N-1 substituted pyrimidin-4-ones: Novel, orally active inhibitors of lipoprotein-associated phospholipase A2

Article abstract of DOI:10.1016/S0960-894X(00)00510-2

From two related series of 2-(alkylthio)-pyrimidones, a novel series of 1-((amidolinked)-alkyl)-pyrimidones has been designed as nanomolar inhibitors of human lipoprotein-associated phopholipase A₂. These compounds show greatly enhanced activity in isolated plasma. Selected derivatives such as compounds 51 and 52 are orally active with a good duration of action. (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0960-894X(00)00510-2

Bioorganic & Medicinal Chemistry Letters 10 (2000) 2557±2561
N-1 Substituted Pyrimidin-4-ones: Novel, Orally Active Inhibitors
of Lipoprotein-Associated Phospholipase A2
Helen F. Boyd, Stephen C. M. Fell, Sean T. Flynn, Deirdre M. B. Hickey, Robert J. Ife,
Colin A. Leach, Colin H. Macphee, Kevin J. Milliner, Kitty E. Moores, Ivan L. Pinto,
Rod A. Porter, D. Anthony Rawlings, Stephen A. Smith,* Ian G. Stans®eld,
David G. Tew, Colin J. Theobald and Caroline M. Whittaker
SmithKline Beecham Pharmaceuticals, New Frontiers Science Park, Third Avenue, Harlow, Essex, CM19 5AW, UK
Received 24 July 2000; revised 28 August 2000; accepted 6 September 2000
AbstractÐFrom two related series of 2-(alkylthio)-pyrimidones, a novel series of 1-((amidolinked)-alkyl)-pyrimidones has been
designed as nanomolar inhibitors of human lipoprotein-associated phopholipase A2. These compounds show greatly enhanced
activity in isolated plasma. Selected derivatives such as compounds 51 and 52 are orally active with a good duration of action.
# 2000 Elsevier Science Ltd. All rights reserved.
The oxidative modi®cation of low density lipoprotein
(LDL) has been suggested to play a key role in athero-
sclerosis through the release of a pro-in¯ammatory oxi-
dative signal.1 Whilst the exact mechanism for the
release of such a message has not been fully char-
acterised, it has been shown that a phospholipase asso-
ciated predominantly with LDL, termed lipoprotein-
associated phospholipase A2 (Lp-PLA2),2,3 is able to
hydrolyse oxidised LDL into lysophosphatidylcholine
(lyso-PtdCho) and oxidised fatty acids. Both of these
hydrolysis products are known to be pro-in¯ammatory
and have been implicated in atherosclerosis.1 Indeed,
the increased levels of lyso-PtdCho in oxidised LDL can
be completely accounted for by Lp-PLA2.1 With this
evidence in hand, we decided that there was a clear need
for inhibitors of Lp-PLA2 to more fully understand the
role of this lipase in the atherosclerotic process.
nitrogen substitution (e.g. 1 R1ˆH) on potency, the
potential for improved in vivo properties and, for selec-
ted compounds, activity in the Watanabe hereditable
hyperlipidaemic rabbit (WHHL rabbit).
In a recent report4 we described the identi®cation of a
series of 2-(alkylthio)-pyrimidones, such as 1a and 1b, as
fully reversible inhibitors of Lp-PLA2. Further evalua-
tion of these compounds however, indicated a low level
of inhibition in whole plasma and as a result, little
activity in vivo. In this communication we describe our
follow-up studies to investigate the e€ect of pyrimidone
Attempts to N-alkylate compound 1a under a range of
conditions (e.g., sodium ethoxide or Hunig's base in
ethanol or dimethylformamide, potassium hydroxide/
tetrabutylammonium bromide in tetrahydrofuran),
indicated that N-3 alkylation predominated (compound
2 was made in this way). As a result, N-1 alkylated
adducts were made via acylisothiocyanate 5 (Scheme 1).5,6
This key intermediate is prepared from propionic ester 3
which is itself made by a Heck reaction with 5-bromo-
pyrimidine followed by reduction. Ester 3 is formylated,
O-alkylated and carefully hydrolysed to acid 4 which is
*Corresponding author. Fax +44-1279-627841; e-mail: Stephen_1_
Smith@sbphrd.com
0960-894X/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(00)00510-2

