An algorithm-directed two-component library synthesized via solid-phase methodology yielding potent and orally bioavailable p38 MAP kinase inhibitors

Article abstract of DOI:10.1021/jm011132l

Previously we reported the identification of RPR200765A, a potent orally bioavailable pyridine - imidazole inhibitor of p38 mitogen-activated protein (MAP) kinase which suppressed paw swelling and joint pathology in streptococcal cell wall-induced arthrit

Full text of DOI:10.1021/jm011132l

J . Med. Chem. 2002, 45, 2173-2184
2173
An Algor ith m -Dir ected Tw o-Com p on en t Libr a r y Syn th esized Via Solid -P h a se
Meth od ology Yield in g P oten t a n d Or a lly Bioa va ila ble p 38 MAP Kin a se
In h ibitor s
J effrey M. McKenna,* Frank Halley,* J ohn E. Souness, Iain M. McLay, Stephen D. Pickett, Alan J . Collis,
Kenneth Page, and Imtiaz Ahmed
Centre de Recherche de Paris, Aventis Pharma S.A., 94403 Vitry sur Seine CEDEX, France
Received December 20, 2001
Previously we reported the identification of RPR200765A, a potent orally bioavailable pyridine-
imidazole inhibitor of p38 mitogen-activated protein (MAP) kinase which suppressed paw
swelling and joint pathology in streptococcal cell wall-induced arthritis. Herein, we report the
use of solid-phase combinatorial organic synthesis for the parallel processing of a related
pyrimidine-imidazole-based library with two points of structural variability. We report also
that the application of a computer algorithm, the Monte Carlo Monomer Selection, maximized
both the combinatorial synthetic efficiency and the bioavailability of the final compounds. In
conjunction with the synthetic protocols, the polymer-supported quench technique was applied
to the purification of the final compounds. Through rapid evaluation of the library using a p38
kinase assay and permeability assays, it was possible to identify a number of potent and orally
bioavailable p38 MAP kinase inhibitors suitable for further biological investigation.
In tr od u ction
(SPCOS) of a pyrimidinoimidazole-based library in
which we made use of a computer algorithm in the
library design to engender greater bioavailability to the
series. Reaction transformations were monitored by
single-bead FT-IR,8 and a polymer-supported quench
(PSQ) technique9 following cleavage from the solid
support allowed for rapid purification of the library.
Thereafter, rapid biological evaluation identified a
number of potent and orally bioavailable p38 MAP
kinase inhibitors.
Pro-inflammatory cytokines tumor necrosis factor R
(TNF-R) and interleukin-1â (IL-1â) are implicated in the
pathology of several chronic inflammatory diseases1
including rheumatoid arthritis.2 They are released from
several inflammatory cells, particularly mononuclear
phagocytes. The synthesis of both TNF-R and IL-1â is
regulated by a kinase referred to as p38, which is a
mitogen-activated protein (MAP) kinase belonging to the
serine/threonine kinase superfamily.3 It is activated by
inflammatory stimuli such as lipopolysaccharide (LPS)
as well as extracellular stresses including heat, UV
light, and osmotic stress and is often referred to as a
stress-activated protein kinase.4 In its activated state,
p38 phosphorylates intracellular protein substrates
which regulate the biosynthesis of TNF-R and IL-1â
post-transcriptionally. Previous studies have demon-
strated that specific inhibitors of this kinase suppress
production of these cytokines from mononuclear cells,
making p38 a very attractive target for novel antiiin-
flammatory therapy. Thus, our aim was to identify
small-molecule inhibitors of p38, for the treatment of
inflammatory conditions resulting from excessive cy-
tokine production.5
Our series of imidazoyl-based cyclic acetals as func-
tional inhibitors of TNF-R originated from our previous
knowledge that the 1,3-dioxane was stable when at-
tached to the imidazole nucleus.6 This knowledge when
coupled to the postulate that the 4-pyridyl group at-
tached to the imidazole nucleus may be a crucial entity
for such biological activity led us to synthesize a large
array of 4-pyridyl-5-(4-fluorophenyl)imidazoyl cyclic ac-
etals.7 Herein, we report our later efforts in the realiza-
tion via solid-phase combinatorial organic synthesis
Ch em istr y
The investigations centered around 4-pyridyl-5-(4-
fluorophenyl)imidazoyl cyclic acetals led to the identi-
fication of RPR200765A (3; Figure 1) as a novel p38
inhibitor and to its being accepted as a development
candidate.7 Our follow-up strategy for this initial work
was to synthesize specifically (i) solution-based librar-
ies10 in which substituted pyrimidines, replacing the
pyridine nucleus of 3, were combined with the most
favorable substituted dioxanes identified during the
discovery of 3 and (ii) a larger library in which we
sought variations in both the amide portion of 3 and
the pyrimidine nucleus of the molecule simultaneously.
This larger library is the subject of the work reported
here and was prepared via SPCOS (Figure 1).
As already stated, we wished to introduce our diver-
sity at the 2-position of the pyrimidine (Nu , R3R4N-)
and the amide unit (Am , R1R2N-) of the dioxane. We
envisaged that the former diversity Nu could be intro-
duced through a nucleophilic cleavage from an ap-
propriately activated resin while the latter Am could
be introduced at a prior point on the resin-bound
substrate. We hoped to maximize the bioavailability
within this series through careful choice of complimen-
tary substituents, a strategic decision which drove us
to develop the Monte Carlo Monomer Selection (MCMS)
* To whom correspondence should be addressed. Please use
e-mail: (F.H.) frank.halley@aventis.com. Phone: +33 (0) 15893 3614.
Fax: +33 (0) 15893 8014.
10.1021/jm011132l CCC: $22.00 © 2002 American Chemical Society
Published on Web 04/27/2002

2174 J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 11
McKenna et al.
F igu r e 1. SPCOS-targeted compounds, trans-1 and cis-2 isomers; based on 3, RPR200765A.
algorithm to direct these variations. It should be noted
that these compounds can exist as cis or trans isomers,
according to their relative stereochemistry about the
dioxane ring. For this study we decided to work with
both isomers, treating each as a separate library. All of
the library design was carried out for the trans isomers
1 and then transferred to the cis isomer 2 library.
F igu r e 2. A virtual data set, a 5 × 6 matrix, with shading
indicating that a monomer coselection (AX, BX) is possible
under the chosen criteria. Darker shading indicates the chosen
subsets, solutions 1 and 2, respectively (vide infra).
In this work we utilized two sets of monomers, both
amines: the first set (Am , R1R2NH) for the amide-
forming reaction, the second set (Nu , R3R4NH) for the
nucleophilic displacement. MCMS applies a scoring
function, controlled by a series of weights, to drive the
monomer selection toward a solution which best satis-
fies a set of selection objectives. It is possible to define
many different objectives for the function, but in this
case the desire was to have a library of acceptable
products (i.e., ones which calculation suggested would
be absorbed from the intestine), which was as close to
a 20 × 20 combinatorial solution as possible and which
allowed each resin to be used exactly 20 times. This
latter criterion was deemed essential on synthetic
grounds as the preparation of the resin is difficult and
time-consuming and to leave resin unused would be
highly wasteful. The method of synthesis for these
libraries did not necessitate a full combinatorial selec-
tion. However, the synthetic ease increased dramatically
the closer the design was to a combinatorial one. Thus,
the criteria with which we set out to constrain the
library can be summarized in the following fashion.
(a ) Syn th etic Con str a in ts. Am (amide variant) to
be equal to 20 (i.e., ideally 20 resins to be prepared) and
Nu (variation in the amine nucleophile) equal to or
greater than 20 (i.e., use whatever nucleophiles neces-
sary, but drive toward the ideal of a combinatorial
solution) while selecting at least 400 compounds and
using each amide resin at least 20 times.
(b) Absor p tion /Bioa va ila bility Requ ir em en ts. A
polar surface area (PSA)11 under 140 Å2 and satisfaction
of a modified rule-of-five in which compounds were
acceptible only if cLogP < 6.2, MW < 550, number of
H-bond donors < 5, and number of H-bond acceptors <
12. These modified values differ from those of the
classical Lipinski rule-of-five12 but were defined through
previous experience with bioavailable compounds in-
vestigated during the identification of 3.
The importance of MCMS in monomer selection has
been reported elsewhere,13 and here is illustrated by the
simplified example in Figure 2.
Assume we require nine products originating from a
virtual library of a possible 30 compounds (a 5 × 6
matrix). It can be seen from Figure 2, in which any
shading indicates a possible selection according to the
operators chosen (e.g., PSA or cLogP), that no 3 × 3
subsets exist; however, several 3 × 4 solutions exist. In
solution 1, as indicated by the heavy shading within the
left-hand figure, one B reagent (B6) is used only once
and the A groups (A1, A4, and A5) are used between
two and four times. However, a solution also exists in
which each A is used three times (A1, A3, and A4) and
each B (B1, B3, B4, and B5) at least twice, indicated
in solution 2. Thus, this latter solution would then be
preferred from a synthetic point of view, in which the
reagents are used maximally, creating a good combina-
torial solution. Hence, in a similar vein, the virtual
library conceived for this work initially contained 1485
compounds (33 Nu × 45 Am matrix). However, follow-
ing filtering by way of PSA and the rule-of-five, it was
reduced to 770. Still, within this subset no 20 × 20 array
could be achieved. A further complication to the ideal
solution being realized was (i) the desire to include some
known active monomers for both Nu and Am and (ii)
the wish to combine each amide resin with 20 nucleo-
philes to use the precious resin efficiently (see Figure 3
for a list).
These multiple objectives were converted into weight-
ings for MCMS which, after a few minutes on an SG
R10000 processor, provided an optimized solution which
utilized 21 Am groups and 24 Nu groups. Within this
selection 89% of the compounds, or 449 of 504 possible,
satisfied the library design criteria. Other programs
have been reported for deriving optimized monomer
selections: PLUMS,14 VOLGA,15 GALOPED,16 and
MFA.17 Invariably these techniques are designed to
produce a full combinatorial solution. MCMS differs in
that it is applicable to situations where a near-combi-
natorial solution is acceptable.
From a chemical strategy viewpoint the retrosynthetic
analysis of the library mandated that key intermediates
were prepared in solution (vide infra), attached to the

