Biological evaluation of 1-alkyl-3-phenylthioureas as orally active HDL-elevating agents

Article abstract of DOI:10.1016/j.bmcl.2005.09.034

A series of 1-alkyl-3-phenylthiourea analogues were prepared and evaluated as HDL- and Apo A-I-elevating and triglyceride-lowering agents. Several derivatives were superior to gemfibrozil (1). The optimal analogue 8d (HDL376) was shown to raise HDL choles

Full text of DOI:10.1016/j.bmcl.2005.09.034

Bioorganic & Medicinal Chemistry Letters 16 (2006) 113–117
Biological evaluation of 1-alkyl-3-phenylthioureas as orally
active HDL-elevating agents
Gary M. Coppola,a,* Robert E. Damon,a J. Bruce Eskesen,a
Dennis S. Franceb and James R. Paterniti, Jr.c
aDepartment of Diabetes and Metabolism, Novartis Institutes for Biomedical Research,
100 Technology Square, Cambridge, MA 02139, USA
bArQule, Inc., 19 Presidential Way, Woburn, MA 01801, USA
cAmylin Pharmaceuticals, 9373 Towne Centre Drive, San Diego, CA 92121, USA
Received 6 June 2005; revised 14 September 2005; accepted 15 September 2005
Available online 10 October 2005
Abstract—A series of 1-alkyl-3-phenylthiourea analogues were prepared and evaluated as HDL- and Apo A-I-elevating and
triglyceride-lowering agents. Several derivatives were superior to gemfibrozil (1). The optimal analogue 8d (HDL376) was shown
to raise HDL cholesterol in the rat, hamster, dog, and monkey models.
Ó 2005 Elsevier Ltd. All rights reserved.
Atherosclerosis is a complex disease where the progres-
sive accumulation of cholesterol within the arterial wall
eventually results in occlusion of the coronary or cere-
bral arteries ultimately leading to myocardial infarction
or stroke. It is well known that major risk factors for
atherosclerotic cardiovascular disease include such
dyslipidemias as elevated low-density lipoprotein
(LDL) cholesterol, low levels of high-density lipopro-
tein (HDL) cholesterol, and high levels of triglycerides.
Extensive epidemiological studies have shown a strong
inverse relationship between serum HDL cholesterol
levels and coronary heart disease.1–3 The Framingham
Heart Study showed that a 10 mg/dL increase in HDL
cholesterol was associated with a 19% decrease in coro-
nary artery disease death and a 12% decrease in all-
cause mortality.4 The Helsinki Heart Study5,6 and the
Veterans Administration HDL Intervention Trial7 dem-
onstrated a reduced incidence of cardiovascular events
with increased HDL cholesterol levels in response to
treatment with the PPARa agonist gemfibrozil (1).
These two clinical trials raised HDL cholesterol by 9%
and 6%, respectively.
is its ability to mediate cholesterol efflux in the reverse
cholesterol transport pathway. The HDL particle ex-
tracts cholesterol from cells,8 thus counteracting the ef-
fects of LDL cholesterol and subsequently preventing
the formation of foam cells, the genesis of atherosclerot-
ic lesions.9,10 The cholesterol-laden HDL particle is then
transported to the liver where the cholesterol is recycled
or removed by excretion in bile.3,10
Apolipoprotein A-I (Apo A-I), the major protein com-
ponent of HDL, actively regulates the function of
HDL. It has been demonstrated that transgenic mice
which overexpress the human Apo A-I gene resulted in
elevated levels of HDL cholesterol and significant pro-
tection from the development of aortic fatty streak
lesions.11 This protection was also seen in spontaneously
atherosclerotic Apo E-deficient mice.12,13
We have previously disclosed a series of N-phenylthio-
ureas as a new class of HDL-elevating agents.14,15 The
initial lead SDZ45-904 (2) was identified from our cor-
porate compound archive. Initial modifications com-
bined structural features from both gemfibrozil (1) and
2 to produce a ÔhybridÕ molecule 3, which was shown
to be as effective as 1 in raising HDL cholesterol and
Apo A-I levels.16 We then focused on probing the
SAR around compound 2. Varying the substitution pat-
tern in the aromatic ring did not improve on the overall
profile of 2 and generally elicited substantial reductions
One of the possible mechanisms for the protective nat-
ure of HDL against the development of atherosclerosis
*
Corresponding author. Tel.: +1 617 871 7310; fax: +1 617 871
7043; e-mail: gary.coppola@novartis.com
0960-894X/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.09.034

