautomer to place the N-H towards the core nitrogen
and mask the H-bond donor via an intermolecular hydrogen bond. This strategy wasn’t
successful when applied to cis-25, but gratifyingly, the trans-26 analog showed improved
permeability compared to 24. In addition, TYK2 potency was partially regained while retaining
low HLM clearance. Finally, the 1-methyl-3-pyrazole analogs were synthesized. In the case of
cis-27 there was a gain in JAK1 and JAK2 selectivity, but this was still a higher clearance compound
ACS Paragon Plus Environment

Page 19 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
based on HLM. Analog trans-28 had weaker TYK2 potency compared to 22 but had similar low
clearance and JAK selectivity.
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
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 54
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 4: SAR of the hinge pyrazole region compounds 23-28.
N
Cis
Trans
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
N
N N
N
N N
N
N
N
N
N
R2
N
R2
Cyclo-
butane
JAK1a
IC50
JAK2a
IC50
JAK3a
MDCK-LE42
(x10-6
HLM
Compound
No.
TYK2a
IC (nM)
TYK2
LipEc
R2
IC50
sfLogDb
(L/min/mg
protein)
Stereo-
5
0
(nM)
(nM)
(nM)
cm/sec)
chemistry
2
2
3
4
Cis
Trans
Cis
8
22
22
280
31
2,242
>10,000
4,448
1.6
1.7
1.7
1.7
1.6
1.6
6.52
5.39
6.37
5.88
6.40
5.64
44
19
20
<8
23
<8
6
7
N
N
H
44
9
398
43
25
7
N
N
H
2
2
2
6
7
8
Trans
Cis
26
10
67
284
261
2,534
172
45
>10,000
>9,588
>10,000
17
14
20
N
N
Trans
467
aCompounds were assayed at least twice, and the IC50 reported as the geometric mean. ATP
b
concentration = 1 mM. Shake-flask LogD (sfLogD) measured between octanol/water phosphate
c
buffered to pH 7.4. LipE = pIC50 – LogD.
A selection of compounds that showed both TYK2 biochemical potency and low in vitro
metabolic clearance (HLM <8 L/min/mg protein) were selected to advance into profiling in
human whole blood (HWB) cytokine assays. These assays were established to identify the
desired candidate profile with respect to efficacy and off-target effects. Inhibition of IL-12-
stimulated phosphorylation of STAT4 was selected as the efficacy driver and blockade of EPO-
ACS Paragon Plus Environment

Page 21 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
induced phosphoSTAT5 was the key selectivity measurement. The project objective was to
achieve potent IL-12 inhibition (IC < 50 nM) with a selectivity profile of IC < 0.30 on EPO when
5
0
xx
set to a calculated IC of IL-12 coverage. For all compounds, human plasma protein binding was
8
0
measured to correct for free-drug potency. The compounds profiled were moderately bound
with fraction unbound ranging from 0.2 - 0.5 (Table 5). In addition, in vitro clearance was further
profiled using human hepatocytes (HHEP) as a complement to the HLM data (CYP450
metabolism) in order to cover both phase I and II metabolic processes.
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
Of the compounds selected for this more extensive profiling, methoxycyclobutane 18 [PF-
0
6799018] showed the largest TYK2/JAK2 selectivity at 10-fold. Unfortunately, even with good
TYK2 biochemical potency (22 nM) there was only moderate inhibition in HWB of IL-12 (IC50,u 186
nM). Compound 18 had a low metabolic clearance (HHEP = 6 L/min/mil cells). Both 26 and 28
had weaker TYK2 potency and more moderate JAK2 selectivity which translated into weak overall
IL-12 inhibition (IC50,u 334 and 254 nM respectfully). Compound 21 was the most potent
compound identified on TYK2 which drove strong potency against IL-12 (IC50,u 9 nM). However
insufficient biochemical selectivity against JAK2 meant there was an unacceptably high coverage
of EPO (IC63). In addition, 21 showed a moderate but increased clearance profile (HHEP =12
L/min/mil cells). The cis-cyclobutane 22 showed weaker TYK2 enzyme activity compared to its
stereoisomer 21, but with more selectivity against JAK2. With this profile, 22 provided the
desired level of IL-12 potency (IC50,u 14 nM). Compound 22 also demonstrated the desired
selectivity profile with an IC for EPO coverage at the IL-12 IC , and a very low in vitro clearance
2
8
80
profile (HHEP <3 L/min/mil cells). Further profiling of 22 against a panel of 231 kinases at non-
physiologically conditions for each kinase Km showed 21 non JAK kinases with >50% at 1 uM
dose. However, a kinase biochemical promiscuity panel against 40 kinase targets at
physiologically relevant conditions with the 1 mM ATP and under those conditions no kinase
target screened in both formats had >50% inhibition. (ThermoFisher Kinase Profiling, See
Supplementary Table 1)45 Based on the strong IL-12 potency, acceptable off-target profile, and
excellent metabolic clearance profile, 22 was selected to advance to in vivo profiling.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 54
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 5: HWB cytokine and HHEP profile of potent TYK2 compounds 18, 21, 22, 26, and 28.
