Design and synthesis of inhaled p38 inhibitors for the treatment of chronic obstructive pulmonary disease

Article abstract of DOI:10.1021/jm200677b

This paper describes the identification and optimization of a novel series of DFG-out binding p38 inhibitors as inhaled agents for the treatment of chronic obstructive pulmonary disease. Structure based drug design and "inhalation by design" principles have been applied to the optimization of the lead series exemplied by compound 1a. Analogues have been designed to be potent and selective for p38, with an emphasis on slow enzyme dissociation kinetics to deliver prolonged lung p38 inhibition. Pharmacokinetic properties were tuned with high intrinsic clearance and low oral bioavailability in mind, to minimize systemic exposure and reduce systemically driven adverse events. High CYP mediated clearance and glucuronidation were targeted to achieve high intrinsic clearance coupled with multiple routes of clearance to minimize drug-drug interactions. Furthermore, pharmaceutical properties such as stability, crystallinity, and solubility were considered to ensure compatibility with a dry powder inhaler. 1ab (PF-03715455) was subsequently identified as a clinical candidate from this series with efficacy and safety profiles confirming its potential as an inhaled agent for the treatment of COPD.

Full text of DOI:10.1021/jm200677b

ARTICLE
pubs.acs.org/jmc
Design and Synthesis of Inhaled p38 Inhibitors for the Treatment of
Chronic Obstructive Pulmonary Disease†,‡
David S. Millan,*,§ Mark E. Bunnage,§ Jane L. Burrows,∞ Kenneth J. Butcher,§ Peter G. Dodd,§
Timothy J. Evans,§ David A. Fairman,^ Samantha J. Hughes,§ Iain C. Kilty,|| Arnaud Lemaitre,§
Russell A. Lewthwaite,§ Axel Mahnke,# John P. Mathias,§ James Philip,|| Robert T. Smith,§
Mark H. Stefaniak,§ Michael Yeadon,|| and Christopher Phillips§
§Worldwide Medicinal Chemistry, Allergy and Respiratory Research Unit, ^Department of Pharmacokinetics, Dynamics, and
Metabolism, #Department of Drug Safety, and ∞Department of Pharmaceutical Sciences, Pfizer Global Research and Development,
Sandwich Laboratories, Ramsgate Road, Sandwich, Kent, CT13 9NJ, U.K.
S Supporting Information
b
ABSTRACT: This paper describes the identification and
optimization of a novel series of DFG-out binding p38 inhibi-
tors as inhaled agents for the treatment of chronic obstructive
pulmonary disease. Structure based drug design and “inhalation
by design” principles have been applied to the optimization of
the lead series exemplied by compound 1a. Analogues have
been designed to be potent and selective for p38, with an
emphasis on slow enzyme dissociation kinetics to deliver
prolonged lung p38 inhibition. Pharmacokinetic properties were tuned with high intrinsic clearance and low oral bioavailability
in mind, to minimize systemic exposure and reduce systemically driven adverse events. High CYP mediated clearance and
glucuronidation were targeted to achieve high intrinsic clearance coupled with multiple routes of clearance to minimize drugꢀdrug
interactions. Furthermore, pharmaceutical properties such as stability, crystallinity, and solubility were considered to ensure
compatibility with a dry powder inhaler. 1ab (PF-03715455) was subsequently identified as a clinical candidate from this series with
efficacy and safety profiles confirming its potential as an inhaled agent for the treatment of COPD.
Chart 1
’ INTRODUCTION
p38 Mitogen activated kinase (MAP) is a ubiquitous target in
the research based pharmaceutical industry. It plays a central role
in the regulation of biosynthesis and actions of key proinflam-
matory mediators such as IL1-β and TNF-α. These cytokines,
and others, are important regulators of the immune response to
infection as well as cellular stress.1 Indeed biologics that bind or
remove TNF-α, such as etanercept, infliximab, and adalimumab,
have shown good efficacy in the clinic.2 Similarly, the IL-1 receptor
antagonist anakinra has proved to be efficacious in rheumatoid
arthritis (RA), suggesting that modulating the levels of these
proinflammatory cytokines provides clinical benefit in inflam-
matory disorders.2
rapidly after cessation of dosing.3 Elevations of liver transami-
nases were also observed for BIRB-796 in a phase I 7-day
multidose study.7 BIRB-796 progressed to a phase II study in
Crohn’s disease patients but failed to show clinical benefit.4 The
compound is no longer listed on the companies' development
pipeline, and it is assumed that it has been discontinued from
further development.8
Small molecule inhibitors of p38, such as VX-7453 and BIRB-
7964 (Chart 1), have advanced to clinical trials in inflammatory
diseases, with VX-745 demonstrating significant clinical benefit
in RA in a 12-week study. However, despite promising preclinical
and clinical efficacy with several other oral p38 inhibitors, their
progression to phase III studies and beyond has been hampered
by adverse findings, such as skin disorders, infection, and a lack of
sustained efficacy in diseases such as RA.5 The aforementioned
inhibitor VX-745 was not progressed beyond the phase II RA
trial because of CNS toxicity observed preclinically.6 Morover,
elevations in liver transaminases were observed, which declined
Perhaps more worrying is a common feature observed in many
of the clinical trials run with several different p38 inhibitors,
which could be linked to the disappointing efficacy observed.
Received: May 27, 2011
Published: September 05, 2011
r
2011 American Chemical Society
7797
dx.doi.org/10.1021/jm200677b J. Med. Chem. 2011, 54, 7797–7814
|

Journal of Medicinal Chemistry
ARTICLE
In many of these studies a drop in C-reactive protein (CRP) levels
(inflammatory biomarker) on drug treatment was observed, but
this was not sustained after 2 weeks of dosing, suggesting possible
redundancy in the p38 pathway.5 At this stage it is unclear
whether this transient effect on serum biomarkers as well as the
plateauing of clinical response is RA specific or not, and there-
fore, there has been significant effort in the pharmaceutical industry
to investigate p38 inhibition in other inflammatory indications.
Many respiratory diseases also show a significant inflammatory
component, suchaschronicobstructivepulmonarydisease(COPD)
and asthma. The global prevalence of COPD has reached epidemic
proportions and is now one of the leading causes of death in the
developed world.9 Existing treatments for COPD patients are
dominated by long acting bronchodilators, with anti-inflamma-
tory agents such as inhaled corticosteroids being largely ineffec-
tive in treating the inflammation or progression of the disease.10
There is therefore a huge unmet medical need with regard to
effective anti-inflammatory agents to treat COPD patients.
Studies have demonstrated that p38 is activated in the alveolar
macrophages of COPD patients, relative to control subjects,
suggesting a link with the disease phenotype.11 Small molecule
inhibitors of p38 have also provided supporting evidence that
modulating this pathway has the potential to show efficacy in
respiratory diseases where inflammation is a key driver. Oral
dosing of SB-239063 has been shown to reduce eosinophil
infiltration following challenge with lipopolysaccharide (LPS)
and also reduced the levels of IL-6 and MMP-9 in lavage fluid.12
Another small molecule p38 inhibitor, SD-282, was efficacious in
a mouse tobacco smoke model in which glucocorticoids were
ineffective (Chart 2).13 Inhalation as a route of delivery has also
been explored with a p38α selective oligonucleotide which has
shown efficacy when delivered to the lung in a model of asthma in
the mouse.14 Similarly, when a small molecule inhibitor of p38
(ARRY-371797, structure not disclosed) was dosed directly to the
lung, it was shown to significantly reduce levels of proinflammatory
cytokines, as well as neutrophil infiltration, in response to LPS
challenge.15
Several oral p38 inhibitors have advanced to human clinical
trials for COPD. PH-797804, losmapimod, and dilmapimod
(Chart 3) have all completed phase II trials for COPD.16 Results
from the phase II trial of PH-797804 suggest real promise.17 In a
6-week study involving 230 patients, PH-797804 showed statis-
tically significant improvements in FEV1 (forced expiratory
volume in 1 s) and clinically meaningful increases were observed
on the transitional dyspnoea index. Interestingly, CRP levels
remained low over the time course of the trial, and therefore, this
result deviates from that seen in RA with other p38 inhibitors.
However, it is reiterated that there is no credible reason why p38
pathway redundancy may not be observed in COPD, and there-
fore, it remains a significant risk. Longer clinical trials will help to
answer this question. PH-797804 was also safe and well tolerated
with p38 related adverse events such as skin rash observed in only
approximately 7% of PH-797804 treated patients. To the best of
our knowledge, data from other COPD studies, with other p38
inhibitors, have not been reported.
This promising result, as well as other preclinical studies
discussed above, has prompted interest in the topical application
of a p38 inhibitor directly to the lung of a COPD patient.
Inhalation of a p38 inhibitor represents an intriguing opportunity
to deliver powerful anti-inflammatory activity coupled with the
potential to minimize systemically driven adverse events that
may be p38 mechanism related.18,19 This paper will detail efforts
to identify a potent, selective, and safe p38 inhibitor with phar-
macokinetic and pharmaceutical properties suitable for lung
delivery. The design principles adopted follow the “inhalation
by design” philosophy outlined in a previous publication.20
’ CHEMISTRY
The synthetic strategy for preparation of target compounds
1aꢀac in this paper is based on the same general disconnection
for each analogue. This involves synthesis of the benzylamines
2aꢀh, which is then coupled with either a commercially available
or synthesized arylpyrazole 3aꢀk through a urea forming step
(Scheme 1).
Chart 2
The arylpyrazole portion of the target molecule was prepared
by cyclocondensation of an arylhydrazine with a ketonitrile. The
ketonitrile either was commercially available, as in 4,4-dimethyl-
3-oxopentanenitrile, or was prepared from the corresponding
commercially available precursors to afford 5a and 5b (Scheme 2).
The ketonitrile 5a was prepared from commercially available
ethyl 2-methyl-2-(methylthio)propionate by reaction with the
anion of acetonitrile. Similarly ketonitrile 5b was prepared by
Chart 3
7798
dx.doi.org/10.1021/jm200677b |J. Med. Chem. 2011, 54, 7797–7814