114
G. M. Coppola et al. / Bioorg. Med. Chem. Lett. 16 (2006) 113–117
in body weight gain.17,18 Chain modifications of 2 led
to conformationally restricted analogue (R)-4, which
was approximately 50% more potent than 1 in elevat-
ing HDL cholesterol and nearly twice that of 1 in rais-
ing Apo A-I levels. Total cholesterol was also raised
significantly with moderate reduction in body weight
gain. The (S)-enantiomer of 4 was equally efficacious
at raising HDL cholesterol but had no effect on apo
A-I.
resolution with (S)-mandelic acid. These were reacted
with 7 to give 9 and 10, respectively.
Cl
Cl
Me
Me
S
S
N
H
N
N
H
N
CH3
CH3
9
10
CH3
Cl
Compounds were administered in diet (ad libitum) to
groups of six male Sprague–Dawley rats for 8 days
as described in an earlier publication.16 Serum lipopro-
teins were separated by classical ultracentrifugation
(UC) techniques and cholesterol was quantitated by
enzymatic colorimetric methods. Serum Apo A-I was
measured by SDS–polyacrylamide gel electrophoresis20
or a rat competition ELISA.16 Concurrently, serum
lipoproteins were also analyzed by FPLC,21 which
was modified for robotic automation.22,23 Within each
study, gemfibrozil (1) was used as an internal positive
control.
S
COOH
OH
O
N
H
N
H
CH3
Cl
CH3
Cl
1
2
OH
S
S
COOH
N
H
N
H
N
H
N
CH3
CH3
3
4
Initial FPLC analysis of 2 provided conflicting results
when compared with the UC method. HDL cholesterol
values obtained by FPLC were significantly higher
than the UC values. We then analyzed the serum sam-
ples using density gradient ultracentrifugation.24 In
contrast to classical UC where a static density of
1.06 g/ml is used, we used seven escalating densities.
As seen in Figure 1, a density of 1.06 g/ml is sufficient
to account for virtually all of the HDL cholesterol of 1
(fractions 20–44) but not for 2 when the same fraction
numbers were collected. The HDL peak is shifted left
(toward the b-VLDL region) indicating the presence
of larger, more buoyant particles. However, all of the
fractions of this peak contain Apo A-I and no Apo
B, thus confirming the presence of HDL particles.
Within the series of compounds listed in Tables 1
and 2 the HDL cholesterol values measured by classi-
cal UC at a density of 1.06 g/ml were underestimated
in more than half of the cases.
A common feature of 2 and its analogues is the primary
alcohol, a potential metabolic liability. We were interest-
ed in exploring whether the alcohol could be replaced
and maintain potency. We also incorporated fast pro-
tein liquid chromatography (FPLC) as another method
for analyzing serum lipoproteins. This report will dis-
cuss the SAR around structural modifications to 2 as
well as differences in results obtained with both ultracen-
trifugation and FPLC techniques.
The aromatic substituted thioureas (6) were readily
prepared by reaction of a phenyl isothiocyanate with
propylamine in methylene chloride (Scheme 1). Con-
versely, chain variants 8 were synthesized by reaction
of 5-chloro-2-methylphenyl isothiocyanate (7) with an
appropriate amine. All amines were commercially avail-
able with the exception of the chiral 2-ethylpiperidine
derivatives. (S)-2-Ethylpiperidine was obtained by
resolution of 2-ethylpiperidine with (R)-mandelic acid.19
The corresponding (R)-enantiomer was obtained by
The first structural change of 2 was the replacement of
the hydroxyl group with methyl (6a, Table 1). Satisfy-
ingly, it exhibited a profile similar to that of 2 with a
smaller increase in total cholesterol (TC) and a smaller
decrease in body weight gain (WG). We then made some
minor changes to the aromatic substituents. Replacing
chlorine with fluorine or bromine (6b and 6c) retained
the HDL effects, however, TC levels were higher than
that of 6a. Replacing methyl with ethyl (6d) resulted in
a further increase of HDL cholesterol, although Apo
A-I levels were statistically no different from controls.
Replacement of methyl with methoxy (6e–g) resulted
in a loss of activity relative to 6a. The 2,5-dimethyl ana-
logue 6h, which has the same substitution pattern as
gemfibrozil, had a similar profile as 6a, although TC
levels were elevated relative to 6a and the WG effects
were decreased. Within this series of compounds, it ap-
pears that the 5-chloro-2-methyl analogue 6a has the
R1
R1
Me
H2N
S
Me
NCS
NCS
N
N
H
H
R2
Cl
R2
Cl
amine
S
R
N
N
H
H
Me
Me
8
7
Scheme 1. General preparation of arylthiourea derivatives.

