Design and synthesis of a Pan-Janus kinase inhibitor clinical candidate (PF-06263276) suitable for inhaled and topical delivery for the treatment of inflammatory diseases of the lungs and skin

Article abstract of DOI:10.1021/acs.jmedchem.6b01634

By use of a structure-based computational method for identification of structurally novel Janus kinase (JAK) inhibitors predicted to bind beyond the ATP binding site, a potent series of indazoles was identified as selective pan-JAK inhibitors with a type 1.5 binding mode. Optimization of the series for potency and increased duration of action commensurate with inhaled or topical delivery resulted in potent pan-JAK inhibitor 2 (PF-06263276), which was advanced into clinical studies.

Full text of DOI:10.1021/acs.jmedchem.6b01634

Subscriber access provided by GAZI UNIV
Article
Design and Synthesis of a pan-Janus Kinase Inhibitor Clinical
Candidate (PF-06263276) Suitable for Inhaled and Topical Delivery
for the Treatment of Inflammatory Diseases of the Lungs and Skin
Peter Jones, R Ian Storer, Yogesh A Sabnis, Florian M Wakenhut, Gavin A. Whitlock, Katherine S
England, Takasuke Mukaiyama, Christoph M Dehnhardt, Jotham Wadsworth Coe, Steven W. Kortum, Jill
E Chrencik, David G. Brown, Rhys M Jones, John R Murphy, Thean Yeoh, Paul M Morgan, and Iain Kilty
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.6b01634 • Publication Date (Web): 16 Dec 2016
Downloaded from http://pubs.acs.org on December 17, 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 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Design and Synthesis of a panꢀJanus Kinase
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
Inhibitor Clinical Candidate (PFꢀ06263276) Suitable
for Inhaled and Topical Delivery for the Treatment
of Inflammatory Diseases of the Lungs and Skin
a,b*
b*,
α
b,
Yogesh A. Sabnis, β Florian M. Wakenhut, Gavin A.
b,γ
Peter Jones, R. Ian Storer,
b,
δ
b,
ε
b,
ζ
a,
η
Whitlock, Katherine S. England, Takasuke Mukaiyama, Christoph M. Dehnhardt,
c
c
c
d,
θ
e
Jotham W. Coe, Steve W. Kortum, Jill E. Chrencik, David G. Brown, Rhys M. Jones, John
f
g
h
h
R. Murphy, Thean Yeoh. Paul Morgan, Iain Kilty,
a
e
h
Medicine Design, Pharmacokinetics, Dynamics and Metabolism, Inflammation and
Immunology Research Unit, Pfizer Inc, 610 Main Street, Cambridge, MA 02139, USA;
c
g
Medicine Design, Medicinal Sciences, Pfizer Inc, 445 Eastern Point Road, Groton, CT, 06340,
b
d
f
USA; Worldwide Medicinal Chemistry, Structural Biology and Biophysics, Pharmaceutical
Sciences, Pfizer Ltd, Ramsgate Road, Sandwich, CT13 9NJ, UK
ABSTRACT: Using a structureꢀbased computational method for identification of structurally
novel Janus kinase (JAK) inhibitors predicted to bind beyond the ATP binding site, a potent
series of indazoles was identified as selective panꢀJAK inhibitors with a type 1.5 binding mode.
Optimization of the series for potency and increased duration of action commensurate with
ACS Paragon Plus Environment
1

Journal of Medicinal Chemistry
Page 2 of 62
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
inhaled or topical delivery resulted in potent panꢀJAK inhibitor 2 (PFꢀ06263276) which was
advanced into clinical studies.
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
Janus kinases, JAKs, are a family of tyrosine kinases comprising four members (JAK1, 2, 3
and Tyk2). They play a critical role in both innate and adaptive immunity since they function as
1
key transducers in the signaling processes of many cytokine receptors. As such, they have
emerged as attractive targets for a number of inflammatory diseases, and numerous inhibitors are
2
under active clinical development. Amongst these, tofacitinib, 1 (Figure 1) was first approved in
3
2
012 by the FDA for the treatment of Rheumatoid Arthritis (RA). The efficacy and safety of
tofacitinib has been studied in several immuneꢀmediated inflammatory diseases, including
4
5
6
7
8
rheumatoid arthritis, psoriasis, ulcerative colitis, Crohn’s disease, psoriatic arthritis, and
9
ankylosing spondylitis.
Using JAK enzymatic assays run at physiologically relevant 1 mM ATP concentration,
tofacitinib displays selectivity for JAK1 inhibition, although it also inhibits JAK3 (3 fold
weaker) and JAK2 (5 fold weaker). In a cellular settting, tofacitinib preferentially inhibits
signalling by heterodimeric cytokine receptors that associate with JAK3 and/or JAK1 with
functional selectivity over cytokine receptors that signal via pairs of JAK2. For example in
human whole blood assays, tofacitinib inhibits ILꢀ15, IFNα and ILꢀ6 signaling with IC50s
between 35ꢀ80 nM whereas the IC50 for JAK2 dependent EPO signaling was ~ 300 nM. Given
the role of JAK2 dependent EPO signaling in red blood cell production, reduction in hemoglobin
2
levels was observed at higher doses with tofacitinib in clinical trials. Despite this, the functional
selectivity for cytokine inhibition for EPO signaling has proven to be manageable at clinical
ACS Paragon Plus Environment
2