Journal of Medicinal Chemistry
ARTICLE
Scheme 1. General Retrosynthetic Strategy for Preparation of Target Compounds
Scheme 2a
available 4-amino-2-chlorophenol, by diazotization, followed by
tin chloride mediated reduction of the diazo intermediate to
furnish the desired hydrazine 4b (Scheme 3). The arylpyrazoles
3aꢀk were prepared by cyclocondensation of the ketonitrile with
the appropriate arylhydrazine, followed by deprotection and
alkylation, affording 3d and 3i, or protection as silyl ethers, affording
3f, 3h, 3j, 3k (Scheme 2).
The benzylamines 2aꢀh were synthesized by reaction of
commercially available 5-bromo-2-hydrazinylpyridine with var-
ious commercially available aliphatic and aromatic aldehydes to
afford the corresponding hydrazones 15aꢀd in good yields
(Scheme 4). The hydrazone was then reacted with iodobenzene
diacetate in a cyclo-oxidation reaction to afford the bromotria-
zolopyridines 16aꢀg. Reaction with commercially available 2-mer-
captobenzyl alcohol in a palladium catalyzed coupling afforded
the benzyl alcohols 17aꢀg. The alcohols 17aꢀg were then
converted to the benzylamines 2aꢀf by reaction with methane-
sulfonic anhydride and displacement of the corresponding
mesylate with ammonia in methanol. Finally, 2g was prepared
by debenzylation of 2d with boron tribromide.
To install the hydroxyethylene group in 2h, 2g was reacted
with di-tert-butyl dicarbonate to doubly protect the amine and
phenol to afford 18. A selective deprotection of the carbonate
was achieved with sodium carbonate to afford 19, followed by
alkylation to give 20. Deprotection of the primary amine and
removal of the THP protecting group revealed the desired coupling
partner 2h (Scheme 5).
In several instances the aldehyde was not commecially avail-
able and was therefore synthesized. The aldehyde 21a was pre-
pared by benzyl protection of the commercially available 3-chlor-
osalicylaldehyde (Scheme 6).
Aldehyde 21b was prepared by taking commercially available
methyl thiosalicylate and alkylating with (2-bromoethoxy)-tert-
butyldimethylsilane to afford 22. The ester was then reduced to
afford the alcohol 23, which was then oxidized to give the desired
aldehyde 21b (Scheme 7).
With benzylamines 2aꢀh and the arylpyrazoles 3aꢀk in hand
we were able to conduct the urea forming reaction to synthesize
target compounds1aꢀac. This was achieved by one of two methods.
We initially utilized carbonyldiimidazole as a carbonyl source to
produce 1aꢀc, 1j, 1nꢀo, 1y, and 24ꢀ30. 24ꢀ30 were then
further manipulated to furnish target compounds 1eꢀf, 1i,
1kꢀ1l, 1p, 1u in good yield. (Scheme 8). Target compound
1y did not require desilylation of the phenolic protecting group,
as it was spontaneously removed during the urea forming step.
The sulfoxides 1g, 1m, and 1q were prepared by oxidation of 1f,
1l, and 1p, respectively.
a Reagents: (a) (i) iPrNEt2, MeSO2Cl, DCM, 0 ꢀC, 2 h; (ii) MeSNa,
MeOH, reflux, 17 h, 17%; (b) NaH, CH3CN, THF, reflux, 5 h, 58ꢀ81%;
(c) arylhydrazine, conc HCl, EtOH, reflux, 6 h, 23ꢀ95%; (d) BBr3,
DCM, 0 ꢀC, 0.5 h, 18ꢀ89%; (e) THPOCH2CH2Br, K2CO3, DMF,
60 ꢀC, 4 h, 71%; (f) TBDMSCl, imidazole, DMF, room temp, 16 h,
17ꢀ93%.
reaction of the commercially available methyl 2,2-dimethyl-3-
hydroxypropionate with methanesulfonyl chloride followed by
nucleophilic displacement with the sodium salt of methanethiol,
affording 6. This was followed by reaction with the anion of
acetonitrile to afford the corresponding ketonitrile 5b.
Arylhydrazines 4a and 4b were not commercially available and
were therefore prepared from the corresponding commercially
available aryl bromides (Scheme 3). First, 5-bromo-2-chloroani-
sole was metalated with n-butyllithium followed by reaction with
di-tert-butyl diazodicarboxylate to afford the corresponding diazo
compound, which was deprotected in situ with hydrochloric acid,
affording the hydrazine hydrochloride salt 4a. 14 was prepared by
a similar procedure as above, but first isolating 7 followed by
in situ deprotection with hydrochloric acid and reaction with
commercially available 4,4-dimethyl-3-oxopentanenitrile. Alterna-
tively, the arylhydrazine 4b couldbepreparedfromthecommercially
For the synthesis of advanced lead compounds such as 1d, 1h,
1rꢀt, 1vꢀx, 1zꢀab we first formed the phenyl carbamates 31aꢀe
of the arylpyrazoles 3c, 3d, 3f, 3h, 3j, respectively, followed by
7799
dx.doi.org/10.1021/jm200677b |J. Med. Chem. 2011, 54, 7797–7814

Journal of Medicinal Chemistry
ARTICLE
Scheme 3a
a Reagents: (a) n-BuLi, di-tert-butyl diazodicarboxylate, THF, ꢀ78 ꢀC, 2 h, 50%; (b) 4 M HCl, 1,4-dioxane, room temp, 16 h, 60%; (c) (i) conc HCl,
NaNO2, H2O, ꢀ10 ꢀC, 0.5 h, (ii) 6 N HCl, SnCl2, ꢀ10 ꢀC, 3.5 h.
Scheme 4a
permeability and solubility. The latter concept could conceivably
lead to increasing the duration of action (DoA) of the inhibitor at
the site of action by retaining the compound in the lung.
In addition to achieving a DoA by physical means, we hypo-
thesized that a slow koff from the enzyme itself could extend
the DoA. The discovery of the oral p38 inhibitor BIRB-796 and
its unique enzyme kinetics suggested that potent and selective
p38α inhibitors could be designed with the potential for a longer
duration of action (DoA).22,23 The slow kon and slow koff rates
observed with the inhibitor were rationalized on the basis of
X-ray crystallographic observations. For BIRB-796 to bind to
p38α, it was reported that the activation loop (consisting of the
conserved residues Asp168-Phe169-Gly170) must undergo a
movement of about 10 Å, creating a lipophilic pocket that was
then filled by the arylpyrazole moiety. High affinity binding to the
so-called DFG-out conformation was reported to deliver the long
koff rates. For this reason we decided to target the DFG-out
a Reagents: (a) EtOH, reflux, 1 h, 86ꢀ99%; (b) PhI(OAc)2, DCM,
conformation.
EtOH, room temp, 18 h, 68ꢀ88%; (c) 2-mercaptobenzyl alcohol, Pd-
At the time we were aware of a triazolopyridine series of DFG-
in binding inhibitors of p38 that were highly selective because of a
(dppf)2Cl2, CsCO3, DMF, 90 ꢀC, 2 h, 17ꢀ77%; (d) (i) (Ms)2O, DCM,
room temp, 2 h, (ii) 7 N NH3/MeOH, room temp, 16 h, 38ꢀ98%;
(e) BBr3, DCM, 0 ꢀC, 0.5 h, 51%.
unique ability to act as a dihydrogen bond acceptor at the hinge
region of the kinase due to an induced flip of Gly110.24 A similar
series containing a triazolopyridine, as exemplified by 41 (Chart 4),
had also been published that contained a urea, similar to that in
BIRB-796.25 41 has been cocrystallized in p38α and was found to
bind in the DFG-in conformation and as expected had relatively
fast kon and koff rates, as determined by surface plasmon resonace
(SPR) (Table 1). This prompted us to consider hydridizing 41
and BIRB-796 to see if a potent DFG-out binding compound
could be achieved with enzyme kinetics similar to those of BIRB-
796 (Figure 1). As can be seen from the overlay of 41 onto the
cocrystal structure of BIRB-796 in p38α, the N-3 of the triazolo-
pyridine in 41 overlays well with the oxygen of the morpholino
group in BIRB-796. These are the key hydrogen bond acceptors
that make an interaction with Met109 at the hinge region of the
kinase. The phenyl ring of 41 interacts with the gatekeeper
residue (Thr106) and overlays well with the naphthalene ring
system of BIRB-796, and finally the ureas of 41 and BIRB-796 are
also suitably aligned. With these structural similarities in mind 1a
(Chart 4) was designed and synthesized and gratifyingly proved
to be a potent p38 inhibitor with enzyme kinetics akin to those of
BIRB-796 (Figure 2, Table 1).
reaction of the carbamates with the benzylamines 2aꢀc and 2g to
afford the desired target compounds in good yields (Scheme 9).
Target compounds 1h, 1vꢀx, and 1aaꢀab did not require
desilylation of the phenolic protecting group that was on 31cꢀe,
as it was spontaneously removed during the urea forming step.
’ RESULTS AND DISCUSSION
Alongside the essential characteristics of a successful drug,
such as efficacy and safety, inhalation as a route of delivery
requires a compound to have certain other attributes to increase
the likelihood of success. These include (1) high potency to
achieve a low inhaled dose (e1 mg of drug product), (2)
duration of action (DoA) for therapeutically relevant period,
(3) high clearance and low oral absorption of any swallowed dose
to mimimize systemic exposure and therefore potential for adverse
events, (4) multiple routes of clearance to minimize drugꢀdrug
interactions (DDIs), and (5) physical form characteristics sui-
table for a dry powder inhaler including high crystallinity and
compatibility with lactose which is the commonly used excipient.
These requirements suggest that designing compounds at the
edge of or outside the “rule-of-five” (Ro5) would potentially
increase the likelihood of achieving the pharmacokinetic profile
mentioned above.21 Non-Ro5 physicochemical space could
potentially deliver a compound with relatively higher clearance,
through high lipophilicity, and a slower rate of absorption of drug
from the lung to the systemic circulation through a reduction in
The slow kon and koff rates presented a screening challenge at
the outset of this program. We decided to use a cell based assay to
routinely assess the potency of target compounds. Compounds
were tested for their ability to inhibit TNF-α production from
LPS stimulated human isolated peripheral blood mononuclear
cells (PBMCs). First we needed to establish how long our whole
7800
dx.doi.org/10.1021/jm200677b |J. Med. Chem. 2011, 54, 7797–7814

Journal of Medicinal Chemistry
ARTICLE
Scheme 5a
a Reagents: (a) (BOC)2O, iPrNEt2, DMF, room temp, 72 h, 100%; (b) Na2CO3, 1,4-dioxane, H2O, 70 ꢀC, 16 h, 87%; (c) THPOCH2CH2Br, NaH,
N-methylpyrrolidinone, 55 ꢀC, 16 h, 55%; (d) 4 M HCl, 1,4-dioxane, room temp, 2 h, 95%.
Scheme 6a
The tert-butyl group of 1a resides in a lipophilic pocket created
by activation loop movement. Introduction of a “soft” metabolic
group within this pocket may render the molecule to site directed
metabolism leading to a more polar metabolite having reduced
binding affinity for p38. Alternatively the isopropyl group on the
triazolopyridine of 1a could be replaced with soft metabolic
groups that may also lead to a less potent metabolite. These
modifications would lead to phase I cytochrome P450 (CYP)
mediated metabolism; therefore, to mitigate any CYP mediated
DDI potential, we were interested in introducing the potential
for phase II mediated metabolic pathways. Glucuronidation is a
phase II metabolic pathway that has high capacity and is a major
route of drug inactivation.27 Phenol groups are well documented
to undergo glucuronidation, and so we intended to explore the
possibility of their introduction into synthetically accessible
regions of the molecule that were also likely to be tolerated from
a potency standpoint. The N-phenyl ring of the arylpyrazole
moiety of 1a was binding into a relatively hydrated pocket and
therefore seemed a sensible site for the introduction of a phenolic
group (Table 2).
As can be seen from compounds 1bꢀf,h,i, generally good levels
of potency could be achieved relative to the starting point 1a.
Introduction of a thiomethyl group at R2, in 1b, led to a slight
drop-off in potency while a methylenethiomethyl, in 1c, group
led to a further drop-off in potency, suggesting that we were
pushing the boundaries of the lipophilic pocket. It was gratifying
to note that introduction of phenols in the meta and para
positions of R1 with 1d and 1e, respectively, did not lead to a
significant drop-off in potency. Moreover, the combination of a
thiomethyl group at R2 with a m-phenol at R1 retained good
potency in 1f alongside high intrinsic clearance. It was also
pleasing to see some level of glucuronidation with this com-
pound, although probably not enough to mitigate a CYP DDI.
Interestingly, the corresponding proposed sulfoxide metabolite
of this compound 1g led to a significant drop-off in potency,
thereby validating the hypothesis above. Introduction of a
chlorine atom adjacent to the phenol in 1h and 1i resulted in
potency akin to the starting point 1a and displayed good levels of
glucuronidation alongside high CYP mediated metabolism.
In parallel to the changes we were making on the arylpyrazole
moiety we were interested in understanding if modifications of
the isopropyl group on the triazolopyridine may also be useful.
a Reagents: (a) benzyl bromide, K2CO3, THF, reflux, 30 h, 92%.
cell PBMC assay would need in terms of preincubation time to
ensure that the experiment was at equilibrium. 1a, for example,
had an IC50 in the whole cell assay of 5 nM after a 10 min
preincubation time, whereas if the preincubation time was
increased to 1 h, the IC50 decreased to 0.3 nM. It did not decrease
further with longer preincubation times. Therefore, a preincuba-
tion time of 1 h was used for future screening. This suggests that
the slow kon and koff were clearly impacting the time to equilibrium
in this experiment. With 1a being potent in a cell based assay and
having the desired enzyme kinetics, we were keen to investigate
the selectivity profile against wider kinases. Through binding to
the DFG-out region of the kinase it was hypothesized26 that a
good selectivity profile could be achieved, as many kinases in the
kinome are potentially unable to adopt a DFG-out conformation.
Interestingly, 41 being a DFG-in binding mode was more
selective than the DFG-out binding mode observed with com-
pound 1a, which is contrary to the hypothesis above. It is possible
that the kinases with significant inhibition at 10 μM (shown in
red) are capable of binding in the DFG-out binding mode in a
similar way to p38 (Figure 3).
With a promising lead in 1a we sought to investigate and
optimize the pharmacokinetic and pharmaceutical properties for
inhalation while maintaining the high levels of potency and
selectivity. As mentioned above, targeting high hepatic clear-
ance was important for minimizing systemic exposure follow-
ing absorption from the lung or absorption from the gut from any
swallowed dose. In the targeting of liver metabolism, it was also
key to ensure that the metabolites were inactive or had reduced
activity and therefore could not contribute to any systemic p38
activity. With these considerations in mind we examined the
cocrystal structure of 1a in p38α which pointed to several
opportunities to tune the pharmacokinetic properties of this
series (Figure 4).
7801
dx.doi.org/10.1021/jm200677b |J. Med. Chem. 2011, 54, 7797–7814

Journal of Medicinal Chemistry
ARTICLE
Scheme 7a
a Reagents: (a) TBDMSOCH2CH2Br, NaH, DMF, room temp, 16 h, 81%; (b) (i) lithium aluminum hydride, THF, 0 ꢀC f room temp, 16 h, 81%, (ii)
Ba(MnO4)2, DCM, room temp, 24 h, (iii) MnO2, DCM, reflux, 16 h, 62%.
Scheme 8a
a Reagents: (a) (i) CDI, DCM, room temp, 16 h, (ii) 2bꢀf, 2h, iPrNEt2, DCM, room temp, 1 h, 23ꢀ100%; (b) BBr3, DCM, 0 ꢀC, 0.5 h, 46ꢀ85%;
(c) H2O2, 1,1,1,3,3,3-hexfluoropropan-2-ol, room temp, 0.5 h, 22ꢀ80%; (d) PTSA, MeOH, 60 ꢀC, 3 h, 67%.
The cocrystal structure of 1a suggested there was space to replace
the isopropyl group with larger groups and in particular with a
phenyl ring. To this end, we designed and synthesized 1jꢀs
(Table 3).
this time retained good potency with only a modest drop-off
from its parent 1p. 1p and others above highlight that this region
of the kinase binding site was very tolerant of functional groups of
varying polarities and size and several different positions and
hence represent a good opporunity to tune the physicochemical
properites and in particular the properties required for inhala-
tion. Finally introduction of a flexible hydroxyethylene group
either on the phenyl ring at the DFG-out binding site, in 1r, or at
the ATP binding site, in 1s, maintained good cellular potency
despite being a relatively large group.
At this stage of the project we were interested to know
whether the optimization of in vitro properties would translate
into high clearance in vivo. To understand this, we investigated
the intravenous and oral pharmacokinetic properties in rats
(Table 4). The rat liver microsome (RLM) intrinsic clearance
was high for 1f, which translated into near liver blood flow
clearance in the in vivo experiment and very high unbound
clearance of >9600 (mL/min)/kg also. The high clearance and
It was pleasing to note that 1j retained good cellular potency
after inclusion of a 2-hydroxyphenyl ring on the ATP binding
portion of the molecule. 1k, while also potent, demonstrated
some level of glucuronidation, albeit low. Interestingly 1l was
more potent than 1c seen previously, suggesting that on this
scaffold the thiomethylmethylene group could be accommo-
dated in the active site. Moreover, the sulfoxide metabolite 1m of
this compound led to a more than 100-fold drop-off in potency,
which was encouraging from a metabolite activity perspective.
The phenolic hydroxyl group could be moved around the phenyl
ring in combination with a chloride substituent, in 1n and 1o,
without loss of cellular potency. Alternatively the 2-hydroxyphe-
nol could be replaced with a thiomethyl group in 1p, again
maintaining good cellular potency. The sulfoxide metabolite 1q
7802
dx.doi.org/10.1021/jm200677b |J. Med. Chem. 2011, 54, 7797–7814

Journal of Medicinal Chemistry
ARTICLE
Scheme 9a
a Reagents:(a) PhO(CO)Cl, pyridine, THF, 0 ꢀC f room temp, 1 h, 86ꢀ100%;(b) iPrNEt2, DMSO, room temp, 16 h, 23ꢀ96%; (c) THPOCH2CH2Br,
K2CO3, DMF, 60 ꢀC, 8 h, 75%; (d) triethylamine trihydrofluoride, MeOH, room temp, 16 h, 67ꢀ75%; (e) BBr3, DCM, 0 ꢀC, 0.5 h, 35ꢀ49%; (f) H2O2,
1,1,1,3,3,3-hexfluoropropan-2-ol, room temp, 0.5 h, 70%; (g) PTSA, MeOH, 60 ꢀC, 3 h, 50ꢀ62%.
Chart 4
low volume of distribution of 1f lead to a very short half-life in rat.
The oral pharmacokinetics of 1f suggests negligable oral bioa-
vailability. This, of course, was exactly the pharmacokinetic profile
we were targeting to minimize systemic exposure either through
a lung absorbed fraction or an orally absorbed fraction following a
swallowed dose.
Alongside favorable pharmacokinetic properties we were
interested to know if the selectivity profile of 1f was maintained
from that of the original lead in 1a. Indeed this was the case, with
1f being only a potent inhibitor of isoforms of p38 (α, β, and γ)
(Figure 5).
pattern of the crystalline 1f suggested a good level of crystallinity,
with a melting point of 220 ꢀC, but the compound had no
detectable solubility in water or in our simulated lung fluid (SLF)
preparation. This was of great concern not only because dis-
solved material is required for a pharmacological response but
also because this series of compounds appeared to have slow kon
and koff rates, necessitating a relatively high local concentration to
drive binding to the kinase. This prompted us to focus on looking
for an inhibitor with an improved solubility profile. As a bench-
mark, we considered the marketed inhaled corticosteroid drug
fluticasone as a good indicator of the required solubility we
should be targeting in our SLF preparation (SLF ≈ 0.5 μg/mL).
In order to improve the solubility, we sought to disrupt the
crystal packing of the molecule to a large enough extent such that
the compound became more soluble but not so much that the
The favorable pharmacokinetic properties led us to evaluate
some of the pharmaceutical properties of 1f to see if the crystal-
linity, solid form properties, and solubility were commensurate
with those required for inhalation. The powder X-ray diffraction
7803
dx.doi.org/10.1021/jm200677b |J. Med. Chem. 2011, 54, 7797–7814

Journal of Medicinal Chemistry
ARTICLE
Table 1. p38α Potency and Enzyme Kinetics
compd
41a
1aa
kon (Mꢀ1 sꢀ1
)
koff (sꢀ1 t1/2 (h) KD (nM) IC50 (nM)b
)
4.3 ꢁ 106
1 ꢁ 105
5 ꢁ 104
0.014
0.01
4.81
3.85
3.3
0.4
1.0
58
0.3
15
4 ꢁ 10ꢀ5
BIRB-796a
5 ꢁ 10ꢀ5
a Enzyme kinetics and KD were determined by surface plasmon reso-
nance using a Biacore instrument, on the inactive form of the kinase.
b Compounds were routinely profiled by assessing their ability to inhibit
TNF-α production from lipopolysaccharide (LPS) stimulated human
isolated peripheral blood mononuclear cells (PBMCs), following a 1 h
preincubation of PBMCs and test compound.
Figure 3. Selectivity profiles of 41 and 1a against a panel 45 kinases at
10 μM.
Figure 1. Overlay of 41 (carbon atoms colored green, derived from
PDB entry 2yix) in the protein DFG-out cocrystal structure of BIRB-796
(carbon atoms colored purple) in p38α (PDB entry 1kv2).
Figure 4. Binding site surface of 1a in p38α cocrystal structure (PDB
entry 2yiw). The region of 1a highlighted in purple is binding into the
pocket created by the DFG loop movement. Binding surface illustrates
lipophilic surface (green/yellow), surface made up of hydrogen accept-
ing oxygen atoms (red) and hydrogen donating atoms (blue).
Figure 2. Overlay of 1a (carbon atoms colored orange, derived from
PDB entry 2yiw) in the protein DFG-out cocrystal structure of BIRB-
796 (carbon atoms colored purple) in p38α (PDB entry 1kv2).
glucuronidation, alongside the other favorable characteristics
observed in compound 1r. 1tꢀac were designed and synthesized
to address the above issues (Table 5).
1t and 1u showed encouraging increases in glucuronidation
rate relative to 1r. However, significant improvement in glucur-
onidation rate came through the addition of 2-chlorophenol ring
systems in 1vꢀx. Interestingly 1v and 1w had differing rates of
glucuronidation, suggesting that a p-phenol was preferred to
maximize the rate of phase II metabolism. 1x in particular was an
interesting lead, as it had one of the highest rates of gluronidation
observed and also had very high cytochrome P450 driven metabo-
lism, but unfortunately the addition of the methylenethiomethyl
group at R2 led to a significant drop-off in potency. Interestingly,
the addition of a 2-fluorophenol in 1y also led to a reduction in
potency relative to the chloro analogue 1v. It is interesting to
note not only that the rates of glucuronidation appear dependent
on the pKa of phenol being glucuronidated but also that the
compounds have SAR. This is probably not a surprise, as the
compound became an amorphous solid. From our SAR studies
we knew that the ATP binding site could tolerate groups such as
hydroxyethylene in 1r and 1s, and we hypothesized that intro-
duction of such flexible groups might disrupt the crystal packing
enough to improve solubility to some extent. To this end, we
crystallized compound 1r and measured its SLF solubility, and
we were pleased to see that we could now detect some dissolved
material (SLF 1.0 μg/mL), which was above that of the bench-
mark fluticasone. As a result, 1r became an attractive lead with
good cellular potency, high intrinsic clearance, but unfortunately
low levels of glucuronidation in the HLM UGT in vitro assay. In
fact modeling of the DDI potential of 1r did suggest that the lack
of multiple routes of clearance for this compound could lead to it
becoming a victim of a DDI. This led us to target higher rates of
7804
dx.doi.org/10.1021/jm200677b |J. Med. Chem. 2011, 54, 7797–7814

Journal of Medicinal Chemistry
ARTICLE
Table 2. p38 Potency, clogP, HLM, and HLM UGT Data for 1aꢀi
a Compounds were routinely profiled by assessing their ability to inhibit TNF-α production from LPS stimulated human isolated PBMCs. b Compounds
were routinely assessed for metabolic stability in human liver microsomes (HLM). c Compounds were assessed for phase II glucuronidation in HLM by
including the cofactor uridine diphosphate glucuronic acid (UDPGA) and excluding the cytochrome P450 dependent cofactor NADPH.
enzyme responsible for transferring the glucuronide to the xeno-
biotic compound will have a specific recognition element and
therefore display SAR.
1v, and so we sought to incorporate a soft metabolic group in
another part of the inhibitor to increase the chances of produc-
ing a significantly less active metabolite. 1p had previously
suggested that a thiomethyl group on the ATP binding side of
the molecule was tolerated from a potency perspective but also
Several of the methylenethiomethyl analogues (R2 = CH2SMe),
such as 1x, were less potent than the tert-butyl analogue (R2 = Me)
7805
dx.doi.org/10.1021/jm200677b |J. Med. Chem. 2011, 54, 7797–7814

Journal of Medicinal Chemistry
ARTICLE
Table 3. p38 Potency, clogP, HLM, and HLM UGT Data for 1jꢀs
a Compounds were routinely profiled by assessing their ability to inhibit TNF-α production from LPS stimulated human isolated PBMCs. b Compounds
c
were routinely assess for metabolic stability in HLM. Compounds were assessed for phase II glucuronidation in HLM by including the cofactor
UDPGA, excluding the cytochrome P450 dependent cofactor NADPH.
7806
dx.doi.org/10.1021/jm200677b |J. Med. Chem. 2011, 54, 7797–7814

Journal of Medicinal Chemistry
ARTICLE
Table 4. In Vitro and in Vivo ADME Properties for 1f
compd
iv clearancea ((mL/min)/kg)
Vss (L/kg)
T1/2 (h)
oral bioavailabilityb (%)
<5
PPB (%)
>99
RLM ((μL/min)/mg protein)
1f
77
0.9
1
200
a Intravenous (iv) pharmacokinetics were determined in rats from a 1 mg/kg dose. b Oral pharmacokinetics were determined from a 3 mg/kg dose.
provides the optimal binding kinetics to overcome competition
with high cellular levels of ATP. The potency and dissociation
t1/2 of 1ab for inhibiting rat p38α kinase were similar to the
results observed for the human enzyme.
1a was shown to bind in the DFG-out binding mode which
translated into slow dissociation kinetics, and so we were keen to
confirm that the DFG-out binding mode was also apparent with
1ab. Indeed, a protein cocrystal structure confirmed this to be the
case (Figure 6).
The previous leads 1a and 1f displayed good selectivity
profiles, and so we were keen to confirm that 1ab was also
selective over other kinases. 1ab has an IC50 at enzyme Km for
ATP of <1 μM for 15 enzymes of the 205 recombinant kinases
screened. Scaling to the p38α IC50 of 0.88 nM, five kinases have a
selectivity ratio of e100ꢁ. Whole cell assays were established
that demonstrated that recombinant enzyme activity does not
translate to the whole cell for three of these kinases, leaving p38β
Figure 5. Selectivity profile of 1f against a panel of 29 kinases, showing
% inhibition at 10 μM.
(enzyme IC50 = 23nM) and misshapen NIK related kinase 1
(MINK) (enzyme IC50 6nM) as potential cross over targets for
1ab. Selectivity over p38β was not thought to represent a toxicity
liability or be required for efficacy. The low systemic exposure
expected for 1ab coupled with the seemingly fast kon and slow koff
kinetics for p38 suggested that inhibition of MINK is unlikely to
represent a toxicity risk. Moreover, 1ab did not inhibit LPS
induced JNK phosphorylation (MINK signals through JNK) in
the human monocytic cell line U937 at 1 μM, suggesting that 1ab
only inhibited p38 dependent end points.
In terms of physicochemical properties, 1ab is a non-Ro5
compliant, high molecular weight (700 Da), highly lipophilic
(clogP = 7.3, log D > 4.1), and weakly acidic molecule exhibiting
an acidic pKa of 7.45 (chlorinated phenol group). It therefore has
low aqueous solubility (<0.1 mg/mL) but with measurable SLF
solubility (3.0 μg/mL) when determined on a crystalline form.
The pharmaceutical properties for 1ab including crystallinity,
stability on lactose, and solubility were considered suitable for
dry powder inhalation.
The pharmacokinetics of 1ab was then assessed in vitro and
in vivo. 1ab was rapidly metabolized in rat, dog, and human liver
microsomes. The HLM glururonidation experiment suggested
that glucuronidation will significantly contribute to the human
clearance of 1ab. In fact the metabolite profile from human
hepatocytes suggested that the glucuronide resulting from trans-
fer to the chlorophenol is the major metabolite and presumably
will be devoid or have greatly reduced p38 activity. In human
hepatocytes, coadministration of 2 μM ketoconazole (CYP3A4
inhibitor) did not lead to any significant reduction in intrinsic
clearance of 1ab over the control incubations, suggesting that
enzymes other than CYP3A4 contributed to the hepatocyte
clearance of 1ab with glucuronidation being a likely alternative
pathway. These data suggest that 1ab should not be a victim of a
DDI. 1ab is a moderate inhibitor of CYP1A2, CYP2C19, and
CYP2D6 and a potent inhibitor of CYP’s 2C9 and CYP3A4.
However, on the basis of the very low projected unbound
plasma Cmax of 1ab, the potential for the compound acting as a
perpetrator of a DDI is low.
that the metabolite 1q only led to a modest drop-off in potency.
This led us to design 1z which incorporates both a soft metabolic
handle and the flexible solubilizing hydroxyethyl chain. To
improve on the glucuronidation rate of 1z, we incorporated the
more acidic phenol in 1aa and 1ab (PF-03715455), with 1ab
having the best balance of potency, HLM intrinsic clearance, and
glucuronidation rate. The sulfoxide metabolite 1ac also showed
an approximate 25-fold drop-off in potency.
At this stage 1ab was a very attractive lead, and so we set about
characterizing it further. In the PBMC cell assay, with a pre-
incubation time of 1 h the potency was determined to be 1.7 nM.
Determining the potency of 1ab against the human p38α enzyme
proved to be complex as a result of the slow dissociation kinetics
observed with this compound. The interaction of 1ab with p38α
was quantified by monitoring the IC50 as a function of the
preincubation time. In the presence of nonsaturating levels of
ATP (10-fold below Km, 5 μM ATP) a decrease in the IC50 for
1ab was observed from 9 nM (no preincubation) to 0.88 nM
(40 min of preincubation).
The binding kinetics of 1ab was determined for the inactive/
nonphosphorylated form of human p38α using SPR. Fast kon and
slow koff kinetics were observed with 1ab for p38α, resulting in a
half-life of approximately 80 h (Table 6). Clearly the enzyme
kinetics for 1ab is significantly different from that determined for
1a, with a faster association rate and a slower dissociation rate,
leading to a much greater affinity of 1ab for the p38α enzyme,
and a much slower half-life. The origin of this difference could
not be rationalized easily from a structural perspective but could
perhaps be due to the larger group at the 2-position of the
triazolopyridine (2-hydroxyethylthiophenyl versus isopropyl)
leading to improved affinity. The mismatch between KD gener-
ated on 1ab from binding kinetic constants and enzymatic IC50 is
attributed to the enzyme assay not being at equilibrium within a
time frame suitable for enzymatic assays (maximum of 40 min).
The relatively fast kon of 1ab coupled with the slow koff rate
7807
dx.doi.org/10.1021/jm200677b |J. Med. Chem. 2011, 54, 7797–7814

Journal of Medicinal Chemistry
ARTICLE
Table 5. p38 Potency, clogP, HLM, and HLM UGT Data for 1tꢀac
a Compounds were routinely profiled by assessing their ability to inhibit TNF-α production from LPS stimulated human isolated PBMCs. b Compounds
c
were routinely assess for metabolic stability in HLM. Compounds were assessed for phase II glucuronidation in HLM by including the cofactor
UDPGA, excluding the cytochrome P450 dependent cofactor NADPH.
7808
dx.doi.org/10.1021/jm200677b |J. Med. Chem. 2011, 54, 7797–7814