G. M. Coppola et al. / Bioorg. Med. Chem. Lett. 16 (2006) 113–117
115
2000
1800
1600
1400
1200
1000
800
1.14
1.12
1.1
1.08
Control
1.06
1.04
1.02
1
Gemfibrozil
SDZ45-904
Density
600
400
0.98
0.96
200
0
1
6
11
16
21
26
31
36
41
Fraction No.
Figure 1. Rat serum density gradient cholesterol of SDZ45-904 (2) and gemfibrozil.
Table 1. Lipid profiles of aromatic variants 6 in cholesterol-fed male rats
6a R1
R2
Dose (mg/kg/day) HDL UC (%)b HDL FPLC (%)b Apo A-I (%)b Total cholesterol (%)b TG (%)b
WG (%)c
a
b
c
Cl
F
Me
Me
Me
Et
55f
57
63
70
83f
152
142
101
99
292f
312
306
369
97
54f
18
6f
54
À54f
À58
À78f
À77
À29
À33
À6
Br
Cl
Cl
46
47
28
À53
d
e
43 (NS)
58 (NS)
19 (NS)
À61
OMe 78
49
60
À43
f
OMe OMe 79
108
84
97
63
À23 (NS)
À54
14
g
h
1d
2e
Me
Me
OMe 71
133
71
29
À16
À38
2
Me
62
50
55
127
104
75
278
139g
241h
72
54
48
À59
À23
39
À29
À53
À104
UC, ultracentrifugation; TG, triglycerides; WG, weight gain.
a Satisfactory elemental analysis obtained for all compounds.
b Relative to untreated controls. Values are significantly different (p < 0.05) relative to untreated controls unless otherwise indicated (NS).
c % change in body weight gain relative to untreated controls.
d Mean value of 160 studies.
e Mean value of 26 studies.
f Mean value of 25 studies.
g Mean value of 97 studies.
h Mean value of 12 studies.
most desirable overall profile. This aromatic substitu-
tion pattern is consistent with earlier analogues reported
by us14–17 as well as others.25
reported17,26 were also the most active in this series,
raising HDL cholesterol 404% and 614%, respectively;
however, TC levels were also dramatically elevated
and large reductions in WG were observed.
We next focused our attention on optimizing the alkyl
chain (Table 2). Either decreasing or increasing the
chain length by one methylene unit (8a and 8b) de-
creased the HDL and Apo A-I response relative to 6a
while raising TC. The t-butyl (8c) and closely related
1,1-dimethylpropyl analogue 8f suffered dramatic loss
of activity. The isobutyl analogue 8d retained HDL
activity, increased Apo A-I relative to 6a, reduced TC le-
vel, and lowered adverse WG parameters to an accept-
able level. Addition of another methyl group to the
end of the chain (8e) resulted in some loss of HDL activ-
ity and an additional methyl next to the nitrogen (8g)
further reduced HDL activity.
In general, nearly all of the thiourea analogues were
effective at lowering TG levels. The liver enzymes
ALT, AST, and ALP were measured for all compounds.
Derivatives 6c, 6e, and 8b elevated ALP 44–49% and 6h,
75%. Elevation of ALT and AST was observed for 8a
(81%, 162%), 8b (93%, 79%), and 8e (111%, 49%). Com-
pounds 9 and 10 elevated both AST (85% and 80%,
respectively) and ALP (129% and 105%, respectively).
Compounds 2, 6a, 8d, 9, and 10 were measured for
PPARa and 6a, 8d, 9, and 10 were measured for PPARd
activity. All of them were found to be inactive in both
assays (EC50 > 100 lM).
The carbon versions (9 and 10) of the most active con-
formationally restricted alcohol analogue (4) previously
Inspection of the data obtained for all compounds
indicated that 8d possessed the optimal overall profile.