2558
H. F. Boyd et al. / Bioorg. Med. Chem. Lett. 10 (2000) 2557±2561
Scheme 1. Reagents: (i) (a) CH2ˆCHCO2Et, Pd(OAc)2, P(o-Tol)3, Et3N, (b) H2, Pd/C, EtOH/Et3N; (ii) (a) HCO2Et, NaH, DME, (b) Me2SO4,
K2CO3, DMF, (c) NaOH, H2O; (iii) (a) (COCl)2, ClCH2CH2Cl, (b) KSCN, CH3CN; (iv) R1NH2, DMF then NaOEt; (v) R2Cl/Br, iPr2NEt, CH2Cl2.
converted to the required acylisothiocyanate. Addition
of an amine to 5 followed by ring closure gives thiour-
acils 6 regiospeci®cally. These were S-alkylated to give
the required pyrimidones 7. For those compounds
bearing additional functionality in R1 (Table 3), further
chemistry, e.g., hydrolysis of an ester (R1=(CH2)n-
CO2Et) followed by amide coupling (carbodiimide/
HOBT), was performed after S-alkylation. Whenever
possible (e.g., optimisation of R2, amide formation within
R1) compounds were prepared using array techniques.
We initially focused our attention on those compounds
bearing a long arylalkyl substituent on the C-2 sulfur
atom. This work showed N-1 substitution (Table 1)
to be preferred over N-3 (cf. compounds 8 and 2). Fur-
thermore, although the introduction of an N-1 sub-
stituent gave some useful enhancements in potency
against the isolated enzyme (cf. compound 1a with 8, 14
and 15), the most marked e€ect was the much improved
activity of these compounds in whole plasma. Potency
in rabbit plasma was generally lower than in human
plasma suggesting that compounds may be more potent
in man than in the rabbit.
All structures in Tables 1±3 were evaluated using human
Lp-PLA2. Assays were performed in duplicate.4 In
order to factor in any non-speci®c binding e€ects in
plasma, the majority of compounds were also assessed
against the plasma enzyme in both whole human and
WHHL rabbit plasma at a single concentration of inhi-
bitor.7 Good activity in rabbit and human plasma was
required before compounds were evaluated in vivo in
WHHL rabbits.8
Table 2. E€ect of N-1 substitution in the S-benzyl series
Inhibition in plasma
No.a
X
R1
IC50 Human % Rabbit %
(nM) @ 300 nM
Table 1. E€ect of N-1 substitution
@ 1 mM
1b
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
H
H
H
H
H
H
H
H
H
4-F
4-Cl
4-Me
3-Cl
2-F
3-F
4-F
3,4-diF
4-F
4-F
3,4-diF
H
Me
1110
660
1200
450
210
300
150
210
350
285
1400
2850
960
450
250
300
235
250
200
110
4
54
27
30
35
18
46
34
59
55
NTb
NT
NT
41
44
49
53
42
60
75
4
33
25
42
73
28
37
40
75
32
NT
NT
NT
40
33
40
47
65
82
87
CH2CH2OMe
CH2Ph
CH2CH2Ph
CH2CH2CH2Ph
CH2(2-furyl)
CH2(2-thienyl)
CH2CH2(2-thienyl)
Me
CH2(2-furyl)
CH2(2-furyl)
CH2(2-furyl)
CH2(2-furyl)
CH2(2-furyl)
CH2(2-furyl)
CH2(2-furyl)
CH2CH2(2-thienyl)
CH2CH2Ph
CH2CH2Ph
Inhibition in plasma
No.a
R1
IC50
(nM)
Human %
@ 300 nM
Rabbit %
@ 1 mM
1a
8
9
10
11
12
13
14
15
H
Me
54
19
39
110
36
34
90
2
53
35
41
29
28
10
36
27
0
23
38
20
14
45
9
Et
CH2CH2OMe
CH2Ph
CH2CH2Ph
CH2CH2CH2Ph
CH2(2-furyl)
CH2(2-thienyl)
16
11
30
31
aAll new compounds gave satisfactory analytical/spectral data.9
bNot tested.
aAll new compounds gave satisfactory analytical/spectral data.9

H. F. Boyd et al. / Bioorg. Med. Chem. Lett. 10 (2000) 2557±2561
Table 3. The introduction of a lipophilic substituent at N-1
2559
Plasma inhib. (%) @ 100 nM
Human Rabbit
No.a
R1
IC50 nM
35
36
37
(CH2)9Ph
n-C11H23
(CH2)2Ph(4-C6H13
25
19
12
10
11
5
10
10
4
)
24
30
33
Me
CH2(2-furyl)
(CH2)2Ph
285
300
200
28
26
27
6
7
29
38
39
40
41
42
43
CH2CONH(CH2)6Ph(4-F)
CH2CONHC7H15
CH2CONHC8H17
2
19
7
0.3
15
1
67
37
43
86
40
79
33
29
38
54
30
39
CH2CONHC12H25
CH2CONMeC8H17
CH2CONMeC12H25
44
45
46
47
48
(CH2)3CONHC6H13
(CH2)3CONHC7H15
(CH2)3CONHC8H17
(CH2)3CONHC10H21
(CH2)3CONMeC7H15
7
2
0.4
0.1
3
63
78
85
94
18
59
62
55
77
45
49
50
51
52
53
(E) CH2CONH(CH2)8CHˆCHC8H17
(Z) CH2CONH(CH2)8CHˆCHC8H17
(E) (CH2)3CONH(CH2)8CHˆCHC8H17
(Z) (CH2)3CONH(CH2)8CHˆCHC8H17
(CH2)3CONHC18H37
2
1
0.7
0.4
4
55
66
71
85
29
32
38
47
59
21
54
55
56
(CH2)5CONHC4H9
(CH2)2NHCOC7H15
(CH2)2NHCONHC7H15
260
140
70
4
8
18
0
9
10
aAll new compounds gave satisfactory analytical/spectral data.9
Although potency against Lp-PLA2 was lower in the S-
benzyl series (Table 2) than for the longer chain com-
pounds of Table 1, activity in human plasma was
broadly similar for compounds with similar N-1 sub-
stituents (cf. compounds 16 and 8, 21 and 14), suggest-
ing a reduced level of non-speci®c binding in plasma for
the S-benzyl series. The e€ect of substitution on the S-
benzyl group was striking, with groups larger than
¯uoro causing a marked reduction in potency (cf. com-
pound 21 with 25±31 and 19 with 33 and 34). For sub-
sequent work the 4-¯uorobenzyl group was chosen on
the grounds of potential metabolic advantage (i.e., a
deactivated aromatic ring) with minimal increase in
molecular weight.
Based on the above results and the knowledge that
further simple modi®cation of the structures of Tables 1
and 2 had not given further enhancements in activity,
we speculated that potent inhibitors would be obtained
if the long arylalkyl chain on sulfur as in Table 1 was
moved to N-1 whilst maintaining the S-benzyl group of
the compounds of Table 2 (see Table 3). Pleasingly,
compounds 35 and 12 showed similar pro®les, con®rm-
ing our hypothesis (note the lower concentration used
for the single point determination of plasma
inhibition).10 We also showed that high potency against
Lp-PLA2 could be achieved by either extending the
phenethyl group of compound 33 or, perhaps more
usefully, by the replacement of the arylalkyl group by a
long alkyl chain (lower MWt, cf. 36 with 35 and 37).
Our real breakthrough however came when, in an e€ort
to make these inhibitors less lipophilic, we introduced
an amide group into the C11 chain of compound 36.
Both acetamide and butyramide derivatives11 proved
Figure 1. Inhibition of plasma Lp-PLA2 in the WHHL rabbit @
10 mg/kg (n=2).