Identification of Potent p38 MAP Kinase Inhibitors
J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 11 2175
F igu r e 3. Targeted structure with the nucleophile (NX) and amide (AX) monomers as defined by the MCMS algorithm. A1
(morpholine) was only used as the amide component for the cis-dioxane series and at the expense of A18 (propylamine).10
polymer, and subsequently functionalized. This desire
when coupled with a requirement to design a single-
compound-single-well methodology meant that we were
to follow a parallel-processing route. An advantage of
this procedure was seen in the analysis of the library,
as each well would be expected to return a single
product of defined identity. Thus, mass spectral analysis
of the crude products would confirm the identity of each
material, and HPLC analysis could be carried out for
product purity. Thus, we used parallel LC-MS in
conjunction with ELS detection to analyze reaction
efficiency, product identity, and homogeneity for all of
the reaction wells. A limiting factor within our series
and hence chemistry strategy was a thermodynamic
mixture of dioxanes being formed during a transaceta-
lation procedure. This both could not be avoided and

2176 J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 11
McKenna et al.
Sch em e 1a
a
Reagents: (i) NaOH, MeI, EtOH; (ii) NaHMDS, THF, 4-F(C6H4)CO2Me; (iii) 48% HBr, DMSO; (iv) NH4OAc, glyoxal dimethyl acetal,
TBME, MeOH; (v) DMF, TsOH, CH(OMe)3, 9.
had to be carried out in solution with the components
being separated chromatographically and then utilized
as the core monomers in our library. Our synthetic
route18 to the imidazole C-2 dimethyl acetal 8 installed
both the acetal and imidazole C-2 carbons in a single
step, using glyoxal dimethyl acetal and ammonium
acetate together with the requisite benzyl 7, which itself
originated from methyl 4-fluorobenzoate and 2-mer-
capto-4-methylpyrimidine (4). This procedure was found
to be preferable to the alternative strategy in which the
acetal was installed through a lithiation protocol utiliz-
ing an SEM-protected imidazole (Scheme 1).7
yielded a spectrum in which the peaks were still
broadened and hence not predictive in either reaction
quantification or analysis, and thus we looked at
alternate means to both follow and characterize our
transformations. Single-bead FT-IR microspectroscopy8
proved to be an excellent qualitative and quantitative
measure for the transformations carried out. To deter-
mine the loading of the resin, oxidation of substrate 17
using a 5-fold excess of mCPBA21,24 and thereafter dis-
placement of the bound substrate from the now acti-
vated sulfonyl resin with sodium thiomethoxide regen-
erated our 2-thiomethyl derivative 11. We determined
the loading of the substrate from the recovered material
to be 0.75 mmol of substrate/g of resin, while FT-IR
analysis of the spent resin showed that complete cleav-
age had occurred. We also conducted this exercise using
benzylamine as the nucleophile, yielding 19, both indi-
cating that cleavage was possible from the resin using
an amine and further determining the loading obtained
from the thiomethoxide experiment (see Scheme 2).
Thus, through a four-step route from commercial
starting material 4, we were able to synthesize the
parent dimethyl acetal 8 in good yield. A simple acid-
catalyzed transetherification under thermodynamic con-
trol yielded the desired 1,3-dioxanes as a mixture (3:2
cis-12/trans-11). The reaction mixture also contained
varying amounts of aldehyde 10; however, all of the
components were separable by flash chromatography,
and further the dioxanes could be interconverted through
reequilibration when necessary. Following oxidation to
their respective sulfonyl derivatives 13 and 14, both the
cis and trans isomers were attached to a Merrifield thiol
resin (16).19 The two resulting fictionalized resins 17
and 18 were then ready for parallel manipulation and
library generation.20 It has been well documented21 that
the sulfonyl group in the 2-position of the pyrimidine is
more reactive than the corresponding chloro derivative;
both of these are still far more reactive than the
corresponding thioether. Hence, the intrinsic advantage
of this strategy allowed us to selectively activate the
2-position and displace the substituent with an alter-
nate thiol, thereby inactivating the substrate to nucleo-
philic cleavage until we wished to cleave our products
from the resin, which would again be through an
oxidation-nucleophilic displacement protocol.22 How-
ever, this “traceless” technology was not without a
problem as simple reaction analysis was not possible
by direct cleavage from the resin. Analysis of the resin
using magic-angle spinning 1H NMR spectroscopy23
Thus, with resin loading and synthetic viability in
hand we could continue on our pathway. Hydrolysis of
the ester 17 was achieved using sodium hydroxide in
THF, giving the carboxylic acid 20 on the solid support,
the first point at which we wished to introduce the
diversity into our route. Thus, division and subsequent
reaction of this acid under optimized conditions was
carried out. The amide formation yielding 21 proceeded
well either through formation of the acid chloride, oxalyl
chloride, and catalytic DMF or by using excess HATU.
Interestingly EDCI proved to be of limited value in our
hands especially for nonnucleophilic amines, which gave
slow and incomplete reactions even over extended
reaction times. Oxidation to the sulfonylpyrimidine
resin 22 could be carried out using mCPBA; however,
there were certain chemoselectivity problems that this
reagent would present in a number of cases in our
library. Thus, to avoid oxidation of alkenes and py-
ridines under these latter conditions, we turned to
OXONE as the preferred oxidant for our libraries.25
Once the resin was cleanly oxidized, cleavage was best

Identification of Potent p38 MAP Kinase Inhibitors
J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 11 2177
Sch em e 2a
a
Reagents: (i) mCPBA, CH2Cl2; (ii) Merrifield thiol resin 16, DMF, NaH; (iii) DMF, NaSMe; (iv) PhCH2NH2, DME; (v) NaOH, THF.
Sch em e 3a
a
Reagents: (i) NaOH, H2O, methanol; (ii) CH2Cl2, (COCl)2, catalytic DMF, amines A1-A21; (iii) mCPBA, CH2Cl2, or OXONE, THF,
MeOH; (iv) NMP, amines N1-N22, 80 °C; then CH2Cl2, isocyanate resin.
achieved using an excess of nucleophile at an elevated
temperature. With DMF as solvent, we observed various
quantities of the 2-dimethylamino compound as an
impurity. Alternative inert solvents for this reaction
were dimethoxyethane (DME) and N-methylpyrrolidi-
none (NMP). However, under the automated conditions
we sought for cleavage of the library from the solid
support, NMP was preferred as it did not develop any
vapor pressure at 80 °C, the chosen reaction tempera-
ture. The conditions used for the nucleophilic cleavage
were a 10-fold excess of nucleophile in NMP for 16 h at
80 °C (Scheme 3).
detection, allowing us to determine reaction efficiency
and purity. The cleavage procedure from the solid
support yielded a 570-member library;26 analysis of
these showed that 85% of the compounds were of greater
than 85% purity, with 90% of these being of 100% purity
(Figure 4).
Biologica l Resu lts a n d Discu ssion
Dissolution of the cleaved products into 10 mM
solutions in DMF allowed rapid biological evaluation of
the library.7,10 Following parallel analyses of the com-
pounds in a primary functional assay, the CaCo-2
monolayer assay, and in an in vitro metabolic stability
study, an assessment of the compound bioavailability
was carried out in a combi-PK cassette-dosing experi-
ment. Our target, through the algorithm-based library
Following cleavage from the resin, the crude com-
pounds were subjected to a PSQ, involving an isocyanate
resin. Analysis of the members of the library was carried
out using a parallel LC-MS in conjunction with ELS