116
G. M. Coppola et al. / Bioorg. Med. Chem. Lett. 16 (2006) 113–117
Table 2. Lipid profiles of chain variants 8, 9, and 10 in cholesterol-fed male rats
8a
R
Dose (mg/kg/day) HDL UC (%)b HDL FPLC (%)b Apo A-I (%)b Total cholesterol (%)b TG (%)b
WG (%)c
a
Et
73
110
100e
30
104f
122
55
215
203e
22
260f
196
39
27
38e
60
77f
89
64
34
27
39
30
54
32
39e
À64
À107
À53e
9 (NS)
À12f
7
b
n-Bu
t-Bu
CH2CHMe2 79f
55e
77
À44
c
À15 (NS)
À11f
3 (NS)
À10
À31
d
À54f
À66
e
CH2CMe3
CMe2Et
88
81
f
À14 (NS)
À35
15
g
CHMeCMe3 82
93
109
136
404
614
139g
15 (NS)
97
À24
41
h
CH2Ph
77
62
63
50
135
124
129
104
À21
9
126
167
À56
À88
À105
2
10
1d
À54
À23
À29
UC, ultracentrifugation; TG, triglycerides; WG, weight gain.
a Satisfactory elemental analysis obtained for all compounds.
b Relative to untreated controls. Values are significantly different (p < 0.05) relative to untreated controls unless otherwise indicated (NS).
c % change in body weight gain relative to untreated controls.
d Mean value of 160 studies.
e Mean value of 2 studies.
f Mean value of 51 studies.
g Mean value of 97 studies.
Table 3. Percent increase of HDL cholesterol of HDL376 (8d) and gemfibrozil in alternate animal models
Model
HDL376
Dose (mg/kg/day)
Gemfibrozil
Dose (mg/kg/day)
NF rat
CF hamster
75
15
80
64
52
35
12
12
25
50
45
NA
NA
NA
NF hamster
NF dog
44
50
45
250
CF rhesus monkey
CF cynomolgus monkey
75
100
NF, normal-fed; CF, chow-fed; NA, not active. Good dose responses were observed in each of the species mentioned in this table.
4. Gordon, T.; Castelli, W. P.; Hjortland, M. C.; Kannel, W.
B.; Dawber, T. R. Am. J. Med. 1977, 62, 704.
HDL cholesterol and Apo A-I levels were elevated near-
ly twice that of gemfibrozil, while TC was reduced by
11%. It was also twice as effective as gemfibrozil at
reducing TG levels. Compound 8d (designated
HDL376) was further profiled against 1 in additional
animal models (Table 3). In all models, HDL376 elevat-
ed HDL cholesterol, whereas 1 was not active in the CF
hamster and NF hamster and dog. This suggests that the
mechanism of action of this class of compounds, while
currently unknown, may be broadly relevant to mam-
mals and not specific to the rat.
5. Frick, M. H.; Elo, O.; Kaapa, K.; Heinonen, O. P.;
Heinsalmi, P.; Helo, P.; Huttunen, J. K.; Kaitaniemi, P.;
Koskinen, P.; Manninen, V.; Maenpaa, H.; