N-4-Pyrimidinyl-1H-indazol-4-amine inhibitors of Lck: Indazoles as phenol isosteres with improved pharmacokinetics

Article abstract of DOI:10.1016/j.bmcl.2007.04.029

2,4-Dianilino pyrimidines are well-known inhibitors of tyrosine kinases including lymphocyte specific kinase (Lck). Structure-activity relationships at the 4-position are discussed and rationalised. Examples bearing a 2-methyl-5-hydroxyaniline substituent

Full text of DOI:10.1016/j.bmcl.2007.04.029

Bioorganic & Medicinal Chemistry Letters 17 (2007) 4363–4368
N-4-Pyrimidinyl-1H-indazol-4-amine inhibitors of Lck:
Indazoles as phenol isosteres with improved pharmacokinetics
Paul Bamborough,* Richard M. Angell, Inder Bhamra, David Brown, James Bull,à
John A. Christopher, Anthony W. J. Cooper, Lynsey H. Fazal,§ Ilaria Giordano,
Lucy Hind, Vipulkumar K. Patel, Lisa E. Ranshaw, Martin J. Sims, Philip A. Skone,
Kathryn J. Smith, Emma Vickerstaff§ and Melanie Washington–
GlaxoSmithKline R&D, Medicines Research Centre, Gunnels Wood Road, Stevenage, Hertfordshire SG1 2NY, UK
Received 14 March 2007; revised 5 April 2007; accepted 8 April 2007
Available online 13 April 2007
Abstract—2,4-Dianilino pyrimidines are well-known inhibitors of tyrosine kinases including lymphocyte specific kinase (Lck). Struc-
ture–activity relationships at the 4-position are discussed and rationalised. Examples bearing a 2-methyl-5-hydroxyaniline substitu-
ent at the 4-position were especially potent but showed poor oral pharmacokinetics. Replacement of this substituent by 4-amino(5-
methyl-1H-indazole) yielded compounds with comparable enzyme potency and improved pharmacokinetic properties.
ꢀ 2007 Elsevier Ltd. All rights reserved.
Lck, a member of the Src family of tyrosine kinases, is
important for the development and activation of T-cells.
Mice which are Lck-deficient or which express a kinase-
inactive mutant Lck show T-cell deficiency.1,2 Antigen
response thresholds are elevated in LckÀ/À mice and re-
turn to normal when Lck is inducibly expressed.3 T-cell
receptor f-subunit phosphorylation and signal transduc-
tion is also regulated by Lck.4 A consequence of Lck
activation is the induction of cytokines including IL-2
which is capable of supporting T-cell proliferation.5
Inhibition of Lck kinase activity therefore offers one ap-
proach to the treatment of T-cell mediated inflammatory
disorders including rheumatoid arthritis, transplant
rejection and inflammatory bowel disease.6
most potent against Lck were compounds with a 2-
methyl 5-hydroxy aniline as a 4-substituent, for example
1 (Lck IC50 = 59 nM).9
The 2,4-dianilino pyrimidines are another intensively
studied series of potent protein kinase inhibitors.10,11
4-Anilino quinolines and the 2,4-dianilino pyrimidines
are likely to bind to Lck so that their 4-substituents
overlay closely. With this expectation, the 2-methyl 5-
hydroxy aniline substituent was taken from 1 and intro-
duced into the 4-position of the pyrimidines (to give e.g.
2). The routes shown in Scheme 1 were used to synthe-
sise the molecules from readily available 2,4-dichloro
pyrimidine or 2-methyl thio-4-hydroxypyrimidine.12
4-Anilino quinolines and quinazolines are potent inhib-
itors of tyrosine kinases.7,8 Of the quinolines, among the
S
O
H
N
N
N
N
H
N
Keywords: 2,4-Dianilino pyrimidines; Lck inhibitors; p56Lck inhibi-
tors; Protein kinase inhibitors.
HN
OH
HN
OH
H2N
O
*
Corresponding author. Tel.: +44 (0) 1438 763246; e-mail:
Paul.A.Bamborough@gsk.com
1
2
Present address: Arrow Therapeutics, Britannia House, 7 Trinity
Street, London SE1 1DA, UK.
´ ´
Present address: Department of Chemistry, Universite de Montreal,
à
§
´
PO Box 6128, Montreal, Que., Canada.