Page 3 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
doses that provide benefit to rheumatoid arthritis and psoriasis patients without reducing
4
a, 10
hemoglobin.
There are a number of JAK inhibitors in clinical development which have varying selectivities
within the JAK family. In a cellular setting the advanced candidates preferentially inhibit JAK1
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
dependent cytokines, including the γꢀcommon chain cytokines, interferons and ILꢀ6. Since
tofacitinib and other JAK inhibitor advanced clinical candidates preferentially inhibit JAK1ꢀ
dependent cytokines little clinical differentiation is expected. To further improve upon the
therapeutic potential of JAK inhibition in general, a greater degree of selectivity over systemic
JAK2 inhibition and EPO signaling in particular would be desirable. Besides attempting to
produce an inhibitor with higher selectivity amongst the JAK family, an alternative approach for
certain diseases would be to target JAK inhibition directly at the site of action in the relevant
organ of interest, while avoiding widespread systemic JAK inhibition. Herein we describe the
1
1
design and development of a novel and potent panꢀJAK inhibitor 2 (PFꢀ06263276, Figure 1),
suitable for inhaled or topical administration as an organ targeted therapy. Compound 2 has the
1
2
potential to treat inflammatory diseases of the lungs (such as COPD) and skin (such as mildꢀtoꢀ
5
a
moderate psoriasis) with negligible systemic JAK inhibition.
Figure 1. Structures of tofacitinib (1) and novel panꢀJAK inhibitor (2).
ACS Paragon Plus Environment
3

Journal of Medicinal Chemistry
Page 4 of 62
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
Results and Discussion
1
3
Following our previously described concept of ‘inhalation by design’, we sought to design a
JAK inhibitor that fulfilled the following criteria: (1) high potency to minimize dose size, given
the limits of dose sizes of typical dry powder inhalation (DPI) devices (<1 mg drug product) and
to minimize the percentage of active ingredient required in topical formulations, (2) high
selectivity for the JAK family of kinases versus the rest of the kinome (3) low systemic exposure
and high metabolic clearance to maximize the therapeutic index by targeting drug only to the site
of action (with no potent active metabolites formed), (4) multiple routes and mechanisms of
metabolism to minimize the potential for any drugꢀdrug interactions, (5) suitable solid form
characteristics to allow for blending with lactose for use in a DPI device and formulation as a
topical ointment or cream, (6) evidence of long duration of action (DoA) of JAK inhibition in
enzymatic assays (through kinetic experiments) or cell based assays (through cell washꢀout
experiments), to allow for sustained target modulation in vivo.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Given the successful clinical precedent of tofacitinib to modulate JAK driven disease, our
initial tactic was to increase the lipophilicity in this template to introduce metabolic instability,
with the aim to maintain or improve the significant JAK potency and selectivity typically
observed with this chemotype, while also assaying for evidence of DoA. We quickly found that a
urea linkage could be used in place of the amide bond in tofacitinib and this allowed for facile
introduction of lipophilic groups that maintained excellent JAK potency and selectivity, while
introducing metabolic instability. For example 3 (PFꢀ00956980), a useful commercially available
JAK inhibitor tool, (Table 1) maintains excellent JAK potency and has high turnover in human
14
liver microsomes (HLM). The pyrrolidine ring of 3 could be further substituted with no loss in
potency by larger groups containing phenols, which were introduced to allow for rapid phase II
ACS Paragon Plus Environment
4

Page 5 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
metabolism of compounds, thus providing a second route of metabolism. For example 4 (Table
1) has high turnover in both HLM and HLM UGT assays and increased JAK enzyme potency
relative to compound 3.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 1. Potency and inꢀvitro metabolic stability data for tofacitinib analogues 3 and 4.
b
c
No.
JAK1
IC50
JAK2
IC50
(nM)
JAK3
IC50
(nM)
Tyk2
IC50
(nM)
HLM
HLM UGT
a
(ꢀL/min/mg (ꢀL/min/mg
(nM)
protein)
82
protein)
<3.5
3
4
37.5
2.3
123.6
21.2
139.3
34.1
1632.2
958.6
131
>200
a
Compounds were routinely assayed using Caliper microfluidic assays for their potency of
inhibition of phosphorylation of a peptide substrate by the appropriate JAK enzyme at 1 mM
ATP concentration. All IC values reported represent geometric means of a minimum of three
50
b
determinations. Compounds were routinely assessed for metabolic stability in human liver
c
microsomes (HLM). Compounds were assessed for phase II glucuronidation in HLM by
including the microsomal activator Brij58 and cofactor uridine diphosphate glucuronic acid
(UDPGA) and excluding the cytochrome P450 dependent cofactor NADPH.
Urea 4 was shown to have a suitable crystalline solid form and acceptable thermodynamic
solubility (20 ꢀg/mL) to meet our initial compound criteria; however in a number of different
assay formats we could not demonstrate evidence for significant DoA with this compound, nor
with any others within this chemotype. To investigate enzyme inhibition kinetics we used a
15
continuous JAK3 dilution recovery assay using microfluidic mobility shift assay technology.
Both tofacitinib 1 and urea 3 showed no evidence of slowꢀoffset from JAK3 in this assay, while
ꢀ1
4
showed an offꢀset rate of 0.024 min that translated to a T of approximately 28 minutes. In a
1/2
ACS Paragon Plus Environment
5