2560
H. F. Boyd et al. / Bioorg. Med. Chem. Lett. 10 (2000) 2557±2561
not only more potent against Lp-PLA2 but were con-
siderably more active in whole plasma (cf. 40 and 44
with 36). The corresponding hexanamides (e.g. 54)
proved much less active and were not considered further,
nor were the reversed amides (cf. 55 and 44) or ureas (cf.
56 and 45).
15 s (rabbit plasma) or 60 s (human plasma). The reaction was
stopped by vortexing with 600 mL of CHCl3/MeOH (2:1) and
the CHCl3 and aqueous layers were separated by centrifuga-
tion. The aqueous layer was removed (250 mL) and vortexed
with 250 mL of CHCl3. The aqueous layer was again removed
and the [3H]acetate determined by scintillation counting. The
inhibitory e€ect of various compounds was determined by
mixing whole plasma (190 mL) and 2 mL of a DMSO solution
of compound and preincubating at 37 ꢁC for 10 min before
running the enzyme assay as above. Results at a single con-
centration proved very reproducible and were used for assess-
ing SAR trends. A full IC50 curve was generated (n=3) for
compounds of most interest (Figs. 2 and 3) and con®rmed
assay reproducibility.
Increasing the length of the alkyl chain increased activ-
ity, giving sub-nanomolar inhibitors in both the acet-
amide and the butyramide series (cf. 39±41, also 44±47),
although there appeared a maximum for this e€ect with
the highly lipophilic C18 derivatives 49±53 proving to be
no more potent than, for example, compounds 41 and
47. Amide N-methylation had little e€ect on potency
against Lp-PLA2. Although a similar e€ect was seen in
whole plasma for the acetamides (cf. 42 and 40, 43 and
41), tertiary butyramides proved signi®cantly less eca-
cious in plasma than their secondary counterparts (e.g.
48 versus 45).
Selected compounds from Table 3 were evaluated in the
WHHL rabbit (at 10 mg/kg po).8 Butyramides 51 and
52 from the C18 amide series proved the most eca-
cious, demonstrating good levels of inhibition alongside
a promising duration of action (ꢀ5 h, Fig. 1).
In conclusion, we have shown that orally active, sub-
nanomolar inhibitors of Lp-PLA2 may be obtained by
the introduction of a suitable amidic N-1 substituent
into our initially optimised series of pyrimidone inhibi-
tors.4 Furthermore, our evaluation of the role of Lp-
PLA2 in the atherosclerotic disease process and as a
target for therapeutic intervention should be markedly
enhanced by the development of potent, orally active
inhibitors such as compounds 51 and 52.
Figure 2.
References and Notes
1. Macphee, C. H.; Moores, K. E.; Boyd, H. F.; Dhanak, D.;
Ife, R. J.; Leach, C. A.; Leakes, D. S.; Milliner, K. J.; Patter-
son, R. A.; Suckling, K. E.; Tew, D. G.; Hickey, D. M. B.
Biochem. J. 1999, 338, 479.
2. Tew, D. G.; Boyd, H. F.; Ashman, S.; Theobald, C. J.;
Leach, C. A. Biochemistry 1998, 37, 10087.
Figure 3.
3. The enzyme is also known as PAF acetyl hydrolase, though
it is clear that the enzyme has much broader substrate speci®-
city than this name implies. See Tew, D. G.; Southan, C.; Rice,
S. Q. J.; Lawrence, M. P.; Haodong, L.; Boyd, H. F.; Moores,
K.; Gloger, I. S.; Macphee, C. H. Atheroscler. Thromb. Vasc.
Biol. 1996, 16(4), 591.
8. Watanabe heritable hyperlipidaemic rabbits were fasted
overnight prior to dosing and were allowed to feed ad lib
immediately following dosing. All dosing was by gavage using
3 mL of PEG300 containing compound. Prior to dosing and at
0.5, 1, 3 and 5 h post dosing, 1.5 mL blood samples were taken
via the ear vein and transferred to Eppendorf tubes containing
30 mL 250 mM EDTA solution. Plasma was immediately pre-
pared by microfuging for 30 s and frozen on dry ice to await
analysis for plasma PLA2 activity using the method of ref 7. Per-
centage inhibitions were determined relative to the pre dose value.
9. Representative examples: Compound 45 (250 MHz) 1H
NMR (CDCl3) d 0.88 (3H, t), 1.27 (8H, m), 1.47 (2H, m),
1.98±2.25 (4H, m), 3.20 (2H, m), 3.69 (2H, s), 3.92 (2H, t), 4.52
(2H, s), 5.42 (1H, bm), 6.98 (2H, m), 7.25 (1H, s), 7.37 (2H,
m), 8.68 (2H, s), 9.08 (1H, s); Mass Spectrum (APCI+)
4. Boyd, H. F.; Flynn, S. F.; Hickey, D. M. B.; Ife, R. J.;
Jones, M.; Leach, C. A.; Macphee, C. H.; Milliner, K. J.;
Rawlings, D. A.; Slingsby, B. P.; Smith, S. A.; Stans®eld, I.
G.; Tew, D. G.; Theobald, C. J. Bioorg. Med. Chem. Lett.
2000, 10, 395.
5. See Agathocleous, D. C.; Shaw, G. J. J. Chem. Soc., Perkin
Trans. 1 1993, 2555.
6. See Shaw, G.; Warrener, R. N. J. Chem. Soc. 1959, 50.
7. Activity in plasma was assessed following measurement of
PAF-AH activity. 10 mL [3H]PAF (50 mM) and whole plasma
were incubated at 37 ꢁC in a ®nal volume of 200 mL for either

H. F. Boyd et al. / Bioorg. Med. Chem. Lett. 10 (2000) 2557±2561
2561
M+1=512, C27H34FN5O2S requires 511. Compound 51
1
10. Compounds 24, 30 and 33 were screened at both con-
centrations and are included in Table 3 for comparison with
the data in Table 2.
11. The corresponding propionamides proved susceptible to
retro-Michael elimination of the N-1 substituent and as a
result were not evaluated.
(250 MHz) H NMR (CDCl3) d 0.88 (3H, t), 1.27 (22H, m),
1.47 (2H, m), 1.99±2.20 (8H, m), 3.20 (2H, m), 3.69 (2H, s),
3.92 (2H, m), 4.46 (2H, s), 5.37 (2H, m), 5.44 (1H, m), 6.99
(2H, m), 7.26 (1H, s), 7.37 (2H, m), 8.71 (2H, s), 9.08 (1H, s);
Mass Spectrum (APCI+) M+1=664, C38H54FN5O2S
requires 663.