In order to focus on a thorough exploration of SAR at
the 4-position, the majority of the early compounds
were made with 3-aminobenzamide as the pyrimidinyl
2-substituent. Greater variation at the 2-position was
Present address: Astex Therapeutics, 436 Cambridge Science Park,
Milton Road, Cambridge CB4 0QA, UK.
–
Present address: Royal Society of Chemistry, Burlington House,
Piccadilly, London W1J 0BA, UK.
0960-894X/$ - see front matter ꢀ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.04.029

4364
P. Bamborough et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4363–4368
NH, respectively). This is in agreement with published
crystal structures of similar compounds bound to
CDK2.14 In these structures the phenyl ring of the 4-ani-
line substituent is directed towards the front of the ATP-
binding site, but in the model used here it is angled to-
wards the back of the site in a predominantly hydropho-
bic location. Its 2-methyl substituent binds in a small
lipophilic hole formed by sidechains of the Lck N-termi-
nal lobe (Val259, Ala271, and Lys273). The 5-hydroxyl
group points towards the backbone of Asp382 and the
sidechain of Glu288. These interactions, illustrated in
Figure 1, governed the choice of aniline reagents for
the pyrimidinyl 4-position.
Cl
Cl
N
MeS
N
N
N
Cl
OH
(a)
(c)
N
H
N
N
N
R2
N
R3
R4
HN
R6
CO2H
OH
R5
In addition to substituents chosen to present differently
shaped hydrophobic groups to the kinase back pocket,
anilines were chosen to provide SAR information about
the nature of the interactions made by the 2-methyl and
the 5-hydroxyl. For the resulting compounds, IC50 val-
ues were measured for inhibition of Lck activity. Aver-
aged values are shown in Table 1.9 3-Aminobenzamide
was used as the pyrimidinyl 2-substituent because it is
directed towards solvent (Fig. 1) and was expected to
be indicative of the activity of 3-aminobenzamides with
larger amide substituents.
(d), (b)
(b)
H
N
N
N
R2
R1
R3
R4
HN
R6
R5
Scheme 1. Synthesis of 2,4-dianilino pyrimidines 2–25. Reagents and
conditions: (a) substituted aniline, NaHCO3, t-BuOH, reflux; (b)
substituted aniline, acetone/water/cHCl (150:100:2), 80 ꢁC; (c) 3-
aminobenzoic acid, diglyme, 180 ꢁC; (d) POCl3, reflux; NH3/dioxane,
rt.
From these data, it can be seen that the 2-methyl con-
tributes about 10-fold to the activity (compare 3 to 2).
Moving the methyl around the ring to the 3- and 4-posi-
tions (4 and 5) did not restore full activity. This was pre-
dicted by the model: the 2-methyl group not only fills the
small lipophilic hole described earlier, but would also af-
fect the orientation of the aromatic ring into the back
pocket. In the docking, the NH linker at the 4-position
of the pyrimidine lies in the plane of the pyrimidine,
while the phenyl sits out of plane. The presence of a 2-
substituent would encourage this geometry. Introduc-
tion of larger 2-substituents than the methyl (6 and 7)
was less potent, showing that they are too large for
the pocket.
incorporated later, but these compounds mostly fall out-
side the scope of this manuscript. Several analogues with
a sulfonamide 2-substituent will be discussed to show
that the amide is not crucial for activity and that the
SAR at the 4-position is consistent.
Substituents to be used at the 4-position were chosen
with the likely binding mode in mind. Compound 2
was docked into the published X-ray structure of Lck
(Fig. 1).13 In the model, the pyrimidine forms two typi-
cal hydrogen bonds to the hinge of the Lck ATP-site
(between the backbone NH and carbonyl of Met319,
and the N1 atom of the pyrimidine and the 2-anilino
The 5-hydroxyl group was also of great importance.
Removing this entirely (8) led to a loss of 7-fold in po-
tency, which was only partially restored by moving the
hydroxyl to the 4- or 6-positions (9 and 10). Replacing
the hydroxyl with F or Cl (11 and 12) resulted in even
lower activity. In an attempt to see whether the acceptor
or donor functionality of the hydroxyl was responsible
for activity, it was replaced by methoxy (13) and amino
(14) groups. Both showed similar activity to the 5-H
compound 8. It seems that for greatest activity the 5-
substituent requires both the acceptor and donor prop-
erties of the phenol. Slightly larger groups containing
H-bond acceptors and donors were tried (e.g., 15–18)
which all led to much reduced activity, suggesting that
space at the 5-position is limited. Compounds lacking
the 2-Me and the 5-hydroxyl were less potent still, gen-
erally having IC50 > 1 lM (e.g., 19–22). No simple ana-
logues were able to achieve the activity of 2.