Human
Plasma
TYK2 IC50
nM
IL-12
Free
IC50
EPO
Free
IC50
HHEP
(l/min/mil
cells)
Compound
No.
ICxx
Protein (TYK2/JAK2
Binding approx. fold
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
EPOc
(
nM)b
(nM)b
Fua
selectivity)
22 (10x)
6 (1x)
1
2
8
1
0.30
0.34
0.38
0.32
0.20
186
9
1,100
22
0.41
0.63
0.28
ndd
6
12
< 3
7
22
26
15 (5x)
14e
334
254
148e
ndd
26, (7x)
2
8
67 (7x)
ndd
ndd
< 2
a
b
Fraction unbound in human plasma. Compounds were assayed at least twice, and the IC
5
0
reported as the geometric mean, Free IC50 calculated using assumed blood to plasma
c
d
e
concentration ratio = 1. ICxx EPO = 1 / ((EPO IC /(4 x IL-12 IC )+1). Not determined. Free IC
5
0
50
50
calculated by whole blood IC50 * Fu / measured blood to plasma concentration ratio.
The physical properties of 22 include a basic pK of <1.7 indicating the compound would
a
be neutral at physiological pH. The measured logD at pH 7.4 was 1.7. Thermodynamic solubility
of the crystalline material was low at 0.8 µM (0.3 µg/mL) at pH 7.4, and a salt formation strategy
to improve solubility could not be pursued due to the lack of ionizable groups. A Spray Dried
Dispersion (SDD) formulation improved experimental intestinal solubility to approximately 160-
180 μM (60-70 µg/mL). Compound 22 showed high passive permeability in MDCK-LE cells (mean
-6
Papp 16.7 x 10 cm/s) but is a substrate for MDR1 (efflux ratio >6) and BCRP (efflux ratio >2)
efflux transport.46
The pharmacokinetics of 22 were studied in Sprague-Dawley rats. Following a 1 mg/kg
intravenous dose, the systemic clearance was 12 mL/min/kg, the steady state volume of
distribution (V ) was 1.1 L/kg, and the AUC was 1,380 ng·h/mL. Following oral administration
ss
inf
of the crystalline (non-SDD) material at 3 and 30 mg/kg in 0.5% methylcellulose, oral
bioavailability was 15 and 8% respectively. Oral bioavailability significantly improved to 58 and
ACS Paragon Plus Environment

Page 23 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
6
3% at 3 and 30 mg/kg respectively when dosed in a SDD at 25% drug load, indicating an
improved fraction absorbed of >75%. At 3 and 30 mg/kg in SDD, the respective Cmax was 369 and
2,780 ng/mL and the respective AUCinf was 2,390 and 25,900 ng·h/mL. Plasma protein binding
was 18% in rat and 38% in human without significant blood to plasma distribution (1.4 in human).
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
The in vitro metabolism of 22 was studied in human liver microsomes (<8 μL/min/mg
protein) and hepatocytes (1.8 μL/min/million cells), both of which indicated low metabolic
turnover (Table 6). Oxidative metabolites of 22 accounted for the primary routes of
biotransformation in rat, monkey, and human in vitro hepatic systems consistent with CYP450
metabolism. Primary metabolic pathways included N-demethylation of the pyrazole,
hydroxylation of the pyrazole, and addition of oxygen and loss of the nitrile groups. Renal (1.3%)
and bile (0.9%) excretion was low in the rat. Considering the importance of CYP450 metabolism,
human clearance predictions were made using both human liver microsomes and hepatocyte
scaling (well-stirred model) indicating a blood clearance of 2.3 and 1.5 mL/min/kg, respectively.
®
The V was predicted to be 1.2 L/kg using SimCYP PBPK modeling, consistent with the neutral
ss
properties of the compound. Oral bioavailability was expected to be dose-dependent based on
low solubility characteristics and the need for a solubility enabling formulation.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 54
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 6: Summary of predicted human blood pharmacokinetics of 22 from in vivo and in vitro
systems
Method
Human Prediction
CL (mL/min/kg)
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
Vss (L/kg)
b
Human Liver Microsomes
Human Relay Hepatocytes
Rat Single Species
2.3
1.5
5.2
2.3
1.2
SimCYP® PBPK modeling
CL : blood clearance; V : volume of distribution at steady state; PBPK: physiologically based
b
ss
pharmacokinetics: Human clearance predictions from in vitro systems were scaled based on the well-
stirred model or single species scaling from allometry47,48
The overall cytokine inhibition profile (IC ) was calculated for 22 based upon the modeled
XX
daily C exposure (i.e. AUC/dosing interval = 56 nM unbound) to cover the calculated IC of IL-
av
80
1
2 (unbound IC50 14 nM x 4 = 56 nM). Relative average daily cytokine inhibition representing
different JAK signaling pairs is shown in Table 7 and Figure 6.
ACS Paragon Plus Environment

Page 25 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 7: Human whole blood potencies for 22 with different cytokine stimuli, and IC calculated
XX
from predicted C exposure to cover IL-12 IC
av
80.
JAK Signaling
Complex
Uncorrected IC50
(nM) Mean ± SDa
53 ± 16
b
Human whole blood
ICxx
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
IL-12-induced pSTAT4 IC (lymphocytes)
TYK2/JAK2
0.80
0.65
0.76
0.33
0.51
0.27
0.05
0.04
0.04
0.05
0.09
5
0
IL-23-induced pSTAT3 IC (lymphocytes)
TYK2/JAK2
112 ± 18
5
0
IFNα-induced pSTAT3 IC (lymphocytes)
TYK2/JAK1
66 ± 21
5
0
IL-10-induced pSTAT3 IC (lymphocytes)
TYK2/JAK1
427 ± 157
5
0
IL-27-induced pSTAT3 IC (lymphocytes)
TYK2/JAK1/JAK2
TYK2/JAK1/JAK2
TYK2/JAK1/JAK2
JAK1/JAK3
201 ± 25
5
0
+
IL-6-induced pSTAT1 IC (CD3 cells)
566 ± 283
5
0
+
IL-6-induced pSTAT3 IC (CD3 cells)
4,370 ± 1,260
4,890 ± 2,550
4,840 ± 2,260
3,740 ± 270
2,220 ± 750
50
IL-21-induced pSTAT3 IC (lymphocytes)
50
IL-15-induced pSTAT5 IC (lymphocytes)
JAK1/JAK3
5
0
IL-2-induced pSTAT5 IC (lymphocytes)
JAK1/JAK3
5
0
+
IFNγ-induced pSTAT1 IC (CD14 )
JAK1/JAK2
5
0
+
EPO-induced pSTAT5 (spiked CD34 progenitor
cells)
JAK2/JAK2
547 ± 195
0.28
aCompound 22 was assayed at least three times in each assay, and the IC50 reported as the
c
geometric mean. SD = standard deviation. b ICxx cytokine = 1 / ((Whole blood IC /(4 x IL-12
50
IC )+1).
5
0
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 54
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
Figure 6. Radial plot of cytokine inhibition IC relative to the IL-12 IC for compound 22, derived
xx
80
from potency in human whole blood assays and predicted average daily human plasma drug
concentration.
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
Efficacy of 22 was evaluated in the mouse imiquimod-induced skin inflammation model
employing a therapeutic dosing paradigm previous used to profile 1 of an anti-p40 antibody.49 In
this mouse skin inflammation model, the subjects are challenged with topically-applied
imiquimod (Toll-Like Receptor 7 and 8 agonist) on the ear to induce inflammation followed by
oral dosing of 22. In this mouse model, the murine TYK2 gene was modified by introducing a
humanizing point mutation in order to change amino acid 980 of murine TYK2 protein from valine
to isoleucine. This was done in order to compensate for an observed species potency shift driven
by this amino acid difference between human and preclinical species.50 Oral treatment with 22
at 30, 10, and 3 mg/kg, once per day (QD) reduced clinical endpoints associated with skin
inflammation in this model. The measured average unbound daily concentrations (Cav,u nM) of 1,
3
, 10, and 30 mg/kg were 14.5, 45.7, 225, and 668 nM respectively, resulting in average IL12
inhibition of 51, 77, 94, and 98%. No significant effect on ear thickness was seen at 1 mg/kg
compared to control animals at the end of the 5-day study.
ACS Paragon Plus Environment

Page 27 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Figure 7: Compound 22 dosed at 30, 10, 3, and 1 mg/kg, PO, QD inhibited ear swelling in the
imiquimod-induced skin inflammation model. Compound 22 significantly inhibited ear swelling
when compared to the vehicle-treated mice (n=12) (30 mg/kg, 46.2%, p=0.0004, 10 mg/kg,
34%, p=0.0145, and 3 mg/kg, 31.3% (0.0198). The effect of the 1 mg/kg dose was not
statistically significant (ns) when compared to the vehicle-treated mice.