2178 J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 11
McKenna et al.
compounds when compared to the similar library in
which an empirical approach was taken.
It is also worth noting that this increased absorption
profile was not at the expense of in vitro potency. Within
the designed library, 85% of the compounds surpassed
the activity of the reference compound RPR200765A for
inhibition of p38 kinase activity while for the empirically
designed library only 60% incidence of increased potency
was achieved.
Following the triage of library compounds through the
outlined assays and decision points, we selected 56
compounds for evaluation in rat pharmacokinetic stud-
ies, with compounds evaluated as mixtures of 14 in a
combi-PK protocol at 1 mg/kg. With the benefit of these
results we were able to identify 10 compounds worthy
of in vivo efficacy studies. These compounds were
relatively structurally diverse and potent and also
exhibited acceptable systemic exposure in the combi-
PK studies (Cp max >25 ng/mL), data that were then
confirmed through a single-compound PK study to
determine relative absorption ranking. The following
data (Table 1) show the structural diversity, potency
both against the isolated enzyme and functionally, and
also the preliminary rat PK data for these compounds.
It is of great interest to see that the compounds which
progressed were obtained from both cis and trans
libraries and utilized six different Nu ’s and six different
Am ’s. This clearly demonstrates the strength of the lead
optimization library. Traditional medicinal chemistry,
in which one variable is changed at a time, could not
have produced such a large number of relatively diverse,
progressable compounds in such a short period of time.
F igu r e 4. Overall purity profile as delineated by parallel LC-
MS and ELS.
Analysis of the results (Table 1) shows the overall
trend that the cis isomers (entries A-C) are less potent
in vitro than the trans isomers (entries D-J ) although
similar in vivo potency is achieved.
In the p38 assay, compounds showing 10-fold im-
proved potencies over 3 have been identified (comparing
entries D-J with entry 3). In the in vitro whole cell
assay, compounds with greater than 10-fold improved
potencies compared to 3 have been synthesized (com-
paring entries D, E, G, I, J , and 3), and compound D
shows a 300-fold improvement in potency over 3.
Overall, four compoundssentries D, E, G, I, and J s
exhibited improved enzyme and whole cell potencies
compared to 3, the other two (entries B and C) being
comparable. Furthermore, in the in vivo LPS-induced
TNF-R release assay in the rat, compounds having 10-
fold improved ED50 compared to RPR200765A have been
identified (compare entries B and F with entry 3), and
the other compounds (except entry E) have ED50 values
comparable to that of 3.
The pharmacokinetic parameters for most of the
compounds are comparable to or better than those
obtained with the free base of RPR200765. Entries A-C
and F -J showed higher Cp(max) values when com-
pared with entry K, and entries D-J show better AUCs
than entry K. Compounds I and J have PK parameters
comparable to that of the bioavailable, optimized me-
sylate salt RPR200765A (entry 3).
F igu r e 5. Comparison of experimental CaCo-2 absorption
data for the empirically designed library and the algorithm-
driven SPCOS library, the latter designed to have high
absorption properties.
design, was to maximize the bioavailability within the
series. It is not easy to demonstrate the success of the
approach without constructing an identical library with
randomly selected monomers. However, at least within
the in vitro setting of the CaCo-2 monolayer assay,
which is a model for intestinal absorption, the designed
objective appears to have been born out with 80% of the
compounds exhibiting a high absorption (>20%).27 To
provide some sort of benchmark, these results were
compared with those of a previously prepared library
of structurally related p38 inhibitors. These had been
designed empirically with only a very simple constraint
on molecular weight (under MW 550). The overall profile
of the algorithm-driven library with regard to the
CaCo-2 assay is compared to that of the manually
chosen library in Figure 5.
The shift of compounds into the higher absorption
category is striking. Thus, through the use of the
algorithm we appear to have significantly improved the
distribution of the library toward highly permeable
The 10 compounds identified have different hydro-
phobic and hydrophilic side chains attached to the
pyrimidine ring and also different small alkyls and alkyl
ether amides linked to the dioxane ring. Secondary

Identification of Potent p38 MAP Kinase Inhibitors
J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 11 2179
Ta ble 1. In Vitro, Functional, in Vivo, and Pharmacokinetic Parameters of the Compounds Following Oral Dosing in the Rat at 1
mg/kg

2180 J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 11
Ta ble 1 (Continued)
McKenna et al.
F igu r e 6. Change in lateral ankle width (mean, injected - uninjected) from sensitized Lewis rats over the 5 days following
intravenous challenge with SCW (100 mg) or phosphate-buffered saline (PBS) on day 21. Vehicle RPR238677 (10, 30, and 100
(mg/kg)/day) or dexamethasone (0.3 (mg/kg)/day) was given orally (as a split dose, b.i.d.) for 5 days from 1 day prior to intravenous
challenge with SCW. n ) 10 rats/group except for PBS where n ) 5 rats/group and vehicle-treated where n ) 15 rats/group.
amines can be introduced in the right-hand side to make
a tertiary amide without loss in activity (entries D, G,
and J ), but only primary amines, linked to the pyrimi-
dine ring on the left-hand side, led to active compounds.
There is no obvious structure-activity relationship
(SAR) that can be discerned from these results, dem-
onstrating that they could not have been obtained as
rapidly following a traditional sequential approach. This
MCMS approach was used to maximize the different in
vitro, in vivo, and PK parameters of the compounds at
the same time, avoiding a more traditional, sequential
SAR approach. Compound J (RPR238677) was selected
for further evaluation on the basis of its good PK profile
and the possibility of preparing a salt formulation.
F igu r e 7. Anatomical assessment of radiographs. Mean (
SEM. One asterisk indicates p < 0.05 and two asterisks
Compound J was further evaluated in vivo in an SCW
(streptococcal cell wall) induced arthritis model in the
Lewis rat. It suppressed paw swelling with an ED50 of
10-30 mg/kg (Figure 6), Furthermore, administration
of RPR238677 was associated with marked preservation
of bone anatomy and a reduction in erosion areas. This
was particularly apparent at the 30 and 100 mg/kg dose
indicate p < 0.01 compared to vehicle control. One-way
ANOVA with correction for Dunnett’s multiple t-test.
levels, but there was some evidence of conservation of
joint anatomy at 10 mg/kg. Periosteal inflammation was
less apparent in the 30 and 100 mg/kg dose groups
(Figure 7).

Identification of Potent p38 MAP Kinase Inhibitors
J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 11 2181
aqueous HBr (3 equiv). The reaction mixture was stirred for
3 h at that temperature and then cooled to ambient temper-
ature. The mixture was poured into ice/water, a solution of
sodium acetate then added dropwise, and subsequently the
pH then adjusted to 7 using solid sodium bicarbonate. The
solid was filtered off, washed with water, and dried, yielding
7 (82%): mp 73-75 °C; 1H NMR (CDCl3) 2.32 (s, 3H), 7.22
(m, 2H), 7.95 (m, 2H), 8.84 (d, J ) 4.8 Hz, 1H); 13C NMR
(CDCl3) 6.5 (d, J ) 21.9 Hz), 13.9, 112.7, 117.64 (d, J ) 4.8
Hz), 129.3, 132.2 (d, J ) 9.8), 157.0, 159.5, 166.7 (d, J ) 258.9
Hz), 174.1, 192.9, 193.5; MS (CI) m/z 277 [MH]+. Anal.
(C13H9N2SO2F) C, H, N.
Con clu sion
As part of a program to identify a backup compound
for the p38 inhibitor 3, an algorithmically guided
SPCOS approach has been used to generate a library
of more than 570 compounds varying structurally in a
combinatorial fashion at two positions. The algorithm
was designed to produce a library which was con-
structed in a synthetically efficient manner and con-
tained a high proportion of compounds with good oral
PK parameters. From this library 10 relatively struc-
turally diverse compounds were identified which pos-
sessed an improved in vitro potency over RPR200765A
and good Cp max values and AUCs in the rat. The
identification of these compounds would not have been
possible without the use of both a solid-phase synthesis
coupled with the Monte Carlo computer algorithm
MCMS. The synthesis made use of a traceless linker
amenable to nucleophilic displacement for the introduc-
tion of diversity within the pyrimidine nucleus, and
purification of the cleaved products of the library
thereafter was achieved through the utilization of a
quench resin.
2-Dim eth oxym eth yl-4-(4-flu or op h en yl)-5-(2-m eth ylsu l-
fa n ylp yr im id in -4-yl)-1H-im id a zole (8). To a solution of 7
and glyoxal dimethyl acetal (1.2 equiv) in tert-butyl methyl
ether was added a solution of ammonium acetate (2.5 equiv)
in methanol. The resulting solution was stirred at ambient
temperature for 4 h. The reaction mixture was partitioned
between ethyl acetate and water, and the combined organic
phases were washed with brine, dried (sodium sulfate), and
concentrated in vacuo to yield an oil. Addition of diethyl ether
induced slow crystallization of a solid, 8 (86%): mp 149-151
°C; 1H NMR (DMSO-d6) 2.04 (s, 3H), 3.38 (s, 6H), 5.50 (s, 1H),
7.26 (app t, J ) 8.8 Hz, 2H), 7.60 (m, 3H), 8.50 (d, J ) 5.2 Hz,
1H), 12.95 (s, 1H); 13C (DMSO-d6) 12.6, 53.4, 98.2, 112.3, 114.6
(d, J ) 21.9 Hz), 126.9, 132.0 (d, J ) 8.3 Hz), 132.5, 145.2,
160.7, 161.9 (d, J ) 250.0 Hz), 170.4; MS (CI) m/z 361 [NH]+.
Anal. (C17H17N4SO2F) C, H, N.
This technique has enabled the rapid identification
of RPR238677 (F ) 20%), a potential drug candidate.
Meth yl Bis(2,2-h yd r oxym eth yl)p r op a n oa te (9). To a
solution of bis(2,2-hydroxymethyl)propionic acid in methanol
at 0 °C was added thionyl chloride. The resulting solution was
heated at reflux for 16 h and evaporated and the residue
distilled (97-99 °C, 0.1 mmHg):
1H NMR (DMSO-d6) 1.04 (s, 3H), 3.40 (d, J ) 10.4 Hz, 2H),
3.50 (d, J ) 10.4 Hz, 2H), 3.57 (s, 3H), 4.66 (br s, 2H); 13C
NMR (DMSO-d6) 16.8, 50.2, 51.2, 63.8, 175.2; MS (CI) m/z 149
[MH]+.
Exp er im en ta l Section
Ch em ica l Meth od s. All reagents were received and used
without further purification. All spectral data were recorded
on a Varian VXR 400 and are reported as chemical shifts in
parts per million (ppm) from an internal standard, tetram-
ethylsilane (δH 0.00) in deuteriochloroform (δH 7.27, δC 77.0)
or deuteriodimethyl sulfoxide (δH 2.49, δC 39.4), with coupling
constants (J ) given in hertz. Mass spectra were recorded on a
VG 7070/250 spectrometer, and microanalyses were performed
on a Carlo-Erba 1106 microanalyzer. Melting points are
uncorrected and were obtained using an electrothermal ap-
paratus. Chromatography was carried out using silica gel
eluting with the indicated solvent systems.
cis- an d tr a n s-2-(5-Meth yl-5-car boxym eth yl-1,3-dioxan -
2-yl)-4-(4-flu or op h en yl)-5-(2-m eth ylsu lfa n ylp yr im id in -4-
yl)-1H-im id a zole (11 a n d 12). To a solution of 8 in DMF were
added 9 (1.6 equiv), trimethylorthoformate (2.5 equiv), and
4-toluenesulfonic acid (2.5 equiv). The reaction mixture was
heated at 80 °C for 4 h before cooling. The reaction mixture
was partitioned between ethyl acetate and sodium bicarbonate
and extracted three times, and the combined organic phases
were washed with brine and dried (sodium sulfate). Following
flash chromatography on silica gel eluting with ethyl acetate
the following were obtained, in order of elution: the hydrolysis
product 10, 11, and 12.
4-Meth yl-2-th iom eth ylp yr im id in e (5). To a solution of
2-mercapto-4-methylpyrimidine hydrochloride (4) in ethanol
and 1.0 M sodium hydroxide (2 equiv) was added methyl iodide
(1 equiv). The reaction was stirred for 16 h, evaporated to half-
volume, and extracted with ethyl acetate. The combined
organic phases were washed with water and brine, dried
(sodium sulfate), and then concentrated to yield 5 (90%) as
an oil: 1H NMR (CDCl3) 2.45 (s, 3H), 2.56 (s, 3H), 6.82 (d, J )
5.1 Hz, 1H), 8.36 (d, J ) 5.1 Hz, 1H); 13C NMR (CDCl3) 14.0,
24.0, 115.9, 156.6, 167.3, 172.2; MS (CI) m/z 141 [MH]+.
2-(2-Th iom et h ylp yr im id in -4-yl)-1-(4-flu or op h en yl)e-
th a n on e (6). To a cooled solution (0 °C) of sodium bis-
(trimethylsilyl)amide (2.2 equiv) in THF was added a solution
of 5, and the resulting solution was stirred for 30 min. A
solution of methyl 4-fluorobenzoate was then added dropwise
and stirred while attaining ambient temperature over 3 h. The
reaction was quenched by addition of ammonium chloride and
extracted into ethyl acetate, and the combined organic phases
were washed with brine, dried (sodium sulfate), and concen-
trated in vacuo to yield, after trituration with hexanes, a solid,
Da ta for 2-Ca r boxya ld eh yd e-4-(4-flu or op h en yl)-5-(2-
m eth ylsu lfa n ylp yr im id in -4-yl)-1H-im id a zole (10): yield
1
16%; mp 187-188 °C; H NMR (DMSO-d6) 2.06 (s, 3H), 7.31
(m, 2H), 7.68 (m, 3H), 8.60 (m, 1H), 9.75 (s, 1H), 14.20 (s, 1H);
13C NMR (DMSO-d6) 12.6, 112.7, 114.9 (d, J ) 21.1 Hz), 126.0,
132.2 (d, J ) 8.3 Hz), 136.4, 145.0, 157.9, 160.0, 162.4 (d, J )
245.2 Hz), 171.0, 181.5; MS (CI) m/z 315 [MH]+. Anal.
(C15H11N4SOF) C, H, N.
Da ta for tr a n s Isom er 11: yield 24%; mp 199-200 °C;
1H NMR (DMSO-d6) 1.54 (s, 3H), 2.02 (s, 3H), 3.67 (s, 3H),
3.96 (d, J ) 11.1 Hz, 2H), 4.16 (d, J ) 11.1 Hz, 2H), 5.68 (s,
1H), 7.27 (m, 2H), 7.58 (m, 3H), 8.50 (m, 1H), 13.15 (s, 1H);
13C NMR (DMSO-d6) 12.6, 19.0, 51.9, 71.1, 95.5, 112.2, 114.6
(d, J ) 21.9 Hz), 127.0, 132.1 (d, J ) 8.3 Hz), 132.4, 132.8,
144.3, 157.4160.6, 162.0 (d, J ) 245.2 Hz), 170.4, 173.1; MS
(CI) m/z 445 [MH]+. Anal. (C21H21N4SO4F) C, H, N.
1
6 (99%): mp 88-90 °C; H NMR (CDCl3) (60% enol) 2.61 (s,
3H), 5.91 (s, 1H), 6.64 (d, J ) 5.3 Hz, 2H), 7.15 (m, 2H), 7.83
(m, 2H), 8.31 (d, J ) 5.3 Hz, 1H); (40% ketone) 2.52 (s, 3H),
4.35 (s, 2H), 6.98 (d, J ) 5.0 Hz, 1H), 7.15 (m, 2H), 8.07 (m,
2H), 8.46 (d, J ) 5.0 Hz, 1H); 13C NMR (CDCl3) 14.1, 47.7,
93.1, 115.6 (d, J ) 21.9 Hz), 112.7, 115.9 (d, J ) 21.9 Hz),
116.5, 128.0 (d, J ) 9.0 Hz), 131.5 (d, J ) 8.3 Hz), 156.3, 157.2,
163.1 (d, J ) 251.3 Hz), 164.4, 165.2 (d, J ) 255.8 Hz), 167.1,
169.7, 172.9, 193.7; MS (CI) m/z 263 [MH]+. Anal. (C13H11N2-
SOF) C, H, N.
1
Da ta for cis Isom er 12: yield 36%; mp 161-163 °C; H
NMR (DMSO-d6) 0.98 (s, 3H), 2.00 (s, 3H), 3.72 (s, 3H), 3.79
(d, J ) 11.1 Hz, 2H), 4.47 (d, J ) 11.1 Hz, 2H), 5.71 (s, 1H),
7.25 (m, 2H), 7.56 (m, 3H), 8.49 (m, 1H), 13.05 (s, 1H); 13C
NMR (DMSO-d6) 12.5, 17.5, 52.2, 72.1, 95.7, 112.2, 114.6 (d,
J ) 21.9 Hz), 127.0, 132.2 (d, J ) 9.0 Hz), 132.5, 132.7, 144.2,
157.3, 160.6, 162.2 (d, J ) 245.0 Hz), 170.4, 174.0; MS (CI)
m/z 445 [MH]+. Anal. (C21H21N4SO4F) C, H, N.
2-Th iom eth ylp yr im id in -4-yl-4-flu or op h en ylben zil (7).
To a solution of 6 in DMSO at 55 °C was added dropwise 48%

2182 J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 11
McKenna et al.
Ta ble 2. FT-IR Frequencies for the Resin Bond Derivatives,
before and after Oxidation with mCPBA
tr a n s-2-(5-Meth yl-5-ca r boxym eth yl-1,3-d ioxa n -2-yl)-4-
(4-flu or op h en yl)-5-(2 -m eth ylsu lfon ylp yr im id in -4-yl)-1H-
im id a zole (13): To a solution of 11 in dichloromethane and
methanol (9:1 v/v) was added mCPBA (2.2 equiv), and the
mixture was stirred for 16 h. The reaction mixture was diluted
with diethyl ether, washed with sodium bisulfite, sodium
bicarbonate, and brine, dried (sodium sulfate), and concen-
trated in vacuo to yield a foam, 13 (85%): 1H NMR (DMSO-
d6) 1.54 (s, 3H), 2.97 (s, 3H), 3.67 (s, 3H), 4.02 (d, J ) 11.1 Hz,
2H), 4.18 (d, J ) 11.1 Hz, 2H), 5.71 (s, 1H), 7.28 (m, 2H), 7.69
(m, 2H), 8.10 (br s, 1H), 8.92 (m, 1H), 13.3 (s, 1H); 13C NMR
(DMSO-d6) 19.0, 38.2, 51.9, 71.1, 95.4, 114.8 (d, J ) 21.9 Hz),
119.0, 126.2, 131.3, 132.1 (d, J ) 8.3 Hz), 145.0, 158.8, 161.4,
162.1 (d, J ) 246.0 Hz), 165.0, 173.1; MS (CI) m/z 477 [MH]+.
Anal. (C21H21N4SO6F) C, H, N.
cis-2-(5-Meth yl-5-ca r boxym eth yl-1,3-d ioxa n -2-yl)-4-(4-
flu or op h en yl)-5-(2- m et h ylsu lfon ylp yr im id in -4-yl)-1H -
im id a zole (14). In a similar fashion using the cis isomer 12,
we obtained 14 as a foam (82%): 1H NMR (DMSO-d6) 0.99 (s,
3H), 2.96 (s, 3H), 3.72 (s, 3H), 3.82 (d, J ) 11.2 Hz, 2H), 4.50
(d, J ) 11.2 Hz, 2H), 5.74 (s, 1H), 7.26 (m, 2H), 7.65 (m, 2H),
8.10 (br s, 1H), 8.92 (d, J ) 5.2 Hz, 1H), 13.2 (s, 1H); 13C NMR
(DMSO-d6) 17.5, 38.2, 42.0, 52.2, 72.1, 95.5, 114.7 (d, J ) 21.9
Hz), 119.0, 126.2, 131.5, 132.2 (d, J ) 8.3 Hz), 134.4, 144.9,
158.7, 161.4, 162.1 (d, J ) 246.0 Hz), 165.0, 174.0; MS (CI)
m/z 477 [MH]+. Anal. (C21H21N4SO6F) C, H, N.
ν(RSMe) ν(RSO2Me)
amine for Am
morpholine
(cm-1
)
(cm-1
)
A1a
A2
A3
A4
A5
A6
A7
A8
A9
A10
A11
A12
A13
A14
A15
A16
A17
1636
1666
1626
1666
1663
1674
1663
1666
1666
1669
1630
1629
1660
1662
1687
1695
1674
1653
1650
1630
1660
1635
1658
1625
1655
1645
1672
1650
1645
1646
1664
1630
1637
1652
1657
1689
1711
1665
1650
1645
1629
1657
benzylamine
piperidine
tetrahydrofurfurylamine
cyclopropylamine
aniline
3-methoxypropylamine
isopropylamine
cyclohexylamine
allylamine
dimethylamine
1-methylpiperazine
1,1-dimethylpropylidenediamine
2-methoxyethylamine
3-aminopyridine
4-aminopyridine
ammonia
A18a propylamine
A19
A20
A21
cyclopropylmethylamine
3-hydroxypyrrolidine
1,1-dimethylethylenediamine
a
A1 (morpholine) was only used as the amide component for
the cis-dioxane series and used at the expense of A18 (propy-
lamine).
P r ep a r a t ion a n d Ch a r a ct er iza t ion of R esin -Bou n d
Su bstr a tes. FT-IR analysis on a number of beads from each
reaction well was performed, ensuring both homogeneity and
identity (ν, cm-1, intensity).
A suspension of the resin 20 in dichloromethane was treated
with a catalytic amount of DMF followed by oxalyl chloride.
The addition of oxalyl chloride was repeated twice more before
the resin was washed with dichloromethane. The resin was
then treated with an appropriate amine Am (A1-A21), the
reaction mixture was agitated for 30 min before being filtered,
washed with THF, methanol, and finally dichloromethane, and
then dried under vacuum at 60 °C. Alternatively, this proce-
dure could be carried out using 20, HATU, and triethylamine
in DMF with agitation for 24 h: the derivatized resins were
subsequently characterized by FT-IR. The carbonyl (CdO)
stretching frequencies are given in Table 2.
To a suspension of the resin 21 (A1-A21) in dichlo-
romethane and methanol (10:1) was added mCPBA (3 equiv).
After 48 h of agitation the resin was filtered off, washed with
dichloromethane, and dried under vacuum at 60 °C: carbonyl
(CdO) stretching frequencies as characterization for 22 are
given in the table. Alternatively, to a suspension of the resin
21 in methanol, THF, and water (1:1:0.1) was added OXONE.
After 24 h of agitation the resin was filtered off, washed with
methanol, THF, and dichloromethane, and dried under vacuum
at 60 °C, yielding 22.
A suspension of Merrifield resin (20 g, 1.1 mmol g-1) in DMF
was treated with potassium thioacetate (5 equiv). After 24 h
at ambient temperature the resulting resin 15 was filtered,
then washed with DMF, THF, water, THF, and finally dichlo-
romethane, and then dried under vacuum at 60 °C: IR 3024s,
2920s, 1691s, 1599m, 1493s, 1453s, 1130m, 758s, 700s.
A suspension of the resin 15 in THF was treated with
lithium borohydride. After 24 h the resin was washed with
THF, 1 N HCl, THF (7:3), water, THF, methanol, and finally
dichloromethane and then dried under vacuum at 60 °C,
yielding 16: IR 3026s, 2924s, 1599w, 1493s, 1452s, 757s, 700s.
A suspension of the thiol resin 16 in DMF was treated with
sodium hydride (60% dispersion in mineral oil) and agitated
for 15 min. It was then treated with a DMF solution of 13 or
14 (2 equiv) and subsequently heated for 20 h. Filtration,
washing with THF, methanol, and finally dichloromethane,
and then drying under vacuum at 60 °C yielded 17 or 18,
respectively: IR 3021m, 2920m, 1732m, 1569m, 1452s, 1091m,
835m, 757s, 699s.
Resin Loa d in g Deter m in a tion . To a suspension of the
resin 17 (100 mg) in dichloromethane and methanol (10:1) was
added mCPBA (3 equiv). After 48 h of agitation the resin was
filtered off, washed with dichloromethane, and dried under
vacuum at 60 °C. The resin was split into two equal portions
and treated with (1) sodium thiomethoxide (3 mmol) in THF
at ambient temperature and (2) benzylamine (3 mmol) in DME
at reflux. Isolation of the products by preparative TLC eluting
with ethyl acetate yielded, respectively, 11, 16.5 mg, indicating
0.75 mmol of substrate/g of resin and trans-2-(5-m eth yl-5-
ca r boxym eth yl-1,3-d ioxa n -2-yl)-4-(4-flu or op h en yl)-5-(2-
N-ben zyla m in op yr im id in -4-yl)-1H-im id a zole (19) as a
foam (15.2 mg, 80% based on previous loading determina-
tion): 13C NMR (DMSO-d6) 19.1, 40.3, 43.5, 51.9, 71.2, 95.7,
114.3 (d, J ) 21.9 Hz), 119.0, 126.2, 131.5, 132.2 (d, J ) 8.3
Hz), 134.4, 140.6, 143.8, 158.0, 161.4, 162.1 (d, J ) 246.0 Hz),
165.0, 173.2; MS (CI) m/z 504 [MH]+. Anal. (C27H26N5O4F
0.4H2O) C, H, N.
Thus, the amines N1-N22 were utilized in the nucleophilic
displacements from the activated 2-sulfonylpyrimidine resins
22. The reaction block (96-well; ACT) was set up with resin
(30 mg) in NMP (1 mL) per well; thereafter was added the
appropriate amine (10 equiv), and the block was heated to 80
°C for 16 h. Each well was flushed with NMP, and the solvents
were then removed in vacuo. Addition of dichloromethane and
isocyanate resin (3 equiv with respect to amine) and further
agitation for 16 h followed by filtration and evaporation of the
solvent in vacuo allowed the isolation of the library products.
The yield of the resin-cleaved products ranged from 1% to 75%,
with between 0.1 and 8.4 mg being isolated with analysis of
the library being carried out by parallel LC-MS and ELS.
Data for the resynthesized compounds are as follows.
Da ta for cis-2-{4-(4-F lu or op h en yl)-5-[2-(2-m eth oxyeth -
yla m in o)p yr im id in -4-yl]-1H-im id a zol-2-yl}-5-m eth yl-[1,3]-
d ioxa n e-5-ca r b oxylic Acid Cyclop r op ylm et h yla m id e
(2{16, 19}, RP R235455): mp 168-170 °C; MS (HRMS) m/z
A suspension of the resin 17 in THF was treated with 1 N
NaOH and stirred for 8 h at 70 °C. The resin was filtered off,
washed with 1 N HCl/THF (1:1), then THF, DMF, methanol,
and finally dichloromethane, and then dried under vacuum
at 60 °C, yielding 20: IR 3060w, 3025m, 2923m, 2852w,
1720w, 1601m, 1571s, 1493s, 1452s, 1335s, 1235m, 1199m,
1183m, 1097s, 838m, 760s, 704s.
1
calcd for C26H32N6O4F (M + H) 511.2469, found 511.2447; H
NMR (DMSO-d6) 0.23 (m, 2H), 0.40 (m, 2H), 0.95-1.10 (m,
1H), 0.98 (s, 3H), 3,17 (t, J ) 6 Hz, 2H), 3.30 (s, 3H), 3.37 (m,
2H), 3.42 (m, 2H), 3,81 (d, J ) 12 Hz, 2H), 4.42 (d, J ) 12 Hz,
2H), 5.75 (s, 1H), 6.05 (m, 1H), 6.91 (br d, J ) 5 Hz, 1H), 7.17
(t, J ) 9 Hz, 2H), 7.31 (m, 1H), 7.68 (dd, J ) 9 and 7 Hz, 2H),