˚
In the model, the oxygen of the 5-hydroxyl is 2.3 A away
from the backbone NH of Asp382. This hydrogen bond
explains the importance of the 5-hydroxyl acceptor
Figure 1. Model of compound 2 docked into the ATP site of Lck.13
Hydrogen bonds are shown as purple dotted lines.

P. Bamborough et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4363–4368
Table 1. Lck inhibition by 2,4-dianilino pyrimidines 2–25 (nM)9
4365
H
N
N
N
R2
R5
R1
R3
R4
HN
R6
Compound
R1
R2
R3
R4
R5
R6
IC50
2
3
CONH2
CONH2
CONH2
CONH2
CONH2
CONH2
CONH2
CONH2
CONH2
CONH2
CONH2
CONH2
CONH2
CONH2
CONH2
CONH2
CONH2
CONH2
CONH2
CONH2
CONH2
SO2NH2
SO2NH2
SO2NH2
Me
H
H
H
Me
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
OH
OH
H
H
H
H
H
H
H
H
OH
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
13
130
73
4
H
OH
OH
5
H
Br
Me
H
H
H
OH
H
H
H
H
H
H
H
H
H
H
H
H
F
190
41
6
OH
OH
7
iPr
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
Me
H
880
89
8
H
9
H
40
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
H
100
270
380
86
F
Cl
OMe
NH2
97
CONH2
CONMe2
NHSO2Me
CH2NH2
OMe
1300
9300
7700
3400
2700
3300
1800
1100
H
F
H
N(Me)COMe
H
OH
F
Me
Me
F
H
H
F
8.5
OMe
H
54
350
functionality. If the hydroxyl lone pair points towards
Asp382, its hydrogen is directed towards the sidechain
Pharmacokinetic properties of the most potent phenol
compounds were determined in mice and found to be
poor (see Table 5) making them unsuitable candidates
for oral administration. In particular, they exhibited
high clearance, short half-life and low oral bioavailabil-
ity. It was thought likely that the 5-hydroxyl aniline at
the pyrimidinyl 4-position was responsible for this, given
the susceptibility of phenols to phase II metabolism.18
The SAR above indicates a preference for a very small
group at the 5-hydroxyl position, ideally with H-bond
acceptor and donor vectors in the correct directions.
Finding a suitable replacement with these properties
was a significant challenge. To try to incorporate these
features, a second array (26–43) was prepared using con-
ditions described above in Scheme 1, in which the phe-
nol group was replaced with fused bicyclic heterocycles
containing at least some of the H-bonding potential of
the 5-hydroxyl group.19
˚
of Glu288. While the 2.7 A distance is quite long for a
direct H-bond, it seems likely that either Glu288 moves
closer when the compound binds or that a water-medi-
ated interaction is present. After this work was com-
pleted, a crystal structure of 1 bound to p38 MAP
kinase was solved which confirmed these predictions,
at least for the quinoline analogue bound to a surrogate
kinase.15 A similar result has been reported for a benz-
imidazole template, the most potent example of which
contained the same 2-methoxy 5-hydroxy aniline. The
position of the aniline in the crystal structure of a benz-
imidazole in Hck agrees with the docking model of 2
and shows that the hydroxyl donates a hydrogen bond
to Glu288.16 This proposed orientation of the bicyclic
aniline is flipped relative to that seen in the X-ray struc-
ture of unactivated Src with a quinazoline bearing a (5-
chloro-1,3-benzodioxol-4-yl)amine 4-substituent.17 If
our binding mode is correct, differences between the ac-
tive Lck conformation of Asp382 and its equivalent in
inactive Src (Asp404) may account for the difference in
compound binding.
Table 2 shows data for selected compounds in which the
4-anilino position of the pyrimidine is a bicyclic ring
fused between the 3- and 4-positions relative to the lin-
ker. These compounds (26–31) were only weakly active
against Lck, with IC50s in the 1–10 lM range.
Analogues of many compounds were prepared with 3-
aminobenzenesulfonamide instead of 3-aminobenza-
mide. Their activity was very similar to or slightly great-
er than their amide analogues. For example, compare 23
to 2, 24 to 13 and 25 to 22.
Table 3 shows data for compounds where the rings were
fused between the 2- and 3-positions relative to the linker.
These were more successful than the 3,4-fused rings, with
several examples showing IC50 < 1 lM (32–34), although

4366
P. Bamborough et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4363–4368
Table 2. Lck enzyme activity of pyrimidines containing 4-substituted
amino bicycles, with rings fused between the 3- and 4-positions (nM)9
Table 3. Lck enzyme activity of pyrimidines containing 4-substituted
amino bicycles, with rings fused between the 2- and 3-positions (nM)9
H
H
N
N
N
N
N
N
R1
HN
HN
R2
H2N
O
R
R1 = CONH2
Compound
R
IC50
3000
Compound
R2
IC50
560
S
*
26
*
N
32
Me
N
N
*
27
28
29
30
31
16,000
3000
1800
4400
8200
*
H
N
N
N
33
34
490
N
*
N
*
H
N
H
N
N
1400
N
N
*
*
N
35
36
37
38
5600
100
12
Me
Me
N
H
H
N
*
*
*
N
N
N
N
H
*
Me
N
Me
N
N
N
H
*
some were still weak (35). The indazole 36 was the most
potent, comparing favourably with the phenol 3.
520
N
N
Me
H
Having established that the indazole ring was an active
isostere of the 5-phenol, the possibility of including a
methyl ortho to the linker was then explored. This was
successfully introduced, giving 37, which has equivalent
Lck activity to 2. As expected, moving the methyl
around the ring led to compounds with lower potency
(38 and 39).
*
N
39
40
55
N
H
Me
*
1600
N
N
Me
Blocking the indazole H-bond donor capability by N-
methylation resulted in over 10-fold loss of activity
(compare 40 to 36). Methylation next to the aniline lin-
ker was unable to rescue the activity (41). The impor-
tance of the H-bond donor NH at the 1-position of
the 1H-indazol-4-amine is further shown by the 14-fold
loss of activity when the indazole NH is moved around
the ring (34).
*
Me
41
2800
N
N
Me
R1 = SO2NH2
*
42
43
33
19
N
As before, 3-aminobenzenesulfonamide was tolerated as
the pyrimidinyl 2-substituent (compare 42 to 36 and 43
to 37).
N
H
*
Me
N
These observations were consistent with the modelling
predictions. Figure 2 shows the docking of 37 compared
to 2. The hydrogen bonding vectors of the indazole mi-
mic those of the phenol very closely, and maintain the
direct interaction with Asp382 and the direct or water-
mediated H-bond to Glu288. The loss of these two inter-
actions explains the lower potency of all of the com-
pounds in which they are disrupted.
N
H
The introduction of the indazole group had an unex-
pectedly beneficial effect on the protein kinase selectivity
profile. The more potent phenol and indazole di-anilino
pyrimidines did not show submicromolar inhibition of

P. Bamborough et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4363–4368
4367
hydroxyaniline. However, these compounds had poor
pharmacokinetic properties. Simple 4-position alterna-
tives were unable to reach the same level of activity.
One of the highest priority isosteric replacements based
on modelling predictions was 5-methyl-1H-indazol-4-
amine. Pyrimidines containing this group at the pyri-
midinyl 4-position retained the high enzyme activity of
the phenols (compounds 37 and 43 had Lck IC50 of
12 nM and 19 nM, respectively) with improved selectiv-
ity. This result provided useful supporting evidence for
the binding mode and the role of hydrogen bonding in
the inner part of the ATP-binding site for Lck potency.
Most significantly, they had much improved pharmaco-
kinetic parameters, including lower plasma clearance,
longer half-life and greater oral bioavailability. As well
as their significance for the 2,4-dianilino pyrimidine ser-
ies, these results may be of interest to other medicinal
chemistry programmes where phenols are important
for activity.
Figure 2. Model of compound 37 (light blue) bound into the ATP site
of Lck. Compound 2 (green) is superimposed to show the similar
hydrogen bonding vectors.
Ser/Thr kinases tested, including GSK-3b, IKK-b and
JNK3. However, the phenols 2 and 23 had appreciable
activity against some tyrosine kinases including Btk,
Lyn, Syk and Txk (Table 4). The direct indazole ana-
logues 37 and 43 exhibited lower activity against these
kinases.
Acknowledgments
The authors wish to thank Rona G. Davis-Ward and
Derek N. Evans for their synthetic assistance.
The pharmacokinetic parameters of key compounds
were determined in mice.20 The data shown in Table 5
summarise the pharmacokinetic differences between di-
rect analogues 23 and 43, where the only structural dif-
ference is the replacement of the phenol with the
indazole isostere. The indazole has a considerably lower
plasma clearance and greater oral bioavailability than
the phenol. This suggests that the phenol was metaboli-
cally more labile and was the main reason for the poor
pharmacokinetic profile of 23.
References and notes
1. Molina, T.; Kishihara, K.; Siderovski, D.; van Ewijk, W.;
Narendran, A.; Timms, E.; Wakeham, A.; Paige, C.;
Hartmann, K.-U.; Veilette, A. Nature 1992, 357, 161.
2. Levin, S.; Anderson, S.; Forbush, K.; Perlmutter, R.
EMBO J. 1993, 12, 1671.
3. Lovatt, M.; Filby, A.; Parravicini, V.; Werlen, G.; Palmer,
E.; Zamoyska, R. Mol. Cell Biol. 2006, 26, 8655.
4. van Oers, N.; Killeen, N.; Weiss, A. J. Exp. Med. 1996,
183, 1053.
In summary, 2,4-dianilino pyrimidines are a class of po-
tent Lck inhibitors. The greatest activity was achieved
when the pyrimidyl 4-substituent was 2-methyl 5-
5. Weiss, A.; Littman, D. Cell 1994, 76, 263.
6. For reviews, see Weil, R.; Veillette, A. Curr. Topics
Microbiol. Immunol. 1996, 205, 63; Bolen, J.; Brugge, J.
Ann. Rev. Immunol. 1997, 15, 371; Kamens, Joanne S.;
Ratnofsky, Sheldon E.; Hirst, Gavin C. Curr. Opin. Invest.
Drugs 2001, 2, 1213.
Table 4. Tyrosine kinase selectivity IC50 values (nM)
Compound 4-substituent Lck
Btk Lyn Syk Txk
7. Golas, J.; Arndt, K.; Etienne, C.; Lucas, J.; Nardin, D.;
Gibbons, J.; Frost, P.; Ye, F.; Boschelli, D.; Boschelli, F.
Cancer Res. 2003, 63, 375.
2
37
Phenol
Indazole
13
12
9
97
3
40
26
220
2
28
8. Hennequin, L.; Thomas, A.; Johnstone, C.; Stokes, E.;
Ple, P.; Lohmann, J.; Ogilvie, D.; Dukes, M.; Wedge, S.;
Curwen, J.; Kendrew, J.; Lambert-van der Brempt, C. J.
Med. Chem. 1999, 42, 5369.
23
43
Phenol
Indazole
8.5 n.t.
19 69
1
430
4
620
0.4
220
n.t., not tested.
9. Lck activity was assessed using a TR-FRET (time-
resolved fluorescence resonance energy transfer) assay.
Murine Lck was overexpressed in insect cells and used as a
semi-purified lysate. Lck was pre-activated at room
temperature for 30 min in the presence of 10 mM MgCl2,
100 lM ATP, then diluted 100-fold in 100 mM Hepes, pH
7.4. 15 ll of activated Lck was added to wells containing
1 ll of various concentrations of compound or DMSO
vehicle (3.3% final) in NUNC 384-well black plates and
incubated for 15 min at room temperature. The reaction
was initiated by the addition of 15 ll of substrate reagent
containing biotinylated peptide, Biotin-EEEEYFELV
(0.5 lM final), ATP (120 lM final) and MgCl2 (20 mM
final) in 100 mM Hepes, pH 7.4. The reaction mixture was
Table 5. Summary of pharmacokinetic parameters determined in
mice20
Compound
Route
23 (phenol)
43 (indazole)
IV PO
IV
PO
Dose (mg/kg)
Clp (ml/min/kg)
Vss (l/kg)
2.5
10
—
—
—
2
1
1
65.5
0.3
22.4
0.7
0.53
743
—
—
—
—
190
25
T1/2 (h)
AUC0–t (ng h/ml)
F (%)
0.12
635
—
<1

4368
P. Bamborough et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4363–4368
incubated for 30 min at room temperature, then termi-
14. Beattie, J.; Breault, G.; Ellston, R.; Green, S.; Jewsbury,
P.; Midgley, C.; Naven, R.; Minshull, C.; Pauptit, R.;
Tucker, J.; Pease, J. Bioorg. Med. Chem. Lett. 2003, 13,
2955.
15. Manuscript in preparation, Bioorg. Med. Chem. Lett.
16. Sabat, M.; VanRens, J.; Laufersweiler, M.; Brugel, T.;
Maier, J.; Golebiowski, A.; De, B.; Easwaran, V.; Hsieh,
L.; Walter, R.; Mekel, M.; Evdokimov, A.; Janusz, M.