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
Chemistry
Synthesis of the [5,6]-fused heterocycle structures 8, 9 and 10 was accomplished as
shown in Scheme 1. Conjugate addition of 4-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)-1H-
pyrazole (29) with tert-butyl 3-(cyanomethylene)azetidine-1-carboxylate (30)51 gave the
elaborated pyrazole boronate 31, which underwent Suzuki-Miyaura coupling independently with
3
2 and 33 to provide 35 and 34, respectively (see Supporting Information). A second Suzuki-
Miyaura coupling of 1-methyl-4-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)-1H-pyrazole with
4 and 35 afforded 36 and 37, respectively. The syntheses were completed by BOC deprotection
3
and azetidine alkylation with 2,2,2-trifluoroethyl triflate to provide 8 from 36 and 9 from 37. The
alkylation of the deprotected azetidines required neutralization of the acid addition salts
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 54
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
obtained by BOC deprotection and elimination of chloride or trifluoroacetate counterions prior
to exposure to 2,2,2-trifluoroethyl triflate. The presence of chloride was found to result in the
formation of unreactive 2,2,2-trifluoroethyl chloride, while the presence of trifluoroacetate ion
was found to result in the formation of 2,2,2-trifluoroethyl trifluoroacetate, which in turn
afforded the azetidine trifluoroacetamides as a serious competing reaction. Similarly, compound
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
8 (see Supporting Information) underwent Suzuki-Miyaura coupling with (1-(tert-
butoxycarbonyl)-1H-pyrazol-4-yl)boronic acid to provide 39 after BOC deprotection in situ.
Conjugate addition of 39 with 30 gave the N-BOC azetidine 40. BOC deprotection followed by
azetidine alkylation with 2,2,2-trifluoroethyl triflate completed the synthesis of 10.
Scheme 1: Synthesis of 8, 9, and 10
R
R
N
N
Boc
N
Cl
N N
c
N
b
N N
N
HN N
B
A
N
a
N
N N
B
N
+
X
A
N
A
N
N
N
O
O
O
O
Cl
N
N
29
31
32 A = N; X = Br
34 R = BOC A = CH
35 R = BOC A = N
36 R = BOC; A = CH
d
d
33 A = CH; X = Cl
8 R = -CH CF ; A = CH
2
3
3
7 R = BOC; A = N
9
R = -CH CF ; A = N
2
3
R
N
HN N
N
Cl
N N
e
f
N
N
40 R = BOC
N
d
N
10
2
R = -CH CF3
N
N
N
N
N
N
N
N
N
38
39
N
N
ACS Paragon Plus Environment

Page 29 of 54
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Reagents and conditions: (a) 30, DBU, MeCN, 50-90 °C, 4-18 h; (b) K CO or Cs CO , Pd(dppf)Cl ,
2
3
2
3
2
1
2
,4-dioxane, H O, 95–100 C, 16 - 48 h; (c) 1-methyl-4-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-
2
-yl)-1H-pyrazole, K CO or CsF, Pd(dppf)Cl , 1,4-dioxane or Et(Me) COH, 70–100 C, 18 - 48 h;
2
3
2
2
(d) 1. MsOH or HCl, MeCN or dioxane, 20–30 C, 16–20 h; 2. CF CH OTf, DIEA, DMF or MeCN, 0–
3
2
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
20 C, 2–24 h. (e) (1-(tert-butoxycarbonyl)-1H-pyrazol-4-yl)boronic acid, Cs CO , Pd(dppf)Cl ,
2
3
2
1
,4-dioxane, H O, 95 C, 16 h; (f) 30, DBU, MeCN, 30 C, 16 h.
2
The pyrazolo[1,5-a]pyrazine analogues (11–22) required a longer synthetic sequence to
allow the selective installation of the differentially functionalized pyrazole substituents at the 4-
and 6-positions of the pyrazolo[1,5-a]pyrazine ring. Initially, the N-methyl pyrazole at the 6-
position was introduced at the beginning of the synthesis so that its placement in the
pyrazolo[1,5-a]pyrazine ring system would be unambiguous following construction of the
pyrazine ring (Scheme 2). Reaction of the α-bromoketone 41 (see Supporting Information)52
with diethyl 1H-pyrazole-3,5-dicarboxylate 42 provided 43. This route was selected to avoid
regioselectivity issues associated with attempts to alkylate ethyl pyrazole-3-carboxylate.53
Ketone 43 was cyclized to the pyrazolo[1,5-a]pyrazine 44 by condensation with ammonium
acetate.54 Saponification to 45 followed by thermal decarboxylation gave the pyrazolo[1,5-
a]pyrazin-4(5H)-one 46, which was treated with phosphoryl chloride to provide the key chloro
pyrazolo[1,5-a]pyrazine intermediate 47.
Scheme 2: Synthesis of 47
ACS Paragon Plus Environment