Journal of Medicinal Chemistry
ARTICLE
Demonstrating intrinsic cell permeability was difficult to
achieve with 1ab because of its low solubility. However, as we
were observing good levels of potency (<10nM) in our whole cell
PBMC experiments, it is assumed that 1ab has some level of
intrinsic cellular permeability, although it is suggested to be low
based on its physicochemistry. 1ab is a high affinity substrate for
P-glycoprotein (P-gp) (Km= 0.12 μM). This, coupled with low
solubility, and presumed low intrinsic permeability translated
into neglibible oral absorption in the rat (Table 7). Moderate
total clearance was observed in rat and dog with 1ab, but when
corrected for high protein binding, this translates into extremely
high unbound clearance values of 33 000 and 15 000 (mL/min)/
kg, respectively. Human clearance predictions from preclinical
data suggest low oral bioavailability and high clearance, resulting
in a short half-life. Therefore, any drug reaching the systemic
circulation, through absorption from either the gut or the lung,
will be rapidly metabolized to relatively inactive metabolites.
With suitable pharmacokinetic properties demonstrated with
1ab, we were then keen to assess if the overall properties of the
molecule would translate into a desirable in vivo efficacy profile.
LPS challenge to the lungs of rats results in lung neutrophilia, a
characteristic of COPD, in addition to induction of both exhaled
nitric oxide and inflammatory mediator expression. When 1ab
was dosed as a micelle solution intratracheally (IT) 2 h prior to
stimulus, it dose-dependently inhibited LPS induced lung neu-
trophilia with an efficacious dose (ED50) of 30 μg. The maximum
inhibition was similar to that of a solution dosed corticosteroid
(fluticasone, dose 100 μg, 64% inhibition) of approximately
50ꢀ70%. To investigate the in vivo DoA of 1ab, the compound
was predosed for increasing times prior to LPS challenge. 1ab
showed significant inhibition of neutrophilia on 4 h of predosing
but did not significantly inhibit lung neutrophilia on 24 h of
predosing. To confirm that the apparent DoA of 1ab was not due
to systemic exposure, systemic anti-inflammatory activity was
assessed in an ex vivo whole blood LPS induced TNFα assay.
While a significant inhibition of ex vivo LPS induced TNFα
production was observed 1 h after IT solution dosing, there was
no significant inhibition in blood collected 4 h after dosing. These
data suggest that the DoA observed in the LPS neutrophilia
model was driven by sustained p38α inhibition in the lung and
was not due to systemic exposure (Figure 7). It is possible that
this observed DoA is made possible by a combination of physical
retention of the compound in the lung, as well as the slow
dissociation kinetics from the p38α enzyme.
1ab was subjected to wide ligand profiling (CEREP) at 10 μM
which revealed <50% inhibition against 66 of the 68 targets
assayed. IC50 values were therefore obtained for 1ab at the
N-type calcium channel (12 μM) and the human VEGFR1-
tyrosine kinase (0.89 μM, consistent with findings in the kinase
selectivity section). These affinities were all >1000ꢁ the primary
pharmacology of 1ab and therefore not considered a significant
risk. Furthermore, 1ab demonstrated little effect on the hERG
current amplitude in the patch-clamp assay (46 ( 8.8% inhibi-
tion at 10 μM (n = 6)) and no in vitro genetic toxicity was
observed. Intravenous and oral in vivo toleration studies in the rat
with 1ab for 4 days did not induce any relevant clinical signs,
clinical pathology, or histopathological evidence of toxicity at
doses up to 2 and 2.8 (mg/kg)/day, respectively. Overall 1ab had
a preclinical safety profile suggesting it would be tolerated in
humans at clinically relevant inhaled doses.
Table 6. p38α Affinity and Enzyme Kinetics of 1ab
compd
kon (Mꢀ1 sꢀ1
3.1 ꢁ 106
)
koff (sꢀ1
)
t1/2 (h)
KD (nM)
1aba
2.4 ꢁ 10ꢀ6
80
0.001
a Enzyme kinetics were determined by surface plasmon resonance on the
inactive form of the kinase using a Biacore instrument.
’ CONCLUSION
We have described our approaches toward the identification of
a novel DFG-out binding lead series, beginning with 1a, followed
by subsequent optimization of various properties to ultimately
identify 1ab. Much of the design philosophy is similar to that
described as “inhalation by design”, which was outlined in a previous
publication.20 1ab is a very potent and selective inhibitor for the
p38α enzyme, which translates into potent inhibition of proin-
flammatory cytokine release in whole cells that are relevant to
COPD. Moreover, the enzyme kinetics of the inactivated form of
the p38α enzyme suggest fast association and slow dissociation,
which should provide optimal binding kinetics to overcome
competition with high cellular levels of ATP. In vivo experiments
Figure 6. Binding site surface of 1ab in p38α cocrystal structure (PDB
entry 2yis). The region of 1ab highlighted in purple is binding into the
pocket created by the DFG loop movement. Binding surface illustrates
lipophilic surface (green/yellow), surface made up of hydrogen accept-
ing oxygen atoms (red) and hydrogen donating atoms (blue).
Table 7. Intravenous (iv) and Oral Rat and Dog Blood Pharmacokinetics of 1ab
species
dose (mg/kg)
iv clearance ((mL/min)/kg)
33.8
Vss (L/kg)
T1/2 (h)
oral absorption (%)
<5
PPB (%)
rat iva
rat pob
dog ivc
1
0.19
1.0
99.8
99.8
99.9
0.5
0.48
13.1
0.06
0.5
a Intravenous pharmacokinetics were determined in rats from a 1 mg/kg dose. b Oral pharmacokinetics were determined from a 0.5 mg/kg dose.
c Intravenous pharmacokinetics were determined in dogs from a 0.5 mg/kg dose.
7809
dx.doi.org/10.1021/jm200677b |J. Med. Chem. 2011, 54, 7797–7814

Journal of Medicinal Chemistry
ARTICLE
Figure 7. (A) Dose dependent inhibition of LPS induced neutrophilia at 2 h predosing with 1ab and (B) demonstration of in vivo duration of action of
greater than 4 h based on predosing studies: /, p < 0.05; NS, not statistically significant.
in rat suggest that 1ab has a a DoA greater than 4 h, which could
be explained by slow dissociation kinetics in combination with
physical retention of compound in the lung. The in vitro and
in vivo pharmacokinetic profile suggests that the compound will
display low oral bioavailability in humans following any swal-
lowed dose, coupled with very high intrinsic clearance such that
any lung absorbed drug will be rapidly metabolized to relatively
inactive metabolites. Data suggest that 1ab is metabolized by
both phase I and phase II pathways and therefore should not be a
victim or a perpetrator of any DDIs. The pharmaceutical proper-
ties of 1ab are suitable for combination with lactose in a dry
powder inhaler, with good stability, crystallinity, and solubility
being key features. Finally the compound has demonstrated a
good preclinical safety profile and may offer an improved safety
profile relative to oral p38 inhibitors. This promising overall
profile has led to the nomination of compound 1ab as a
development candidate for the treatment of COPD. Further
information relating to the research and development of com-
pound 1ab will be reported in due course.
before recovery of samples of supernatant. TNF-α in the samples was
determined using an enzyme-linked immunosorbent assay (ELISA)
(Invitrogen kit no. CHC-1754; Invitrogen Carlsbad, CA) and following
the manufacturer's instructions. Dose response curves were constructed
from which IC50 values were calculated. At least n = 2 determinations
were made from a single donor of PBMCs. For compounds such as 1ab,
n = 6 determinations were conducted from six different donors, with
IC50 values ranging from 1.39 to 2.02 nM.
In Vivo Pharmacological Activity. ED50 Determination of
Inhibtion of LPS-Induced Lung Neutrophilia in Rats. The
primary objective of this procedure was to determine the anti-inflam-
matory activity of the compounds of formula 1, when administered
directly into the lungs via the trachea, and also to determine the duration
of effect. Test compounds were dissolved, or prepared as fine suspen-
sions, in phosphate buffered saline containing 0.5% (w/v) Tween-80
to provide a range of dose levels. Male CD SpragueꢀDawley rats (300ꢀ
450 g) were randomized to study groups of n = 6 and then briefly
anesthetized in an anesthetic chamber with 5% isoflurane in 3 L/min O2.
One of the test compound formulations or dose vehicle (100 μL) was
injected directly into the trachea of each anesthetized rat using a
Hamilton syringe. The animals were then allowed to recover from the
anesthetic. Dependent on the study design, animals were predosed 2, 4,
or 24 h prior to inflammatory challenge. The rats were placed into a
chamber (300 mm ꢁ 300 mm ꢁ 450 mm), connected to an ultrasonic
nebulizer and a small animal rodent ventilator set to the maximum tidal
volume and rate (5 mL, 160 strokes/min). An amount of 10 mL of
1 mg/mL LPS (Sigma-Aldrich, L2630) dissolved in saline, prewarmed to
37 ꢀC, was nebulized into the chamber. After 15 min the ventilator and
nebulizer were turned off, and the animals remained in the chamber to
breathe the mist for a further 15 min before being returned to the home
cage. Four hours after the end of the LPS treatment the animals were
terminally anesthetised with 1 mL/kg Pentoject ip. The trachea was
cannulated. The lungs were lavaged with 4 ꢁ 2.5 mL of PBS containing
2.6 mM EDTA, and the lavage fluid was collected. An amount of 1 mL of
bronchioalveolar lavage (BAL) was added to 125 μL of 40% bovine
serum albumen (BSA) and the cellular count determined using an Advia
120 hematology system (Siemens). Also, a terminal blood sample was
collected from each rat, and IC50 was determined for inhibition of LPS
induced TNFα release from whole blood. Dose response curves were
constructed for inhibition of LPS-induced lung neutrophils, and half-
maximal effect doses (ED50) were estimated from the fitted curves.
p38α Enzyme Kinetics (kon and koff) for PF-03715444
(1ab), As Determined by Surface Plasmon Resonance Using
a Biacore Instrument, on the Inactive (Nonphosphorylated)
Form of the Kinase. Biacore uses a technique based upon surface
plasmon resonance to follow the real-time binding of compounds in a
’ EXPERIMENTAL SECTION
Biology. In Vitro Pharmacological Activity. IC50 Determi-
nation of Inhibition of TNF-α Release from Isolated Human
Peripheral Blood Mononuclear Cells (PBMCs) after LPS
Stimulation. Peripheral venous blood from healthy, nonmedicated
donors was collected using ethylenediaminetetraacetic acid (EDTA) as
the anticoagulant. For PBMC preparation, samples of blood were
diluted 1:1 with sterile phosphate buffered saline and then separated
using ACCUSPIN System-Histopaque-1077 tubes (Sigma-Aldrich, St.
Louis, MO), centrifuged at 400g for 35 min. Buffy coat cells were
removed into PBS, centrifuged at 200g for 10 min, and resuspended in
PBMC assay buffer (Hanks balanced salt solution, 0.28% [w/v] 4-[2-
hydroxyethyl]-1-piperazineethanesulfonic acid [HEPES], 0.01% [w/v]
low-endotoxin bovine serum albumin [BSA]). A differential white cell
count was performed, and PBMCs were diluted to 1 ꢁ 106 lymphocytes
per mL in PBMC assay buffer. Test compounds were dissolved in
DMSO and diluted in PBMC assay buffer (final DMSO concentra-
tion 1%) to cover an appropriate concentration range, e.g., 0.001ꢀ
10000 nM. Samples of test compound solution or vehicle (20 μL) were
added into 96-well tissue culture treated plates (Corning), and PBMC
(160 μL) added to each well. The assay mixtures were incubated at 37 ꢀC
for 1 h in a humidified incubator containing an atmosphere of air sup-
plemented with 5% CO2 before adding LPS (20 μL of 100 ng/mL). Plates
were returned to the incubator for a further 18 h and then centrifuged
7810
dx.doi.org/10.1021/jm200677b |J. Med. Chem. 2011, 54, 7797–7814

Journal of Medicinal Chemistry
ARTICLE
liquid milieu to immobilized proteins on a carboxymethyl dextran surface.
The compounds flow across the protein surface, and as they bind, there
is an increase in signal (RU). Then, as the compounds dissociate from
the surface, the signal decreases and returns to baseline. These binding
studies were performed at 25 ꢀC using a running buffer of 10 mM
HEPES, 150 mM NaCl, 0.005% P20 with a flow rate of 20 μL/min. For
immobilization, a CM5 chip (Biacore, Piscataway, NJ) was precondi-
tioned and then streptavidin immobilized (13000ꢀ14000 RU) using
amine coupling with N-hydroxysuccinimide and N-ethyl-N0-(3-dimeth-
ylaminopropyl)carbodiimide according to Biacore methods. Nonpho-
sphorylated BirA-MK2 (control flow cell) or BirA-p38α were then
captured to a level of 10000ꢀ11000 RU. These particular binding
experiments were done slightly differently from typical Biacore studies.
To the immobilized, nonphosphorylated, and biotinylated p38α was
added SC-71830 (200 nM) which was first allowed to bind (three times)
to achieve a maximal control response. SC-71830 is a proprietry ATP
competitive p38α inhibitor that displays fast on and fast off p38α
enzyme kinetics. In theory any fast on and fast off inhibitor that is ATP
competitive could be used. Following dissociation of the final pass of SC-
71830, 1ab (1000 nM) was injected for 5 min. At various times after the
injection of 1ab, SC-71830 (200 nM) was injected and the binding
response measured. As 1ab dissociated from the immobilized p38α, the
binding of SC-71830 increased. The increase in binding response of SC-
71830 over time was used to measure the dissociation of 1ab, and these
data were used to calculate the kon and koff for 1ab.
Chemistry. Unless otherwise indicated, all reactions were carried
out under a nitrogen atmosphere, using commercially available anhy-
drous solvents. “Ammonia” refers to commercially available aqueous
ammonia solution of 0.880 specific gravity. Thin-layer chromatography
was performed on glass-backed precoated Merck silica gel (60 F254)
plates, and compounds were visualized using UV light, 5% aqueous
potassium permanganate, or chloroplatinic acid/potassium iodide solu-
tion. Silica gel column chromatography was carried out using 40ꢀ63 μm
silica gel (Merck silica gel 60). Proton NMR spectra were measured on a
Varian Inova 400 or Varian Mercury 400 spectrometer in the solvents
specified. In these NMR spectra, exchangeable protons that appeared
distinct from solvent peaks are also reported. Low resolution mass
spectra were recorded on either a Fisons Trio 1000, using thermospray
positive ionization, or a Finnigan Navigator, using electrospray positive
or negative ionization. High resolution mass spectra were recorded on a
Bruker Apex II FT-MS using electrospray positive ionization. Test
compound purity of g95% was determined using combustion analysis
conducted by Exeter Analytical U.K. Ltd., Uxbridge, Middlesex, U.K. In
the absence of test compound purity data determined by combustion
analyis, all target compounds were purified until at least g95% purity as
judged by HPLC utilizing an ELSD detection system.
dropwise to the mixture, followed by sodium hydroxide (1M, 3 mL). The
mixture was then stirred for 30 min at room temperature. The mixture was
filtered through Celite, and the Celite was then washed with ethyl acetate.
The organics were then concentrated in vacuo to give a clear oil (5.90 g,
1
81%). H NMR (400 MHz, CDCl3) δ 7.39ꢀ7.45 (m, 1H), 7.32ꢀ7.35
(m, 1H), 7.16ꢀ1.26 (m, 2H), 4.74 (s, 2H), 3.71 (t, 2H), 2.99 (t, 2H), 2.25
(br, 1H), 0.84 (s, 9H), 0.00 (s, 6H) ppm. LRMS: m/z 299 [M + H]+.
2-[(2-{[tert-Butyl(dimethyl)silyl]oxy}ethyl)thio]benzaldehyde
(21b). To a solution of 23 (5.90 g, 19.87 mmol) in dichloromethane
(100 mL) was added barium permanganate (25.32 g, 99.27 mmol), and
the mixture was stirred at room temperature for 24 h. The mixture was
filtered through Celite and concentrated in vacuo. Crude analysis indicated
that significant starting material remained. The residue was then diluted
with dichloromethane (100 mL), and manganese dioxide (25 g, 288 mmol)
was added. The mixture was heated to reflux overnight. The mixture was
allowed to cool to room temperature and then filtered through Celite.
The residue was concentrated in vacuo to give a purple oil (3.66 g, 62%).
1H NMR (400 MHz, CDCl3) δ 10.39 (s, 1H), 7.78 (d, 1H), 7.46 (d, 1H),
7.24ꢀ7.28 (m, 1H), 3.78 (t, 2H), 3.05 (t, 2H), 0.84 (s, 9H), 0.00 (s, 6H).
LRMS: m/z 297 [M + H]+.
2-[(2-{[tert-Butyl(dimethyl)silyl]oxy}ethyl)thio]benzal-
dehyde (5-bromopyridin-2-yl)hydrazone (15a). To a solu-
tion of 21b (3.61 g, 12.24 mmol) in ethanol (10 mL) was added
(5-bromopyridin-2-yl)hydrazine (2.30 g, 12.24 mmol), and the mixture
was heated to reflux for 3 h. The solvent was then removed in vacuo to
1
give the product as a pink solid (5.63 g, 99%). H NMR (400 MHz,
CDCl3) δ 9.20 (br, 1H), 8.39 (s, 1H), 8.13 (d, 1H), 7.88ꢀ7.91 (m, 1H),
7.66 (dd, 1H), 7.43ꢀ7.45 (m, 1H), 7.31 (d, 1H), 7.21ꢀ7.28 (m, 1H), 3.73
(t, 2H), 2.98 (t, 2H), 2.74ꢀ2.96 (br, 1H), 0.84 (s, 9H), 0.00 (s, 6H) ppm.
LRMS: m/z 466ꢀ468 [M + H]+.
6-Bromo-3-{2-[(2-{[tert-butyl(dimethyl)silyl]oxy}ethyl)thi-
o]phenyl}[1,2,4]triazolo[4,3-a]pyridine (16a). To a solution of
15a (5.63 g, 12.08 mmol) in dichloromethane (50 mL) and methanol
(15 mL) was added iodobenzene diacetate (4.67 g, 14.50 mmol), and the
mixture was stirred at room temperature overnight. The mixture was then
diluted with dichloromethane (50 mL), and the mixture was washed with
sodium hydroxide solution (1M, 100 mL), brine (100 mL), and subse-
quently dried (MgSO4). The organics were then concentrated in vacuo
and purified by column chromatography on silica gel, eluting with a
gradient of ethyl acetate in heptane from 0% to 100%. Fractions contain-
ing product were concentrated in vacuo to leave a pink oil (4.30 g, 77%).
1H NMR (400 MHz, CDCl3) δ 7.92 (dd, 1H), 7.76 (dd, 1H), 7.65 (dd,
1H), 7.57ꢀ7.58 (m, 2H), 7.36ꢀ7.43 (m, 2H), 3.69 (t, 2H), 2.93 (t, 2H),
0.84 (s, 9H), 0.00 (s, 6H) ppm. LRMS: m/z 464ꢀ466 [M + H]+.
{2-[(3-{2-[(2-{[tert-Butyl(dimethyl)silyl]oxy}ethyl)thio]phe-
nyl}[1,2,4]triazolo[4,3-a]pyridin-6-yl)thio]phenyl}methanol
(17a). A solution of 16a (2.0 g, 4.31 mmol) in N,N-dimethylformamide
(70 mL) was degassed with argon. 2-Mercaptobenzyl alcohol (724 mg,
5.17 mmol), cesium carbonate (2.81 g, 8.62 mmol), and [1,10-bis-
(diphenylphosphino)ferrocene]dichloropalladium(II) (169 mg, 431 μmols)
were then added, and the mixture was degassed an additional time. The
mixture was then heated to 90 ꢀC for 4 h. The mixture was concentrated
in vacuo, and the residue was partitioned between ethyl acetate (50 mL)
and water (100 mL). The aqueous phase was then re-extracted
with ethyl acetate (50 mL). The combined organics were washed with
brine (100 mL), dried (MgSO4), and concentrated in vacuo. The
residue was purified by column chromatography on silica gel, eluting
with ethyl acetate in heptane in a gradient from 0% to 100%. Fractions
containing product were concentrated in vacuo to leave a brown gum
Methyl 2-[(2-{[tert-Butyl(dimethyl)silyl]oxy}ethyl)thio]ben-
zoate (22). Sodium hydride (60% in oil, 1.70 g, 29.9 mmol) was
washed with heptane and subsequently suspended in N,N-dimethylfor-
mamide (25 mL). To this was added methyl thiosalicylate (5.0 g,
29.9 mmol), and the mixture was allowed to stir for 1 h. A solution of
(2-bromoethoxy)-tert-butyldimethylsilane (7.16 g, 29.9 mmol) in N,N-
dimethylformamide (5 mL) was added dropwise, and the mixture was
allowed to stir overnight at room temperature. The mixture was diluted
with ethyl acetate (100 mL), and the organics were washed with brine
(100 mL) and dried (MgSO4). Concentration in vacuo gave a colorless
1
oil (7.90 g, 81%). H NMR (400 MHz, CDCl3) δ 7.85 (dd, 1H),
7.32ꢀ7.39 (m, 2H), 7.97 (dt, 1H), 3.85 (s, 3H), 3.78 (t, 2H), 3.03
(t, 2H), 0.83 (s, 9H), 0.00 (s, 6H) ppm. LRMS: m/z 327 [M + H]+.
{2-[(2-{[tert-Butyl(dimethyl)silyl]oxy}ethyl)thio]phenyl}me-
thanol (23). To a solution of 22 (7.90 g, 24.37 mmol) in tetrahydrofuran
(50 mL) at 0 ꢀC was added lithium aluminum hydride (1 M in tetra-
hydrofuran, 24.37 mL, 24.37 mmol), and the mixture was allowed to warm
to room temperature with stirring overnight. Water (1 mL) was added
1
(1.32 g, 59%). H NMR (400 MHz, CDCl3) δ 7.78 (d, 1H), 7.24
(d, 1H), 7.52ꢀ7.64 (m, 4H), 7.38 (dd, 1H), 7.33 (dd, 1H), 7.27ꢀ7.30
(m, 2H), 7.17 (dd, 1H), 4.85 (s, 2H), 3.67 (t, 2H), 2.90 (t, 2H), 0.84
(s, 9H), 0.00 (s, 6H) ppm. LRMS: m/z 524 [M + H]+.
7811
dx.doi.org/10.1021/jm200677b |J. Med. Chem. 2011, 54, 7797–7814

Journal of Medicinal Chemistry
ARTICLE
1-{2-[(3-{2-[(2-{[tert-Butyl(dimethyl)silyl]oxy}ethyl)thio]-
phenyl}[1,2,4]triazolo[4,3-a]pyridin-6-yl)thio]phenyl}met-
hanamine (2a). To a solution of 17a (1.32 g, 2.52 mmol) in dichlor-
omethane (20 mL) was added N,N-diisopropylethylamine (0.88 mL,
5.05 mmol) followed by methanesulfonic anhydride (666 mg, 3.79 mmol),
and the mixture was stirred for 2 h at room temperature. The solution
was then poured into a solution of methanolic ammonia (7 M, 36 mL,
252.2 mmol), and the mixture was stirred for 4 h. The solvent was
removed in vacuo, and the residue was partitioned between ethyl acetate
(50 mL) and ammonium hydroxide (20 mL). The organics were then
washed with brine (50 mL), dried (MgSO4), and concentrated in vacuo.
Purification was effected using column chromatography on silica gel,
eluting from 0% to 10% methanol in (99:1) ethyl acetate with ammonia.
Fractions containing product were concentrated in vacuo to leave a
brown oil (904 mg, 68%). 1H NMR (400 MHz, CDCl3) δ 7.82ꢀ7.83
(m, 1H), 7.76 (dd, 1H), 7.63 (dd, 1H), 7.58 (dd, 1H), 7.52 (dd, 1H),
7.45 (d, 1H), 7.39 (dd, 1H), 7.27ꢀ7.32 (m, 1H), 7.21ꢀ7.23 (m, 2H),
7.16 (dd, 1H), 4.07 (s, 2H), 3.66 (t, 2H), 2.90 (t, 2H), 0.85 (s, 9H), 0.00
(s, 6H) ppm. LRMS: m/z 523 [M + H]+.
2-Chloro-4-hydrazinophenol (4b). To a suspension of 4-ami-
no-2-chlorophenol (25 g, 174 mmol) in water (40 mL) was added
concentrated hydrochloric acid (63 mL, 783 mmol). The mixture was
cooled to ꢀ10 ꢀC under nitrogen. A solution of sodium nitrite (12.0 g,
174 mmol) in water (35 mL) was added dropwise. After addition was
complete the mixture was stirred for an additional 30 min at ꢀ10 ꢀC.
Tin(II) chloride dehydrate (98.1 g, 435 mmol) was dissolved in 6 M
hydrochloric acid (100 mL), and the diazotization mixture was poured
into this solution. The mixture was stirred for 3.5 h at ꢀ10 ꢀC. The solids
were then filtered off and dried overnight. The material was not purified
further but carried over into the next step. LRMS: m/z 159 [M + H]+.
4-(5-Amino-3-tert-butyl-1H-pyrazol-1-yl)-2-chlorophenol
(12). To a solution of 4b (174 mmol) (carried over from the previous
step) in ethanol (250 mL) was added 4,4-dimethyl-3-oxopentanenitrile
(21.9 g, 174 mmol), and the mixture was heated to reflux overnight. The
solvent was removed in vacuo. The residue was then triturated using
diethyl ether (125 mL) and heptane (250 mL). After the mixture was
stirred vigorously for 2 h a fine powder was obtained. The powder was
filtered off and resuspended in diethyl ether (100 mL) and heptane
(200 mL), and the mixture was stirred for an additional 2 h. The solid
was filtered off and dried in a vacuum oven (40 ꢀC) to afford the desired
compound as a brown solid (22 g, 48% over two steps). 1H NMR (400
MHz, DMSO-d6) δ 11.10 (br, 1H), 7.10ꢀ7.67 (m, 3H), 5.61 (s, 1H),
1.28 (s, 9H) ppm. LRMS: m/z 266ꢀ268 [M + H]+.
phenyl chloroformate (1.46 mL, 11.6 mmol). After addition was complete
the mixture was allowed to warm to room temperature for 3 h. The
mixture was then partitioned between ethyl acetate (100 mL) and water
(100 mL). The organics were separated, washed with brine (40 mL),
dried (Na2SO4), and concentrated in vacuo to yield an orange oil. The
oil was triturated with heptane (60 mL) and the triturate stirred for 1 h.
The desired product was then filtered off as a white solid (4.5 g, 86%).
1H NMR (400 MHz, CDCl3) δ 7.55 (d, 1H), 7.37ꢀ7.42 (m, 2H),
7.27ꢀ7.32 (m, 2H), 7.23ꢀ7.26 (m, 1H), 7.10ꢀ7.18 (m, 2H), 7.02
(d, 1H), 6.46 (br, 1H), 1.36 (s, 9H), 1.07 (s, 9H), 0.28 (s, 6H) ppm.
LRMS: m/z 500 [M + H]+.
1-[3-tert-Butyl-1-(3-chloro-4-hydroxyphenyl)-1H-pyrazol-
5-yl]-3-{2-[(3-{2-[(2-{[tert-butyl(dimethyl)silyl]oxy}ethyl)thio]-
phenyl}[1,2,4]triazolo[4,3-a]pyridin-6-yl)thio]benzyl}urea(36).
To a solution of 31c (100 mg, 0.2 mmol) in dimethylsulfoxide (2 mL)
was added 2a (105 mg, 0.2 mmol), and the mixture was stirred overnight
at room temperature. The mixture was diluted with ethyl acetate
(30 mL) and water (30 mL). The organics were separated, washed with
brine (20 mL), dried over (Na2SO4), filtered, and concentrated in vacuo.
The residue was purified using column chromatography on silica gel,
eluting with ethyl acetate in cyclohexane, 0ꢀ100%. Product containing
fractions were combined and concentrated in vacuo to give the desired
product as a white solid (110 mg, 68%). 1H NMR (400 MHz, methanol-d4)
δ 7.78ꢀ7.82 (m, 1H), 7.72ꢀ7.76 (m, 2H), 7.62ꢀ7.67 (m, 1H), 7.58
(dd, 1H), 7.29ꢀ7.49 (m, 7H), 7.21ꢀ7.24 (m, 1H), 7.02 (d, 1H), 6.28
(s, 1H), 4.51 (s, 1H), 3.73 (t, 2H), 3.00 (t, 2H), 1.34 (s, 9H), 0.84
(s, 9H), 0.00 (s, 6H) ppm. LRMS: m/z 814 [M + H]+.
1-[3-tert-Butyl-1-(3-chloro-4-hydroxyphenyl)-1H-pyrazol-
5-yl]-3-{2-[(3-{2-[(2-hydroxyethyl)thio]phenyl}[1,2,4]triaz-
olo[4,3-a]pyridin-6-yl)thio]benzyl}urea (1ab). To a solution
of 36 (90 mg, 0.1 mmol) in methanol (10 mL) was added triethylamine
trihydrofluoride (18 μL, 0.1 mmol), and the mixture was stirred at
room temperature overnight. The mixture was concentrated in vacuo.
Then 7 M methanolic ammonia (20 mL) was added, and the reaction
mixture was concentrated again in vacuo. The residue was preabsorbed
onto silica gel and purified using column chromatography, eluting with
a gradient of ammonia in dichloromethane (0.3:99.7) to a mixture of
dichloromethane, methanol, and ammonia (94.5:5:0.5). Product con-
taining fractions were combined and concentrated in vacuo to give the
1
desired product as a white solid (51 mg, 73%). H NMR (400 MHz,
methanol-d4) δ 7.72ꢀ7.76 (m, 1H), 7.64ꢀ7.68 (m, 2H), 7.58ꢀ7.63
(m, 1H), 7.53 (d, 1H), 7.37ꢀ7.44 (m, 3H), 7.23ꢀ7.36 (m, 4H), 7.17
(d, 1H), 6.97 (d, 1H), 6.23 (s, 1H), 4.46 (s, 2H), 3.56 (d, 2H), 2.94
(d, 2H), 1.29 (s, 9H) ppm. LRMS: m/z 700 [M + H]+. Anal.
(C35H34ClN7O3S2) C, H, N.
3-tert-Butyl-1-(4-{[tert-butyl(dimethyl)silyl]oxy}-3-chloro-
phenyl)-1H-pyrazol-5-amine (3f). To a solution of 12 (40 g,
0.15 mol) in N,N-dimethylformamide (100 mL) were added tert-butyl-
(dimethyl)silyl chloride (23.8 g, 0.16 mol) and imidazole (11.2 g,
0.17 mol). The mixture was then stirred at room temperature overnight.
The reaction was quenched by addition of methanol (40 mL), and sub-
sequently the mixture was concentrated in vacuo. The residue was partitioned
between ethyl acetate (300 mL) and aqueous saturated sodium bicarbonate
solution (100 mL). The aqueous was then re-extracted with ethyl acetate
(2 ꢁ 300 mL). The organic washes were then combined and dried
(Na2SO4). After concentration in vacuo, the residue was purified by
column chromatography on silica gel, eluting with ethyl acetate in
hexane (1:6). Fractions containing product were concentrated in vacuo
and then crystallized from ethyl acetate and hexane, giving fine white
’ ASSOCIATED CONTENT
S
Supporting Information. Additional experimental pro-
b
cedures and analytical data for intermediates and final com-
pounds. This material is available free of charge via the Internet at
http://pubs.acs.org.
Accession Codes
†PDB codes are the following: 2yis for 1ab, 2yiw for 1a, and 2yix
for 41.
1
needle crystals (9.6 g, 17%). H NMR (400 MHz, DMSO-d6) δ 7.64
(d, 1H), 7.48 (dd, 2H), 7.10 (d, 1H), 5.41 (s, 1H), 5.22 (br, 2H), 1.24
(s, 9H), 1.05 (s, 9H), 0.29 (s, 6H) ppm. LRMS: m/z 380ꢀ382 [M + H]+.
Phenyl [3-tert-Butyl-1-(4-{[tert-butyl(dimethyl)silyl]oxy}-
3-chlorophenyl)-1H-pyrazol-5-yl]carbamate (31c). To a 0 ꢀC
cooled solution of 3f (4.0 g, 10.5 mmol) in tetrahydrofuran (30 mL) was
added pyridine (1.20 mL, 14.7 mmol) followed by dropwise addition of
’ AUTHOR INFORMATION
Corresponding Author
*To whom correspondence should be addressed. Phone: +44
1304 643497. Fax: +44 1304 651987. E-mail: david.millan@
gmail.com.
7812
dx.doi.org/10.1021/jm200677b |J. Med. Chem. 2011, 54, 7797–7814