Journal of Medicinal Chemistry
Page 6 of 62
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
PBMC cellꢀbased assay we could measure inhibition of the cellular pSTAT5 response to ILꢀ2
stimulation with 3 and 4 giving IC values of 27 nM and 343 nM respectively. Using this assay,
50
we investigated cellular DoA by incubating compounds for 1 hour at 10ꢀfold the IC , washing
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
cells extensively and then monitoring recovery of the cellular pSTAT5 response to ILꢀ2
stimulation over time. Both 3 and 4 showed no sign of pSTAT5 inhibition in this assay
immediately following cell washing. Assessing this data together, we did not view these
compounds as having significant potential to show clinical duration of action as inhaled or
topical agents in vivo, and thus we sought an alternative chemical series.
Examination of overlaid Xꢀray crystal structures of tofacitinib 1 and 4 bound to JAK2 (Figure
2
) shows that they both adopt the same ‘typeꢀ1’ inhibitor binding mode with the methionine
1
6
gatekeeper residue flanking the back of the pocket in both cases. The larger molecule 4 shows
some displacement of the glycine rich Pꢀloop of the kinase relative to tofacitinib, although as
demonstrated by our data, this binding mode does not appear to result in any significant kinetic
advantage, likely due to high flexibility of the protein in this region.
ACS Paragon Plus Environment
6

Page 7 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2. Overlaid Xꢀray coꢀcrystal structures illustrating binding mode of tofacitinib (1) (PDB
entry 3LXK) and 4 (PDB entry 5TQ7) bound to JAK2.
It has been well documented by us and others that certain kinase inhibitor binding modes
appear to be correlated to slowꢀoffset binding kinetics, in particular the ‘typeꢀ2’ inhibitor binding
1
3c, 16
mode.
This binding mode results in a portion of the inhibitor binding at a remote distance
from the hinge binding motif, behind the gatekeeper amino acid, into an induced ‘back pocket’
of the kinase, resulting in significant movement of the DFG portion of the activation loop and
stabilization of an inactive form of the kinase. In order to find a JAK inhibitor series with
evidence of DoA in enzyme kinetics and cell based experiments, we sought to leverage our
ACS Paragon Plus Environment
7

Journal of Medicinal Chemistry
Page 8 of 62
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
internal database of kinase Xꢀray coꢀcrystal structures, along with kinase coꢀcrystal structures
1
7
available from the PDB, to identify a subset of kinase inhibitors with the potential to bind
behind the gatekeeper residue, and to screen this subset for potency against the JAK kinases.
Using approximately 2500 inꢀhouse kinase inhibitor crystal structures, together with
approximately 1500 additional structures available from the PDB, we computationally collected
and aligned all structures into a common orientation. A computational triage was then performed
to identify any inhibitor templates that were predicted to bind to the hinge region of the kinase,
and within 2Å of the methionine gatekeeper residue in JAK3. We reasoned that any compounds
found to be potent JAK inhibitors from this selection would likely bind in a manner necessitating
a conformational change in the kinase, placing binding elements in the back pocket of the kinase,
potentially leading to slower offset kinetics. This screening effort resulted in the discovery that
the indazoleꢀphenol 5 (Table 2), a compound in our corporate collection from a previous
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
8
research program, is a potent panꢀJAK inhibitor with IC values of <1 nM to 10 nM against
50
the JAK kinase family at 1 mM ATP concentration. 5 displayed an enzymatic kinetic offꢀset rate
ꢀ1
of 0.006 min that translated to a T of approximately 114 minutes for JAK3. It also showed a
1
/2
significant time dependent inhibition of the cellular pSTAT5 response to ILꢀ2 stimulation
following compound incubation and cell washing, with a T1/2 of 6.5 hours.
ACS Paragon Plus Environment
8

Page 9 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 2. Profile of indazoleꢀphenol 5.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
b
c
No.
JAK1
IC50
JAK2
IC50
(nM)
JAK3
IC50
(nM)
Tyk2
IC50
(nM)
HLM
HLM UGT
a
(ꢀL/min/mg (ꢀL/min/mg
(nM)
protein)
35
protein)
49
5
0.4
2.2
8.3
9.1
a
Compounds were routinely assayed using Caliper microfluidic assays for their potency of
inhibition of phosphorylation of a peptide substrate by the appropriate JAK enzyme at 1 mM
ATP concentration. All IC values reported represent geometric means of a minimum of three
determinations. Compounds were routinely assessed for metabolic stability in human liver
microsomes (HLM). Compounds were assessed for phase II glucuronidation in HLM by
5
0
b
c
including the microsomal activator Brij58 and cofactor uridine diphosphate glucuronic acid
(UDPGA) and excluding the cytochrome P450 dependent cofactor NADPH.
We were pleased to find that an Xꢀray crystal structure of 5 bound to JAK2 (Figure 3) showed
a novel binding mode for the JAK kinases, compared to the previously observed typeꢀ1 binding
mode observed with tofacitinib. The methionine gatekeeper residue moved approximately 4.3Å
from the known typeꢀ1 binding mode orientation, allowing the phenolic portion of the molecule
to bind behind the gatekeeper in an induced back pocket. The indazoleꢀbenzimidazole portion of
the molecule forms three hydrogen bonding interactions with the hinge region of the kinase,
while the phenol group makes two hydrogen bonding interactions in the backꢀpocket, with the αꢀ
Glu (E898) of the Cꢀhelix and the backbone NH of the Phe (F995) of the DFG motif. The ethyl
group of the phenol ring occupies the lipophilic space left by movement of the methionine side
chain and helps induce a twist across the biaryl bond to the indazole ring. It is interesting to note
ACS Paragon Plus Environment
9