Bioorg. Med. Chem. Lett. 2006, 16, 5973.
17. Hennequin, L.; Allen, J.; Breed, J.; Curwen, J.; Fennell,
M.; Green, T.; Lambert-van der Brempt, C.; Morgentin,
R.; Norman, R.; Olivier, A.; Otterbein, L.; Ple, P.; Warin,
N.; Costello, G. J. Med. Chem. 2006, 49, 6465.
18. Jeong, E.; Liu, X.; Chen, J.; Hu, M. Curr. Drug Metab.
2005, 6, 455.
nated by the addition of 30 ll of read reagent containing
60 mM EDTA, 150 mM NaCl, 25 nM Streptavidin APC
(Perkin-Elmer, Beaconsfield, Bucks, UK), 0.4 nM anti-
phosphotyrosine antibody labelled with W-1024 europium
chelate (Wallac OY, Turku, Finland) in 40 mM Hepes, pH
7.4, 0.25% BSA. The reaction mixture was further
incubated for 15 min at room temperature. The degree
of phosphorylation of Biotin-EEEEYFELV was measured
using a Packard Discovery plate reader as a ratio of
specific 665 nm energy transfer signal to reference Eu
620 nm signal. All values quoted are means of IC50 results.
The average standard deviation of pIC50 retests was 0.3.
Human Lck IC50 values were found to be in close
agreement with mouse Lck, typically within twofold
(e.g., 2 had human Lck IC50 of 14 nM).
19. Compounds in Tables 2 and 3 were prepared as described
in Scheme 1, except for: Compound 36—the second
aniline displacement step was performed in a refluxing
mixture of n-BuOH/MeOH (1:1); 3,4-dimethoxy benzyl N-
protection of the indazole nucleophile was employed in the
preparation of 37 and 43.
10. See numerous patents, including Davis, P.; Moffat, D.;
Davis, J.; Hutchings, M. WO1997019065, 1997.
11. Dev, I.; Dornsife, R.; Hopper, T.; Onori, J.; Miller, C.;
Harrington, L.; Dold, K.; Mullin, R.; Johnson, J.; Crosby,
R.; Truesdale, A.; Epperly, A.; Hinkle, K.; Cheung, M.;
Stafford, J.; Luttrell, D.; Kumar, R. Br. J. Cancer 2004,
91, 1391.
12. Compound 14 was formed from the carboxylic acid
precursor under standard Curtius rearrangement condi-
tions: (a) diphenylphosphoryl azide, Et3N, t-BuOH,
reflux; (b) 4 M HCl in dioxane, CH2Cl2, rt. Compounds
15 and 16 were made from the carboxylic acid precursor
under standard amide coupling conditions. Phthalimide
protection was employed to protect the primary amine
functionality in compound 18.
13. Coordinates of Lck were taken from a complex with
AMP-PNP (PDB entry 1qpc) Zhu, X.; Kim, J.; Rose, P.;
Stover, D.; Toledo, L.; Zhao, H.; Morgenstern, A.
Structure 1999, 7, 651; Gold version 1.2 (CCDC, Cam-
bridge, UK) Jones, G.; Willet, P.; Glen, R.; Leach, A.;
Taylor, R. J. Mol. Biol. 1997, 267, 727, was used to dock
compounds, with H-bonding constraints applied to ensure
that the pyrimidine interacted with the hinge.
20. The pharmacokinetic parameters in female BalbC mice
were determined following intravenous (IV) and oral (PO)
administration at various doses between 1 and 10 mg/kg.
Compounds were administered as solutions in formula-
tions typically consisting of 15% DMSO:45% cyclodextrin
(25% aq):40% water or 5% DMSO:40% Vit E/EtOH
(60:40):40% PEG400:15% mannitol and blood collected
over an 8-h time period. Plasma was prepared following
centrifugation and compound extracted from 100 ll
plasma using protein precipitation with acetonitrile. Sam-
ples were then evaporated under nitrogen and re-sus-
pended in 100 ll of 10:90 acetonitrile:water. Analysis was
performed using LC–MS/MS on the API365 with a
5.6 min fast gradient comprising 0.1% formic acid in
water and 0.1% formic acid in acetonitrile (mobile phases),
40 ll injection volume, flow rate 4 ml/min and ODS3
Prodigy column (5 cm · 2.1 mm, 5 lm). Pharmacokinetic
data were generated using PKTools.