Identification of Potent p38 MAP Kinase Inhibitors
J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 11 2183
8.19 (d, J ) 5 Hz, 1H), 11.20-12.40 (m, 1H). Anal. (C26H31
-
C26H30N6O3F (M + H) 493.2363, found 493.2401; 1H NMR
(DMSO-d6) 0.12 (m, 2H), 0.37 (m, 2H), 0.50 (m, 2H), 0.66 (m,
2H), 0.93 (m, 1H), 1.50 (s, 3H), 2.69 (m, 1H), 3.00 (m, 2H),
4.00 (d, J ) 12 Hz, 2H), 4.08 (d, J ) 12 Hz, 2H), 5.61 (s, 1H),
6.31 (m, 1H), 6.94 (m, 1H), 7.19 (t, J ) 9 Hz, 2H), 7.31 (br s,
1H), 7.67 (dd, J ) 9 and 6.5 Hz, 2H), 8.20 (d, J ) 5 Hz, 1H),
12.00-12.60 (br m, 1H). Anal. (C26H29FN6O3‚H2O) C, H, N.
FN6O) C, H, N.
Da t a for cis-2-[5-(2-Allyla m in op yr im id in -4-yl)-4-(4-
flu or op h en yl)-1H-im id a zol-2-yl]-5-m eth yl[1,3]d ioxa n e-5-
ca r b oxylic Acid (2-Met h oxyet h yl)a m id e (2{8, 14},
RP R235315): mp 188-190 °C; MS (HRMS) m/z calcd for
C
25H30N6O4F (M + H) 497.2313, found 497.2345; 1H NMR
(DMSO-d6) 1.00 (s, 3H), 3.24 (s, 3H), 3,42 (m, 2H), 3.49 (t, J )
6 Hz, 2H), 3.70-4.00 (m, 2H), 3.82 (d, J ) 12 Hz, 2H), 4.40 (d,
J ) 12 Hz, 2H), 5.02 (br d, J ) 10 Hz, 1H), 5.13 (br d, J ) 16
Hz, 1H), 5.75 (s, 1H), 5.75-6.00 (m, 1H), 6.00-6.50 (m, 2H),
7.17 (t, J ) 9 Hz, 2H), 7.33 (m, 1H), 7.68 (dd, J ) 9 and 6.5
Hz, 2H), 8.21 (d, J ) 5 Hz, 1H), 11.50-12.50 (m, 1H). Anal.
(C25H29FN6O4) C, H, N.
Da ta for tr a n s-2-{4-(4-F lu or op h en yl)-5-[2-(3-m eth oxy-
p r op yla m in o)p yr im id in -4-yl]-1H-im id a zol-2-yl}-5-m eth -
yl-[1,3]d ioxa n e-5-ca r boxylic Acid P r op yla m id e (1{11,
18}, RP R239939): mp 176-177 °C; MS (HRMS) m/z calcd
for C26H33N6O4F (M + H) 513.2626, found 513.2623; 1H NMR
(DMSO-d6) 0.90 (t, J ) 7.5 Hz, 3H), 1.45-1.60 (m, 2H), 1.53
(s, 3H), 1.69 (m, 2H), 3.11 (m, 2H), 3.19 (broad m, 2H), 3.27
(s, 3H), 3.36 (t, J ) 6.5 Hz, 2H), 4.03 (d, J ) 12 Hz, 2H), 4.11
(d, J ) 12 Hz, 2H), 5.63 (s, 1H), 6.30 (m, 1H), 6.94 (m, 1H),
7.19 (t, J ) 9 Hz, 2H), 7.30 (m, 1H), 7.71 (dd, J ) 9 and 6.5
Hz, 2H), 8.20 (d, J ) 5 Hz, 1H), 12.00-12.50 (br m, 1H). Anal.
(C26H33FN6O4) C, H, N.
Da ta for cis-2-{4-(4-F lu or op h en yl)-5-[2-(3-m eth oxyp r o-
p yla m in o)p yr im id in -4-yl]-1H -im id a zol-2-yl}-5-m et h yl-
[1,3]d ioxa n e-5-ca r boxylic Acid (2-Meth oxyeth yl)a m id e
(2{11, 14}, RP R235316): mp 159-161 °C; MS (HRMS) m/z
1
calcd for C26H34N6O5F (M + H) 529.2575, found 529.2598; H
NMR (DMSO-d6) 0.98 (s, 3H), 1.69 (m, 2H), 3.19 (m, 2H), 3.22
(s, 3H), 3.25 (s, 3H), 3,36 (m, 2H), 3.40 (t, J ) 5.5 Hz, 2H),
3.47 (t, J ) 6 Hz, 2H), 3.80 (d, J ) 12 Hz, 2H), 4.40 (d, J ) 12
Hz, 2H), 5.73 (s, 1H), 6.32 (m, 1H), 6.91 (m, 1H), 7.19 (t, J )
9 Hz, 2H), 7.44 (m, 1H), 7.69 (dd, J ) 9 and 6.5 Hz, 2H), 8.19
(d, J ) 5 Hz, 1H), 11.50-12.50 (br m, 1H). Anal. (C26H33FN6O5)
C, H, N.
Da ta for tr a n s-2-[5-[2-(Cyclop r op ylm eth yla m in o)p y-
r im idin -4-yl]-4-(4-flu or oph en yl)-1H-im idazol-2-yl]-5-m eth -
yl-[1,3]d ioxa n e-5-ca r boxylic Acid Dim eth yla m id e (1{9,
11}, RP R237897): mp 196-198 °C; MS (HRMS) m/z calcd
for C25H30N6O3F (M + H) 481.2363, found 481.2347; 1H NMR
(DMSO-d6) 0.17 (m, 2H), 0.40 (m, 2H), 0.97 (m, 1H), 1.65 (s,
3H), 3.00 (s, 6H), 3.06 (m, 2H), 4.13 (d, J ) 12 Hz, 2H), 4.20
(d, J ) 12 Hz, 2H), 5.61 (s, 1H), 6.16 (m, 1H), 6.93 (m, 1H),
7.17 (t, J ) 9 Hz, 2H), 7.68 (dd, J ) 9 and 6.5 Hz, 2H), 8.20 (d,
J ) 5 Hz, 1H), 11.50-12.50 (br m, 1H). Anal. (C27H29FN6O4‚
3H2O) C, H, N.
Da ta for tr a n s-2-{4-(4-F lu or op h en yl)-5-[2-(3-cyclop r o-
p ylm et h yl)p yr im id in -4-yl]-1H -im id a zol-2-yl}-5-m et h yl-
[1,3]d ioxa n e-5-ca r boxylic Acid 4-Meth ylp ip er a zin a m id e
1
(1{9, 12}, RP R238677): mp 229-230 °C; H NMR (DMSO-
d6) 0.15 (m, 2H), 0.41 (m, 2H), 0.98 (m, 1H), 1.63 (s, 3H), 2.25
(s, 3H), 2.38 (t, J ) 5 Hz, 4H), 3.06 (m, 2H), 3.55 (t, J ) 5 Hz,
4H), 4.15 (m, 4H), 5.63 (s, 1H), 6.13 (br m, 1H), 6.93 (br m,
1H), 7.17 (t, J ) 9 Hz, 2H), 7.69 (dd, J ) 9 and 5.5 Hz, 2H),
8.19 (d, J ) 5.5 Hz, 1H), 11.50-12.50 (br m, 1H); MS (CI) m/z
536 [MH]+. Anal. (C28H34FN7O3) C, H, N.
Biologica l Meth od s. Kin a se Assa y. The p38 enzyme
assay is carried out at room temperature for 1 h, using 40 ng/
well of the mouse enzyme. The substrate 50 µg/mL ATF-2 is
coated onto 96-well plates, the assay is carried out in 25 mM
Hepes buffer, pH 7.7, containing 25 mM magnesium chloride,
2 mM dithiothreitol, 1 mM sodium orthovanadate, and 100
µM ATP. Phosphorylated ATF-2 is quantitated using a phos-
pho-specific ATF-2 primary antibody (rabbit anti-human)
followed by a europium-labeled secondary antibody (sheep
anti-rabbit IgG) with addition of the DELFIA enhancement
solution, resulting in fluorescence. ERK was measured using
a [33P]ATP filtration assay format used for the substrate
myelin basic protein. ZAP-70, Syk, and Lck kinase activities
were measured using the homogeneous time-resolved fluores-
cence (HTRF) methodology with the catalytic domains of each
of the tyrosine kinases, biotinylated, specific peptide sub-
strates, streptavidin-linked APC, and europium cryptate-
conjugated anti-phosphotyrosine antibody. Results represent
means ( SEM (n ) 3).
Hu m a n Mon ocytic Cell Lin e (THP -1) TNF r Relea se
Assa y. THP-1 cells were seeded into 96-well plates at 1 × 106-
cells/mL (200 µL/well) in medium containing FCS (1%) and
incubated overnight. Following pretreatment with compounds
for 1 h, the cells were incubated with LPS (20 µg/mL) for a
further 24 h. TNFR release was measured in the supernatants
by sandwich ELISA. IC50 values shown from repeat experi-
ments are means (n ) 3).
Mou se TNF r Relea se Assa y. Compound was adminis-
tered orally to balb/c mice 30 min prior to LPS (0.1 mg/kg ip)
challenge. Serum TNFR levels were determined 90 min after
LPS insult. Results represent means (n ) 3).
Ra t Dr u g Distr ibu tion Stu d ies. Three rats were dosed
orally (1 mg/kg) with a suspension of compound in methylcel-
lulose or iv (1 mg/kg) with a solution of compound in N-
methylpyrrolidone. Blood samples were taken at intervals over
24 h and analyzed for compound. Averaged results are reported
(n ) 3).
Ra t SCW-In d u ced Ar th r itis Mod el. This model was
performed using Lewis rats in either a prophylactic or thera-
peutic regimen with 10 animals per group. On day 0 an intra-
articular injection of 10 µg of SCW 100P fraction (in 10 µL)
was made into the tibiotarsal joint of one hind limb of
anaesthetized female rat. Any animals which had not devel-
oped a swollen ankle 2 days later were excluded from the
study. Challenge with iv SCW was made on day 21, and
Da ta for tr a n s-2-(5-(4-F lu or op h en yl)-4-{2-[(fu r a n -2-yl-
m eth yl)a m in o]p yr im id in -4-yl}-1H-im id a zol-2-yl)-5-m eth -
yl-[1,3]d ioxa n e-5-ca r boxylic Acid P r op yla m id e (1{14,
18}, RP R239938): mp 158-160 °C; MS (HRMS) m/z calcd
for C27H30N6O4F (M + H) 521.2313, found 521.2310; 1H NMR
(DMSO-d6) 0.91 (t, J ) 7.5 Hz, 3H), 1.45-1.60 (m, 2H), 1.52
(s, 3H), 3.13 (m, 2H), 4.04 (d, J ) 12 Hz, 2H), 4.12 (d, J ) 12
Hz, 2H), 4.39 (m, 2H), 5.64 (s, 1H), 6.15 (m, 1H), 6.30-6.70
(br m, 1H), 6.34 (br s, 1H), 6.80-7.30 (br m, 1H), 7.13 (t, J )
9 Hz, 2H), 7.46 (br s, 1H), 7.69 (dd, J ) 9 and 6.5 Hz, 2H),
8.23 (d, J ) 5 Hz, 1H). Anal. (C27H29FN6O4‚2H2O) C, H, N.
Da ta for tr a n s-2-[5-[2-(Cyclop r op ylm eth yla m in o)p y-
r im idin -4-yl]-4-(4-flu or oph en yl)-1H-im idazol-2-yl]-5-m eth -
yl-[1,3]d ioxa n e-5-ca r boxylic Acid Isop r op yla m id e (1{9,
8}, RP R239497): mp 138-140 °C; MS (HRMS) m/z calcd for
C
26H32N6O3F (M + H) 495.2520, found 495.2531; 1H NMR
(DMSO-d6) 0.15 (m, 2H), 0.40 (m, 2H), 0.97 (m, 1H), 1.16 (d, J
) 7 Hz, 6H), 1.51 (s, 3H), 3.05 (m, 2H), 3.96 (m, 1H), 4.04 (d,
J ) 12 Hz, 2H), 4.09 (d, J ) 12 Hz, 2H), 5.63 (s, 1H), 6.14 (m,
1H), 6.87 (m, 1H), 6.92 (m, 1H), 7.17 (t, J ) 9 Hz, 2H), 7.69
(dd, J ) 9 and 6.5 Hz, 2H), 8.19 (d, J ) 5 Hz, 1H), 11.00-
12.50 (br m, 1H). Anal. (C26H31FN6O3‚2.5H2O) C, H, N.
Da ta for tr a n s-2-(4-(4-F lu or op h en yl)-5-{2-[(p yr id in -2-
ylm eth yl)am in o]pyr im idin -4-yl}-1H-im idazol-2-yl)-5-m eth -
yl-[1,3]d ioxa n e-5-ca r boxylic Acid Dim eth yla m id e (1{21,
11}, RP R238778): mp 120-122 °C; MS (HRMS) m/z calcd
for C27H29N7O3F (M + H), 518.2316, found 518.2291; 1H NMR
(DMSO-d6) 1.63 (s, 3H), 3.00 (s, 6 H), 4.13 (d, J ) 12 Hz, 2H),
4.20 (d, J ) 12 Hz, 2H), 4.51 (m, 2H), 5.62 (s, 1H), 6.71 (m,
1H), 6.97 (m, 1H), 7.10 (t, J ) 9 Hz, 2H), 7.15-7.30 (m, 2H),
7.60-7.75 (m, 3H), 8.24 (d, J ) 5 Hz, 1H), 8.54 (br d, J ) 4.5
Hz, 1H), 11.70-12.40 (br m, 1H). Anal. (C27H28FN7O3‚0.5H2O)
C, H, N.
Da ta for tr a n s-2-[5-[2-(Cyclop r op ylm eth yla m in o)p y-
r im idin -4-yl]-4-(4-flu or oph en yl)-1H-im idazol-2-yl]-5-m eth -
yl-[1,3]dioxan e-5-car boxylic Acid Cyclopr opylam ide (1{9,
5}, RP R239457): mp 186-189 °C; MS (HRMS) m/z calcd for

2184 J ournal of Medicinal Chemistry, 2002, Vol. 45, No. 11
McKenna et al.
(12) Lipinski, C. A.; Lombardo, F.; Dominy, B. W.; Feeney, P. J .
Experimental and computational approaches to estimate solubil-
ity and permeability in drug discovery and development settings.
Adv. Drug Delivery Rev. 1997, 23, 3-25.
(13) Pickett, S. D.; McLay, I. M.; Clark, D. E. Enhancing the Hit-to
Lead properties of Lead Optimization Libraries. J . Chem. Inf.
Comput. Sci. 2000, 40, 263-272.
(14) Bravi, G.; Green, D. V. S.; Hann, M. M.; Leach, A. R. PLUMS:
a program for the Rapid Optimization of Focussed libraries J .
Chem. Inf. Comput. Sci. 2000, 40, 1441-1448.
(15) Pozzan, A.; Leach, A. R.; Feriani, A.; Hann, M. Virtual
Optimization of Chemical Libraries Using Genetic Algorithms.
Book of Abstracts, 218th ACS National Meeting, New Orleans,
Aug 22-26, 1999; American Chemical Society: Washington, DC,
1999.
compounds were given on days 20-24 (po, b.i.d.) for the
prophylactic studies and 21-24 for the therapeutic studies.
Paw swelling was calculated by summation of daily measure-
ment of the paw medial and lateral malleoli width postchal-
lenge with iv SCW until termination of the study. Radiography
of tibio-tarsal joints was undertaken using a Phillips Practix
X-ray unit. The radiographs were scored using a system which
sums lesion scores for each of the bonesswhich are divided
into quadrants.
The results represent means ( SEM. One asterisk indicates
p < 0.05, and two asterisks indicate p < 0.01, ANOVA with
post hoc Dunnett’s test compared to vehicle-treated animals.
Ack n ow led gm en t. We thank Adnan Al-Shaar and
Mustafa Bhimani for the synthesis of the compounds,
Nicola Wilsher and Clive Brealey for DMPK studies,
Nick Hopkins, Peter Lockey, Martyn Foster, and Maria
Belvisi for the biology results, Mireille Robin, Serge
Sable´, Gilles Barbalat, and Bertrand Monegier for
spectroscopy, and Fabrice Debu and Zofia Kochanek for
microanalysis.
(16) Brown, R. D.; Martin, Y. C. Designing Combinatorial Library
Mixtures Using a Genetic Algorithm. J . Med. Chem. 1997, 40,
2304-2313.
(17) Zheng, W.; Cho, S. J .; Tropsha, A. Rational Combinatorial
Library Design. 1. Focus-2D: A New Approach to the Design of
Targeted Combinatorial Chemical Libraries. J . Chem. Inf.
Comput. Sci. 1998, 38, 251-258.
(18) Cook, D. C.; Lythgoe, D. J .; Thompson, D. M.; Wallace, P. A.;
Walton, J . B. Chemical Development and Scale-up of
RPR200765A, a Novel p38 Anti-inflammatory Candidate. In
press.
(19) Kobayashi, S.; Hachiya, I.; Suzuki, S.; Moriwaki, M. Polymer-
supported silyl enol ethers. Synthesis and reactions with imines
for the preparation of an amino alcohol library. Tetrahedron Lett.
1996, 37, 2809-12.
(20) (a) Gayo, L. M.; Suto, M. J . Traceless linker: oxidative activation
and cleavage of a sulphur-based linker. Tetrahedron Lett. 1997,
38, 211-214. (b) Obrecht, D.; Abrecht, C.; Grieder, A.; Villal-
gordo, J . M. A novel and efficient approach for the combinatorial
synthesis of structurally diverse pyrimidines on solid support.
Helv. Chim. Acta 1997, 80, 65-72. (c) Masquelin, T.; Sprenger,
D.; Baer, R.; Gerber, F.; Merchadal, Y. A novel solution and solid-
phase approach to 2,4,5-tri- and 2,4,5,6-tetrasubstituted pyri-
midines and their conversion to condensed heterocycles. Helv.
Chim. Acta 1998, 81, 646-660.
(21) Brown, D. J .; Ford, P. W. Simple Pyrimidines. X. The formation
and reactivity of 2,4- and 5-pyrimidinyl sulfones and sulfoxides.
J . Chem. Soc. C 1967, 568-72.
(22) Gacek, M.; Undheim, K. Chemoselectivity and regioselectivity
in reactions of pyrimidines. Acta Chem. Scand., Ser. B 1985,
B39, 691-6.
(23) Fitch, W. L.; Detre, G.; Holmes, C. P.; Shoolery, J . N.; Keifer, P.
A. High-resolution 1H NMR in Solid-Phase Organic Synthesis
J . Org. Chem. 1994, 59, 7955-6.
(24) (a) Deady, L. W. Ring nitrogen oxidation of amino substituted
nitrogen heterocycles with m-chloroperbenzoic acid. Synth.
Commun. 1977, 7, 509-514 (b) Brown, D. J .; Ford, P. W.
Pyrimidine Reactions. XIX. Aminolysis and hydrolysis of sul-
foxides and sulfones derived from 2-(p-substituted phenylthio)-
pyrimidines. J . Chem. Soc. C 1969, 2720-3.
Refer en ces
(1) Lee, J . C.; Badger, A. M.; Griswold, D. E.; Dunnington, D.;
Truneh, A.; Votta, B.; White, J . R.; Young, P. R.; Bender, P. E.
Bicyclic imidazoles as a novel class of cytokine biosynthesis
inhibitors Ann. N. Y. Acad. Sci. 1993, 696, 149-170.
(2) Muller-Ladner, U. Molecular and cellular interactions in rheu-
matoid synovium. Curr Opin. Rheumatol. 1996, 8, 210-220.
(3) (a) Lee, J . C.; Laydon, J . T.; McDonnell, P. C.; Gallagher, T. F.;
Kumar, S.; Green, D.; McNulty, D.; Blumenthal, M. J .; Heys, J .
R.; Landvatter, S. W.; Strickler, J . E.; McLaughlin, M. M.;
Siemens, I. R.; Fisher, S. M.; Livi, G. P.; White, J . R.; Adams, J .
L.; Young, P. R. A protein kinase involved in the regulation of
inflammatory cytokine biosynthesis. Nature (London) 1994, 372,
736. (b) Han, J .; Lee J .-D.; Bibbs, L. Ulevitch, R. J . A MAP
Kinase Targeted by Endotoxin and Hyperosmolarity in Mam-
malian Cells. Science 1994, 265, 808.
(4) Raingeaud, J .; Gupta, S.; Rogers, J . S.; Dickens, M.; Han, J .;
Ulevitch, R. J .; Davis, R. J . Pro-inflammatory cytokines and
environmental stress cause p38 mitogen-activated protein kinase
activation by dual phosphorylation on tyrosine and threonine.
J . Biol. Chem. 1995, 270, 7420.
(5) Boehm, J . C.; Adams, J . L. New inhibitors of p38 kinase. Exp.
Opin. Ther. Patents 2000, 10, 25 and references therein.
(6) Astles, P. C.; Ashton, M. J .; Bridge, A. W.; Harris, N. V.; Hart,
T. W.; Parrott, D. P.; Porter, B.; Riddell, D.; Smith, C.; Williams,
R. J . Acyl-CoA: Cholesterol O-acyltransferase (ACAT) inhibitors.
2.2-(1,3-Dioxan-2-yl)-4,5-diphenyl-1H-imidazoles as potent in-
hibitors of ACAT. J . Med. Chem. 1996, 39, 1423.
(7) McLay, I. M.; Halley, F.; Souness, J . E.; McKenna, J . M.;
Benning, V.; Birrell, M.; Burton, B.; Belvisi, M.; Collis, A.;
Constan, A.; Foster, M.; Hele, D.; J ayyosi, Z.; Kelley, M.; Maslen,
C.; Miller, G.; Ouldelheim, M.-C.; Page, K.; Phipps, S.; Pollock,
K.; Porter, B.; Ratcliffe, A. J .; Redford, E. J .; Webber, S.; Slater,
B.; Thybaud, V.; Wilsher, N. The discovery of RPR200765A, a
p38 MAP Kinase inhibitor displaying a good oral anti-arthritic
efficacy. Bioorg. Med. Chem. 2001, 9, 537-554.
(8) Yan, B.; Kumaravel, G.; Anjaria, H.; Wu, A.; Petter, R. C.; J ewell,
C. F., J r.; Wareing, J . R. Infrared Spectrum of a single resin
bead for real-time monitoring of solid-phase reactions. J . Org.
Chem. 1995, 60, 5736-8.
(9) Booth, R. J .; Hodges, J . C. Poymer-Supported quenching Re-
agents for Parallel Purification. J . Am. Chem. Soc. 1997, 119,
4882-6.
(25) Liverton, N. J .; Butcher, J . W.; Claiborne, C. F.; Claremon, D.
A.; Libby, B. E.; Nguyen, K. T.; Pitzenberger, S. M.; Selnick, H.
G.; Smith, G. R.; Tebben, A.; Vacca, J . P.; Varga, S. L.; Agarwal,
L.; Dancheck, K.; Forsyth, A. J .; Fletcher, D. S.; Frantz, B.;
Hanlon, W. A.; Harper, C. F.; Hofsess, S. J .; Kostura, M.; Lin,
J .; Luell, S.; O’Neill, E. A.; Orevillo, C. J .; Pang, M.; Parsons, J .;
Rolando, A.; Sahly, Y.; Visco, D. M.; O’Keefe, S. J . Design and
Synthesis of Potent, and Orally Bioavailable Tetrasubstituted
Imidazole Inhibitors of p38 Mitogen-Activated Kinase. J . Med.
Chem. 1999, 42, 2180-2190.
(26) The difference between this and the targeted 800 compounds
was
a
venting phenomenon within the automated reaction
high local vapor pressure forced the
chamber, in which
a
contents of a well to waste. Hence, upon completion of the
reaction the well contained dry resin and no solvent, reagent,
or product.
(10) Collis, A.; Foster, M.; Halley, F.; Maslen, C.; McLay, I. M.; Page,
K.; Wilsher, N.; Redford, E. J .; Souness, J . E. RPR203494 a
Pyrimidine Analogue of the p38 Inhibitor RPR200765A with an
improved in vitro Potency. Bioorg. Med. Chem. Lett. 2001, 11,
693-696.
(11) Palm, K.; Stenberg, P.; Luthman, K.; Artursson, P. Polar
molecular surface properties predict the intestinal absorption
of drugs in humans. Pharm. Res. 1997, 14, 568-71.
(27) This classification into high (>20%), medium (2-20%), and low
(<2%) is used by the group in which this assay was carried out
for all programs and compounds and thus is a fair judgment of
the library with regard to the assay.
J M011132L