Journal of Medicinal Chemistry
Page 10 of 62
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
that despite significant binding in the backpocket, and the movement of the gatekeeper
methionine, the DFG motif has not moved significantly and that this structure does not represent
a typeꢀ2 binding mode. Similar binding modes have been described previously for other kinases
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
19
as type 1.5. Based on this binding mode, combined with evidence of slower offꢀset kinetics,
and more significant cellular DoA, we chose to pursue 5 as an alternative lead for the program.
Figure 3. Xꢀray coꢀcrystal structure of 5 bound to JAK2 (PDB entry 5TQ3) overlaid on protein
from coꢀcrystal with 4 to illustrate movement of the methionine gatekeeper residue. Numbers
correspond to hydrogen bonding distances in Angstroms.
Considering 5 as a lead, it already fulfilled our JAK enzyme potency requirements and
displayed significant DoA in both enzyme and cellular assays. However, the aqueous solubility
ACS Paragon Plus Environment
10

Page 11 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
of 5 was very poor, to the extent that an accurate measurement could not be determined. In
addition, despite a significant increase in lipophilicity compared to the tofacitinib series, turnover
in HLM was only modest, where high metabolism was desired. It was however of significant
promise that the molecule already contained a phenol group which brought modest turnover in an
in vitro assay (HLM UGT) which contained cofactors for phase 2 glucuronidation. Given the
apparent importance of the hydrogen bonding interactions of the phenol group in binding to
JAK, we postulated that any phase 2 metabolites of the phenol group (for example the
glucuronide or sulfate) would be largely inactive as JAK inhibitors, thus greatly reducing any
liability from systemic JAK inhibition of potential circulating metabolites. The crossꢀkinase
promiscuity (Figure 4) appeared less optimal for 5, compared to the tofacitinib chemotype, as it
inhibited 9 of 35 additional kinases at the ATP Km concentration in a screening panel with >50%
inhibition at 1 ꢀM compound concentration, although only 1 of these (ZC1/HGK – also known
as MAP4K4) was inhibited at >85%.
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. Heatmap showing kinase inhibition profile of 5 dosed at 1ꢀM against a panel of 36
kinases at the ATP Km concentration. Colors correspond to % inhibition of a given kinase as per
the color legend shown.
Initial structure activity relationships (Table 3) were investigated around the phenol ring in
order to probe for improvements in JAK kinase potency and selectivity, as well as to increase the
ACS Paragon Plus Environment
11

Journal of Medicinal Chemistry
Page 12 of 62
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
overall metabolic rate. We employed enzymatic assays for each JAK kinase at 1mM ATP
concentration rather than at the Km concentrations to gain a better understanding of compound
20
potency and selectivity under more physiologically relevant conditions. In addition assays in
primary PBMC cells were used to measure inhibition of the pSTAT5 response to ILꢀ15 for
cellular DoA, as well as inhibition of the IFNγ response to ILꢀ2 for potency assessment.
The location of the phenol OH group was shown to be preferred in the para position to the
biaryl linkage, as the equivalent meta phenol 6 was a significantly weaker inhibitor against JAK
enzymes and in the PBMC cellular assay. This is consistent with the observed Xꢀray coꢀcrystal
structure of 5 where the two hydrogen bonds to the para phenolic OH group appear to be close to
optimal distances. Not surprisingly, removal of the OH group or methylation to provide a para
methoxyl group resulted in significant loss in potency (data not shown). Increasing the size of the
substituent on the phenol ring to either an isopropyl or nꢀpropyl group (7 and 8) resulted in a
drop in both enzyme and cellular potency, while increasing the size of this group still further led
to even more significant loss in potency (data now shown). The presence of the ethyl group was
found not to be an essential feature for good potency against JAK kinases, as a number of
compounds without the ethyl group showed good enzyme and cellular potency. However, these
compounds all had far poorer kinase selectivity (data now shown). This suggests that an ethyl
substituent is close to the optimal size for a group at this position to maintain good potency while
also providing the benefit of improved kinase selectivity. Consistent with this observation, a
methoxyl group was also well tolerated at this position, as in 9, albeit with some loss in potency
compared to 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
ACS Paragon Plus Environment
12

Page 13 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 3. SAR around phenol ring: JAK kinase potency and selected cellular potency data.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
a
b
Enzyme IC at 1mM ATP (nM)
Cellular IC (nM)
5
0
50
No.
R
JAK1
JAK2
JAK3
TYK2
9.1
PBMC
5
0.4
2.2
8.3
10.5
6
7
68.6
26.8
100.8
79.6
493.1
293.4
291.2
392.3
572.7
250.3
8
9
9.7
17.3
166.0
437.5
75.4
6.1
nt
5.4
nt
43.5
nt
102.1
nt
57.8
204.8
47.9
1
0
11
0.8
7.9
16.7
15.3
ACS Paragon Plus Environment
13

Journal of Medicinal Chemistry
Page 14 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
a
Compounds were routinely assayed using Caliper microfluidic assays for their potency of
inhibition of phosphorylation of a peptide substrate by the appropriate JAK enzyme at 1 mM
ATP concentration. All IC values reported represent geometric means of a minimum of three
determinations. Cellular potency was routinely assayed in peripheral blood mononuclear cells
(
interleukin 2 (IL2). All IC values reported represent geometric means of a minimum of three
determinations. nt denotes ‘not tested’.
50
b
PBMCs) measuring inhibition of interferon gamma (IFNγ) release following stimulation with
50
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
At this point we also explored adding groups such as chloro, cyano and fluoro adjacent to the
13b
phenolic OH group in an attempt to modulate its glucuronidation rate by acidification.
However, to maintain good JAK potency, only fluoro groups were found to be well tolerated in
these positions (10 and 11), with only a small loss in potency compared to 5. This is again
consistent with the observed binding mode as these positions lie in very close proximity to the
protein surface, allowing only for a small atom to be substituted. It appears as though the
position para to the ethyl group has slightly more space in the Xꢀray structure, which is reflected
in the better cell potency observed for 11 over 10. Interestingly, the fluoro substitution para to
the ethyl group also resulted in improved kinase selectivity, as 11 inhibited only 3 kinases of a
panel of 36 at >50% at 1 ꢀM concentration, with only JAK3 (other JAK family members were
not present in the screening subset) in the panel being significantly inhibited at >85% as
expected (kinase heat maps for a number of key analogues can be found in the supporting
information). Gratifyingly, this change also dramatically increased the glucuronidation rate for
1
1, which showed a turnover rate in the HLM UGT assay of >360 ꢀL/min/mg protein. This was
not the case for 10, which showed no increase in glucuronidation rate (46 ꢀL/min/mg protein)
versus 5, despite the two phenols 10 and 11 having very similar calculated pKa values. Clearly
structural effects around the phenol as well as acidification are important contributing factors for
overall glucuronidation rates. Fluoroꢀphenol 11 also showed good DoA in our cellular washꢀout
assay, inhibiting the pSTAT5 response with a T1/2 of 3.5 hours, consistent with the earlier
ACS Paragon Plus Environment
14

Page 15 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
observation for compounds in this series. However, aqueous solubility was again shown to be
low and an accurate value could not be measured in a thermodynamic solubility assay. In
addition, we could not formulate these molecules successfully for in vivo efficacy evaluation
when attempting to use solution dosing via the inhaled route. While indazoleꢀbenzimidazoles
such as 11 showed some promise to meet our criteria, they were clearly not soluble enough to be
considered for dosing via the inhaled route, although useful SAR information had been gained.
Moving forward we needed to address the poor solubility while maintaining the excellent
potency and encouraging DoA of this series.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
In an attempt to improve compound solubility, while maintaining an otherwise favorable
2
1
profile, we sought to reduce the overall aromatic ring count in our series. SAR investigations in
this series were carried out using the nonꢀfluorinated phenol template. Truncating the
benzimidazole back to imidazole 12 (Table 4) resulted in no loss of potency in both kinase
enzymes and cell based assays. Encouragingly, in addition to excellent potency, we could also
see an improvement in thermodynamic solubility (4.1 ꢀM) measured from a crystalline solid
sample of 12 in water buffered at pH 7.4. Not surprisingly, truncating further to remove the
imidazole completely (R=H) resulted in significant loss of potency given that a key Hꢀbonding
interaction with the hinge motif of the kinase had been removed (data not shown). Alternative
heterocycles that maintained an Hꢀbonding interaction with the hinge motif, either through an Nꢀ
H group or aromatic CꢀH group, in general showed good enzyme inhibition potency, but a
reduced cellular potency relative to imidazole 12. Further truncation of the imidazole to the
amide functionality, as in primary amide 13 and methyl amide 14, maintained the excellent
potency profile while providing a further improvement in solubility (15 ꢀM and 14 ꢀM
respectively measured from crystalline solid samples in water buffered at pH 7.4). Methyl amide
ACS Paragon Plus Environment
15

Journal of Medicinal Chemistry
Page 16 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
4 in particular showed an excellent kinase selectivity profile with only JAK3 out of a panel of
36 kinases inhibited at >50% at 1 ꢀM concentration. An Xꢀray coꢀcrystal structure of primary
amide 13 binding to JAK2 (Figure 5) showed the expected binding mode for the series, with
three key hydrogen bonds in a donorꢀacceptorꢀdonor motif from the indazole NꢀH and lone pair
and NꢀH of the amide to the hinge region of the kinase, as well as the two hydrogen bonds from
the phenol group to the back pocket residues as noted previously. Not surprisingly, the dimethyl
amide 15 showed significant loss of potency, consistent with the requirement for the amide NꢀH
hydrogen bonding interaction with the kinase. However, despite adopting the same binding mode
as the indazoleꢀbenzimidazole 5, none of these potent and more soluble truncated compounds
showed a significant DoA in our cellular assay, nor significant slow offꢀset in the kinase kinetic
assay. Subsequent compounds that maintained good DoA (vide infra) suggested that additional
interactions between the ligand and kinase protein have been lost with this series of smaller
amide compounds, leading to the loss of DoA.
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 4. Replacements for benzimidazole: JAK kinase potency, cellular potency and solubility
data.
a
Cellular IC50
Enzyme IC at 1mM ATP (nM)
b
5
0
(nM)
No.
R
JAK1
JAK2
JAK3
Tyk2
PBMC
ACS Paragon Plus Environment
16

Page 17 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
5
0.4
2.2
8.3
9.1
10.5
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
2
3
0.9
1.7
0.8
4.2
0.8
4.3
4.6
10.4
19.1
1
54.5
14
7.5
40.9
125.0
796.2
9.1
1
5
6
2756.5 >10000 >10000 >10000
>10000
277.8
1
60.0
58.8
418.3
815.9
a
Compounds were routinely assayed using Caliper microfluidic assays for their potency of
inhibition of phosphorylation of a peptide substrate by the appropriate JAK enzyme at 1 mM
ATP concentration. All IC values reported represent geometric means of a minimum of three
determinations. Cellular potency was routinely assayed in peripheral blood mononuclear cells
(PBMCs) measuring inhibition of interferon gamma (IFNγ) release following stimulation with
interleukin 2 (IL2). All IC values reported represent geometric means of a minimum of three
5
0
b
50
determinations.
We reasoned that further binding potency and improved DoA could be achieved by expanding
the amide series into the pocket occupied by the benzimidazole group in our initial lead 5, while
maintaining improved solubility. Of a number of larger amide analogues prepared, we found that
only those made from aromatic amines, for example 16 (Table 4), maintained any reasonable
kinase inhibition potency. Upon inspection of the Xꢀray crystal structure of 16 bound to JAK2
(
Figure 5), it is clear that the region of the kinase where the amide heterocyclic group is bound is
restricted in size by a triad of surrounding amino acids, Leuꢀ855, Tyrꢀ931 and Glyꢀ935. We
ACS Paragon Plus Environment
17

Journal of Medicinal Chemistry
Page 18 of 62
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
found that amides substituted with a methyl group (as in 14) or a planar group (as in 16) were
tolerated in this position, allowing access through the LeuꢀTyrꢀGly triad of amino acids to the
more open space beyond the hinge region. Heterocyclic amide 16 showed reasonable JAK
inhibition potency in both enzymatic and cellular assays, as well as significant DoA in our
cellular washꢀout assay, with a T1/2 of approximately 4.5 hours. In addition, kinase selectivity for
this compound was good, with only 2 kinases (besides JAK3) in a panel of 36 kinases showing
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
>
50% inhibition at 1 ꢀM concentration. Solubility for amide 16, along with most other larger
heterocyclic amides prepared, although improved over 5, was however found to be poor (0.5
M). Even though we had reduced the aromatic ring count relative to 5, amide 16 is still a rather
ꢀ
2
2
planar molecule where crystal packing forces in the solid state are likely to be significant. In
addition, chemical stability towards hydrolysis was found to be less than optimal for the majority
of these heterocyclic amides over a relatively short time (>10% degradation seen after 1ꢀ2 days
at pH 8.0 at 40 ˚C), raising concerns for potential solid form stability and formulation problems
in this series. Taken together these data prompted us to pursue a related but ultimately more
successful subꢀseries as described below.
ACS Paragon Plus Environment
18

Page 19 of 62
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 5. Crystal structures of amides 13 (upper picture, PDB entry 5TQ4) and 16 (lower
picture, PDB entry 5TQ5) bound to JAK2. The triad of amino acids restricting exit of more
sterically demanding amides from the binding pocket are illustrated in cyan. Numbers
correspond to hydrogen bonding distances in Angstroms
ACS Paragon Plus Environment
19

Journal of Medicinal Chemistry
Page 20 of 62
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
We pursued an alternative tactic to improving solubility, while maintaining the overall
3
molecular size and similar shape of the indazole benzimidazole lead, by increasing the sp
carbon count in our molecule through saturation of the second aromatic ring of 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
23
benzimidazole system. The imidazoꢀpiperidine motif 17 (Figure 6) serves this purpose,
maintaining the donorꢀacceptorꢀdonor motif required for hinge binding, as well as the planarity
required to maintain binding through the LeuꢀTyrꢀGly triad of the kinase. In addition, the amine
functionality of the piperidine allowed us to further build out from this portion of the molecule to
increase binding affinity and DoA.
3
Figure 6. Benzimidazole to imidazoꢀpiperidine transformation. Increasing saturation and sp
count in our lead series.
Our SAR exploration was carried out using both the nonꢀfluorinated and fluorinated (at R2)
phenol templates (Table 5). The imidazoꢀpiperidine was shown to be an effective replacement
for the benzimidazole group in original lead 5, with amine 18 showing good enzyme and cell
potency. Both amide 19 and sulfonamide 20 groups could be prepared with no loss in potency,
although the amides typically showed greater cellular potency. This synthetic strategy allowed
for a wide range of substituents to be explored, enabling modulation of various properties that
could increase HLM and to explore effects on potency, kinase selectivity and DoA. The methyl
and acetamide analogues (18 and 19) showed improved solubility over the earlier
ACS Paragon Plus Environment
20

Page 21 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
benzimidazoles as expected (10ꢀ20 ꢀM). However, their kinase selectivity profiles were
generally much poorer, with at least 10 of a panel of 36 kinases showing >85% inhibition for
each of these compounds. Larger substituents, such as benzyl 21 and benzamides 22 and 2,
maintained excellent enzymatic and cellular JAK potency, while also introducing significantly
higher HLM turnover, in keeping with their increased lipophilicity. We were pleased to find that,
as previously discovered, introduction of a fluorine atom para to the ethyl group greatly
increased the glucuronidation rate for these compounds. As shown for the fluorinated
piperidinylꢀpyrazine 2, the turnover rate in the HLM UGT assay is 129 ꢀL/min/mg protein,
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
compared with 17 ꢀL/min/mg protein for the nonꢀfluorinated equivalent (compound not shown).
These compounds with larger substituents on the piperidine ring also showed significant cellular
DoA, typically with T of >6 hours.
1
/2
Table 5. Imidazoꢀpiperidine replacement for benzimidazole: JAK kinase potency and selected
cellular potency data.
Cellular
Μicrosomal
clearance assays
(ꢀL/min/mg protein)
Enzyme IC at 1mM ATP
5
0
a
IC50
(nM)
b
(nM)
JAK JAK JAK TYK
HLM
UGT
c
No.
R1
R2
H
PBMC
18.6
HLM
56
d
1
2
3
2
1
8
2.1
6.4
2.9
20.1
<1.9
ACS Paragon Plus Environment
21