Journal of Medicinal Chemistry
ARTICLE
’ ACKNOWLEDGMENT
(11) Renda, T.; Baraldo, S.; Pelaia, G.; Bazzan, E.; Turato, G.; Papi,
A.; Maestrelli, P.; Maselli, R.; Vatrella, A.; Fabbri, L. M.; Zuin, R.;
Marsico, S. A.; Saetta, M. Increased activation of p38 MAPK in COPD.
Eur. Respir. J. 2008, 31, 62–69.
We thank Jonathan Duckworth for assistance with ADME
experiments. Deborah Bruce, Iva Navratilova, and Klaus Rumpel
are thanked for assistance with Biacore generated data. Suengil
Han, Eric Marr, Marie Anderson, and Stephen Irving are thanked
for assistance with X-ray crystallographic data. William Hood is
thanked for help with progression curve experiments and anal-
ysis. Joe Monahan is also thanked for useful advice and guidance.
(12) Underwood, D. C.; Osborn, R. R.; Kotzer, C. J.; Adams, J. L.;
Lee, J. C.; Webb, E. F.; Carpenter, D. C.; Bochnowicz, S.; Thomas, H. C.;
Hay, D. W. P.; Griswold, D. E. SB 217063, a potent p38 MAP kinase
inhibitor, reduces inflammatory cytokine production, airways eosinophil
infiltration, and persistence. J. Pharmacol. Exp. Ther. 2000, 293, 281–288.
(13) Medicherla, S.; Fitzgerald, M.; Spicer, D.; Woodman, P.; Ma,
J. Y.; Kapoun, A. M.; Chakravarty, S.; Dugar, S.; Protter, A. A.; Higgins,
L. S. p38α selective MAP kinase inhibitor, SD-282, reduces inflamma-
tion in a sub-chronic model of tobacco smoke-induced airway inflam-
mation. J. Pharmacol. Exp. Ther. 2008, 324, 921–929.
’ DEDICATION
‡This paper is dedicated to the memory of Mark H. Stefaniak.
(14) Duan, W.; Chan, J. H.; McKay, K.; Crosby, J. R.; Choo, H. H.;
Leung, B. P.; Karras, J. G.; Wong, W. S. F. Inhaled p38alpha mitogen-
activated protein kinase antisense oligonucleotide attenuates asthma in
mice. Am. J. Respir. Crit. Care Med. 2005, 171, 571–578.
(15) Array Biopharm. http://www.arraybiopharma.com (accessed
April 13, 2011).
’ ABBREVIATIONS USED
ATP, adenosine triphosphate;CNS, central nervous system;CRP,
C-reactive protein; COPD, chronic obstructive pulmonary disease;
CYP, cytochrome P450; DDI, drugꢀdrug interaction; DoA, dura-
tion of action; FEV1, forced expiratory volume in 1 s; hERG, human
ether-a-go-go-related gene; HLM, human liver microsomes; IL-1β,
interleukin 1β; IL-6, interleukin 6; LPS, lipopolysaccharide;
MMP-9, matrix metalloprotease 9; MINK, misshapen NIK related
kinase 1; MAP, mitogen activated kinase; PBMC, peripheral
blood mononuclear cells; PDB, Protein Data Bank; P-gp,
P-glycoprotein; PDE4, phosphodiesterase 4; PPB, plasma
protein binding; RA, rheumatoid arthritis; Ro5, rule-of-five; SLF,
simulated lung fluid; SAR, structureꢀactivity relationship; TBDMS,
tert-butyldimethylsilyl; THP, tetrahydropyran-2-yl; TNF-α, tumorne-
crosis factor α; VEGFR1, vascular endothelial growth factor receptor 1
(16) ClinicalTrials.gov. http://www.clinicaltrials.gov (accessed April 13,
2011).
(17) (a) MacNee, W.; Allan, R.; Jones, I.; De Salvo, M. C.; Tan, L. A
Randomised, Placebo Controlled Trial of 6 Weeks’ Treatment with a
Novel Oral p38 Inhibitor in Patients with COPD. Presented at the
European Respiratory Society Meeting, Barcelona, Spain, 2010. (b) Selness,
S. R.; Devraj, R. V.; Devadas, B.; Walker, J. K.; Boehm, T. L.; Durley, R. C.;
Shieh, H.; Xing, L.; Rucker, P. V.; Jerome, K. D.; Benson, A. G.; Marrufo,
L. D.; Madsen, H. M.; Hitchcock, J.; Owen, T. J.; Christie, L.; Promo, M. A.;
Hickory, B. S.; Alvira, E.; Naing, W.; Blevis-Bal, R.; Messing, D.; Yang, J.;
Mao, M. K.; Yalamanchili, G.; Vonder Embse, R.; Hirsch, J.; Saabye, M.;
Bonar, S.; Webb, E.; Anderson, G.; Monahan, J. B. Discovery of PH-797804,
a highly selective and potent inhibitor of p38 MAP kinase. Bioorg. Med.
Chem. Lett. 2011, 21, 4066–4071.
’ REFERENCES
(18) Chopra, P.; Kanoje, V.; Semwal, A; Ray, A. Therapeutic
potential of inhaled p38 mitogen-activated protein kinase inhibitors
for inflammatory pulmonary diseases. Expert Opin. Invest. Drugs 2008,
17, 1411–1425.
(1) Dinarello, C. A. Inflammatory cytokines: interleukin-1 and
tumour necrosis factor as effector molecules in autoimmune diseases.
Curr. Opin. Immunol. 1991, 3, 941–948.
(2) Lee, S. J.; Kavanaugh, A. A. Pharmalogical treatment of estab-
lished rheumatoid arthritis. Best Pract. Res., Clim. Rheumatol. 2003,
17, 811–829.
(19) Millan, D. S. What is the potential for inhaled p38 inhibitors in
the treatment of chronic obstructive pulmonary disease? Future Med.
Chem. 2011, 3, 1635ꢀ1645.
(3) Weisman, M.; Furst, D.; Schiff, M.; et al. A Double-Blind,
Placebo Controlled Trial of VX-745, an Oral p38 Mitogen-Activated
Protein Kinase (MAPK) Inhibitor in Patients with Rheumatoid Arthritis
(RA). Presented at the Annual Congress of the European League against
Rheumatism, Stockholm, Sweden, Jun 12ꢀ15, 2002; FRI0018.
(4) Schreiber, S.; Feagan, B.; D’Haens, G.; Colombel, J.-F.; Geboes,
K.; Yurcov, M.; Isakov, V.; Golovenko, O.; Bernstein, C. N.; Ludwig, D.;
Winter, T.; Meier, U.; Yong, C.; Steffgen, J. Oral p38 mitogen-activated
protein kinase inhibition with BIRB 796 for active Crohn’s disease: a
randomized, double-blind, placebo-controlled trial. Clin. Gastroenterol.
Hepatol. 2006, 4, 325–334.
(5) Goldstein, D. M.; Kuglstatter, A.; Lou, Y.; Soth, M. J. Selective
p38α inhibitors clinically evaluated for the treatment of chronic
inflammatory disorders. J. Med. Chem. 2010, 53, 2345–2353.
(6) Vertex Pharmaceuticals Inc: Vertex moves to re-allocate re-
sources from VX-745 in p38 MAP kinase program to accelerate
development of second generation drug candidates VX-702 and VX-
850. Press Release, September 24, 2001.
(7) Wood; Yong; Madwed; Gupta, A. Safety, pharmacokinetics and
pharmacodynamics of an oral p38 MAP kinase inhibitor (BIRB 796 BS),
administered once daily for 7 days. J. Allergy Clin. Immunol. 2002,
109, 993.
(8) Boehringer Ingelheim. http://www.boehringer-ingelheim.com
(accessed April 13, 2011).
(9) Mannino, D. M.; Buist, A. S. Global burden of COPD: risk
factors, prevalence, and future trends. Lancet 2007, 370, 765–773.
(10) Barnes, P. J.; Adcock, I. M. Glucocorticoid resistance in
inflammatory diseases. Lancet 2009, 373, 1905–1917.
(20) Glossop, P. A.; Lane, C. A. L.; Price, D. A.; Bunnage, M. E.;
Lewthwaite, R. A.; James, K.; Brown, A. D.; Yeadon, M.; Perros-Huguet,
C.; Trevethick, M. A.; Clarke, N. P.; Webster, R.; Jones, R. M.; Burrows,
J. L.; Feeder, N.; Taylor, S. C. J.; Spence, F. J. Inhalation by design: novel
ultra-long-acting β2-adrenoreceptor agonists for inhaled once-daily
treatment of asthma and chronic obstructive pulmonary disease that
utilize a sulfonamide agonist headgroup. J. Med. Chem. 2010, 53,
6640–6652.
(21) Lipinski, C. A.; Lombardo, F.; Dominy, B. W.; Feeney, P. J.
Experimental and computational approaches to estimate solubility and
permeability in drug discovery and development settings. Adv. Drug
Delivery Rev. 1997, 23, 3–25.
(22) Regan, J.; Capolino, A.; Cirillo, P. F.; Gilmore, T.; Graham,
A. G.; Hickey, E.; Kroe, R. R.; Madwed, M.; Moriak, M.; Nelson, R.;
Pargellis, C. A.; Swinamer, A.; Torcellini, C.; Tsang, M.; Moss, N.
Structureꢀactivity relationships of the p38α MAP kinase inhibitor 1-
(5-tert-butyl-2-p-tolyl-2H-pyrazol-3-yl)-3-[4-(2-morpholin-4-yl-ethoxy)-
napthalen-1-yl]urea (BIRB 796). J. Med. Chem. 2003, 46, 4676–4686.
(23) Regan, J.; Paregllis, C. A.; Cirillo, P. F.; Gilmore, T.; Hickey,
E. R.; Peet, G. W.; Proto, A.; Swinamer, A.; Moss, N. The kinetics of
binding to p38 MAP kinase by analogues of BIRB 796. Bioorg. Med.
Chem. Lett. 2003, 13, 3101–3104.
(24) McClure, K. F.; Abramov, Y. A.; Laird, E. R.; Barberia, J. T.; Cai,
W.; Carty, T. J.; Cortina, S. R.; Danley, D. E.; Dipesa, A. J.; Donahue,
K. M.; Dombroski, M. A.; Elliott, N. C.; Gabel, C. A.; Han, S.; Hynes,
T. R.; LeMotte, P. K.; Mansour, M. N.; Marr, E. S.; Letavic, M. A.; Pandit,
J.; Ripin, D. B.; Sweeney, F. J.; Tan, D.; Tao, Y. Theoretical and
7813
dx.doi.org/10.1021/jm200677b |J. Med. Chem. 2011, 54, 7797–7814

Journal of Medicinal Chemistry
ARTICLE
experimental design of atypical kinase inhibitors: application to p38
MAP kinase. J. Med. Chem. 2005, 48, 5728–5737.
(25) Braganza, J. F.; Letavic, M. A.; McClure, K. F. Preparation of
Triazolo[4,3-a]pyridines as MAP Kinases, in Particular p38 Kinase
Inhibitors, and Their Use as Antiinflammatory Agents. WO2004072072,
2004.
(26) Liu, H.; Kuhn, C.; Feru, F.; Jacques, S. L.; Deshmukh, G. D.; Ye,
P.; Rennie, G. R.; Johnson, T.; Kazmirski, S.; Low, S.; Coli, R.; Ding,
Y.-H.; Cheng, A. C.; Tecle, H.; English, J. M.; Stanton, R.; Wu, J. C.
Enhanced selectivity profile of pyrazole-urea based DFG-out p38α
inhibitors. Bioorg. Med. Chem. Lett. 2010, 20, 4885–4891.
(27) Fisher, M. B.; Paine, M. F.; Strelevitz, T. J.; Wrighton, S. A. The
role of hepatic and extrahepatic UDP-glucuronosyltransferases in human
drug metabolism. Drug Metab. Rev. 2001, 33, 273–297.
7814
dx.doi.org/10.1021/jm200677b |J. Med. Chem. 2011, 54, 7797–7814