Journal of Medicinal Chemistry
Page 22 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
9
0
H
0.5
3.3
5.6
7.1
1.3
22.7
25.1
4.5
42
5.5
2
H
F
F
1.3
1.8
39.1
30.4
35.0
18
75
48
7.3
>360
37
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
21
15.9 31.0 57.5
2
2
<0.1
0.3
1.7
0.8
2
F
2.2
23.1 59.9 29.7
70.3
101
129
a
Compounds were routinely assayed using Caliper microfluidic assays for their potency of
inhibition of phosphorylation of a peptide substrate by the appropriate JAK enzyme at 1 mM
ATP concentration. All IC values reported represent geometric means of a minimum of three
determinations. Cellular potency was routinely assayed in peripheral blood mononuclear cells
PBMCs) measuring inhibition of interferon gamma (IFNγ) release following stimulation with
interleukin 2 (IL2). All IC values reported represent geometric means of a minimum of three
determinations Compounds were routinely assessed for metabolic stability in human liver
microsomes (HLM). Compounds were assessed for phase II glucuronidation in HLM by
including the microsomal activator Brij58 and cofactor uridine diphosphate glucuronic acid
UDPGA) and excluding the cytochrome P450 dependent cofactor NADPH.
50
b
(
50
c
d
(
Fluorinated piperidinylꢀpyrazine 2 met all of our inꢀvitro assay criteria and was selected for
more advanced studies. Overall it demonstrated good JAK potency, kinase selectivity, metabolic
turnover through two different routes of metabolism, as well as sufficient solubility in various
formulations for in vivo testing, even though thermodynamic solubility in aqueous solution was
low (~0.5 ꢀM). In addition, a crystalline solid form for 2 was found with excellent chemical and
physical stability. Most noteworthy, 2 displayed a very long cellular DoA with a T1/2 of 9 hours
in the PBMC cellular washꢀout assay. Profiling 2 against a panel of 36 kinases at 1 ꢁM, a
concentration that significantly exceeds the JAK cellular and enzymatic IC ’s at 1mM ATP,
5
0
ACS Paragon Plus Environment
22

Page 23 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
demonstrated that it has suitable selectivity over nonꢀJAK kinases as only JAK3 was inhibited
more than 85% at this concentration (Figure 7).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 7. Heatmap showing kinase inhibition profile of 2 dosed at 1ꢀM against a panel of 36
kinases. Colors correspond to % inhibition of a given kinase as per the color legend shown.
An Xꢀray crystal structure of 2 bound to JAK2 (Figure 8) showed the expected binding mode
as described previously, with some additional interactions observed with two of the residues
flanking the front pocket. Firstly, the piperidinylꢀpyrazine heterocyclic moiety is wrapped
alongside and engaged in a πꢀstacking interaction with the amide carbonyl groups of Leuꢀ855 in
the glycine rich Pꢀloop, which also bridges through water to interact with the imidazole ring.
Secondly, the amide carbonyl in 2 bridges via another conserved water to interact with Tyrꢀ931.
This additional network of binding interactions might explain the longer DoA observed with this
larger class of compound within this series.
ACS Paragon Plus Environment
23

Journal of Medicinal Chemistry
Page 24 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 8. Xꢀray coꢀcrystal structure of 2 bound to JAK2 (PDB entry 5TQ8).
JAK Inhibitory Profile of 2 in Human Whole Blood (HWB)
In order to test functional potency and to provide insight into the JAK inhibitory profile of 2,
studies in human whole blood investigated inhibition of STAT phosphorylation induced by a
selection of cytokines known to signal through various combinations of JAK enzymes (Table 6).
In each assay the inhibition by 2 of STAT phosphorylation was measured in a specific subꢀ
population of white blood cells and IC50 values were determined. Free IC50 values were
calculated by taking into account plasma protein binding, where the fraction unbound in plasma
is 0.001. These data confirmed that 2 is a functional panꢀJAK inhibitor in human whole blood.
ACS Paragon Plus Environment
24

Page 25 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Table 6. Inhibitory potency of 2 in human whole blood versus various stimulatory cytokines
a
Cytokine
JAK dimer
White blood cells
pSTAT IC (ꢀM)
Free IC (nM)
5
0
50
IFNα
ILꢀ23
ILꢀ4
JAK1/TYK2
JAK2/TYK2
JAK1/JAK3
JAK1/JAK2
Lymphocytes
Lymphocytes
Monocytes
Monocytes
Monocytes
pSTAT3
pSTAT3
pSTAT6
pSTAT3
pSTAT5
0.62
4.2
1.8
2.1
5.2
0.62
4.2
1.8
2.1
5.2
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ILꢀ6
GMꢀCSF JAK2/JAK2
a
All IC values reported represent geometric means of a minimum of three determinations.
5
0
Pharmaceutical Properties of 2
Fluorinated piperidinylꢀpyrazine 2 is a lipophilic compound with a cLogP of 4.4 and LogD of
3.9 (measured by a shake flask method). The molecular weight of the compound is 566 Da and
2
the topological polar surface area (TPSA) is calculated to be 126 Å . The compound contains
both a weakly basic imidazole group with a measured pK of 4.5 and a weakly acidic
a
fluorophenol group with a measured pK of 8.7. Following extensive screening efforts, including
a
efforts to generate various salt forms of the compound, a stable neutral nonꢀsolvated crystalline
form was found that could be crystallized directly from the final reaction step to provide a means
of isolating the compound with excellent purity (Figure 9, left hand panel). This form has very
low hygroscopicity and is chemically stable both at room temperature and at 40 ˚C with 75%
relative humidity for up to 6 weeks, underwriting its long term stability. The material is also
stable to jet milling to provide solid material with regular and consistent particle size in the
2
4
respirable range (~2 ꢀm). The resulting micronized material (Figure 9, right hand panel) has
been found to be chemically and physically stable when blended with lactose for up to four
weeks, demonstrating its suitability for use in a dry powder inhalation device.
ACS Paragon Plus Environment
25

Journal of Medicinal Chemistry
Page 26 of 62
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 9. SEM images of solid 2 following crystallization (left hand panel) and subsequent
micronization (right hand panel)
In addition this crystalline material was also amenable to preparation as a 4% solution in a base
formulation typically used in topical applications to the skin as shown below (Table 7).
Table 7. Components of 4% dermal formulation of 2
Component
2
% wt/wt
4
Propylene Glycol
30
Polyethylene Glycol
3
0
4
00
Dimethyl Isosorbide
Glycerin
10
25.9
0.1
BHA
In vitro metabolism and pharmacokinetic properties of 2
Following in vitro incubation of 2 in human hepatocytes, the major metabolites observed were
produced by direct glucuronidation and P450 metabolism, consistent with the presence of the
fluorinated phenol group functioning as an effective handle for phase II conjugation. As
discussed previously in relation to the binding mode of 2, these phenol conjugated metabolites
ACS Paragon Plus Environment
26

Page 27 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
are expected to show no inhibition of JAK kinases once formed. This metabolic profile was
consistent across rat and dog in vitro hepatocyte and microsomal metabolism studies.
The pharmacokinetic properties of 2 were assessed in vivo and were found to be consistent
with its high lipophilicity and neutral physicochemistry (Table 8). Consistent with the high inꢀ
vitro intrinsic clearance values previously described 2 exhibits very high unbound clearance in
the rat and dog and has an oral bioavailability of <5% following oral solution administration to
the rat.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 8. In Vivo Rat and Dog plasma pharmacokinetic data of 2 following intravenous (i.v.) and
oral (p.o.) dosing
c
b
Species
Dose
Cl
Clunbound
Vss
Terminal
Oral
b
(
mg/Kg) (mL/min/Kg) (mL/min/Kg) (L/Kg) T1/2 (h) Bioavailability
Rat i.v.
Rat p.o.
Dog i.v.
0.5
1
48
ꢀ
>48,000
ꢀ
0.8
ꢀ
2.1
ꢀ
ꢀ
<5%
ꢀ
a
0.1
18
>18,000
1.0
2.0
a
b
c
30 min infusion; parameters from i.v. PK profile; plasma protein binding =>99.9%
Based on these human in vitro data and the in vivo rat and dog pharmacokinetic data, 2 was
predicted to be cleared in humans by glucuronidation and CYP3A4 metabolism, exhibiting low
levels of unbound circulating parent following inhaled or dermal delivery due to high unbound
clearance. Due to the predicted negligible oral absorption and lack of observed circulating
metabolites in animal Pk studies the potential for significant levels of circulating active
metabolites in human was deemed to be low risk. Therefore, following topical dosing to either
the lung or skin, 2 demonstrated suitable pharmacokinetic properties to provide high confidence
that it would demonstrate an organ targeted effect, with negligible systemic exposure of parent
ACS Paragon Plus Environment
27

Journal of Medicinal Chemistry
Page 28 of 62
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
compound and no active circulating metabolites. Based on the multiple routes of metabolism
(P450 and glucironidation) and projected low levels of unbound exposure in human the potential
for 2 to be either a victim or perpetrator of a drugꢀdrug interaction was also deemed to be low.
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
Biological in vivo activity of 2
The in vivo activity of 2 dosed as a solution via the intraꢀtracheal (i.t.) route was demonstrated
in a mouse pharmacodynamic model of ILꢀ6 induced pSTAT3 response in the lungs, mediated
via the JAK1/JAK2 mechanism (Figure 10). 2 caused a doseꢀrelated inhibition of the ILꢀ6ꢀ
elicited increase in lung pSTAT3, giving an ED of approximately 3 ꢁg/animal. For comparison,
5
0
the inhaled corticosteroid, fluticasone propionate, demonstrates an ED of between 3 ꢁg/animal
50
and 30 ꢁg/animal across a variety of mouse models when dosed i.t.
2
500
0
0
2000
5
1
500
000
500
0
1
1
00
150
0- 1. 1
01
11 0
1 20 0
Compound 2 (ꢀg per animal)
Compound 2 (ꢀg per animal)
Figure 10. Effects of intraꢀtracheal administration of 2 on ILꢀ6 (i.t.) induced increases in lung
phosphorylated STAT3 in the mouse: Left hand panel represents absolute quantity of
ACS Paragon Plus Environment
28

Page 29 of 62
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
phosphorylated pSTAT in mouse lung; Right hand panel represents the percentage inhibition ±
95% CI calculated from the difference in geometric means between vehicle/ILꢀ6ꢀtreated and
vehicle/salineꢀtreated groups.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
In vivo activity was also assessed via topical application of 2 to the skin of a mouse ear in a
dermal model of ILꢀ23 induced inflammation (Figure 11). ILꢀ23 is known to be a major driver of
2
5
psoriasis pathology mediated via the JAKꢀSTAT pathway. For mice ears treated topically with
2
as a 4% solution, a significant reduction of ear swelling (48%) was observed over a period of
1 days following multiple ILꢀ23 injections in the ear, consistent with inhibition of the
1
JAK2/TYK2 pathway. For compound treated mice, swelling typically occurred to less than 50
m, while ears of control mice receiving ILꢀ23 injections with no compound treatment typically
ꢀ
swelled to over 100 ꢀm.
Topical Vehicle/ILꢀ23
(4% solution in Vehicle)/ILꢀ23
Topical Vehicle/Saline
2
150
1
00
50
0
48%
0
1
2
3
4
5
6
8
7
9
10 11
Days of study
muILꢀ23 challenge by interdermal injection
ACS Paragon Plus Environment
29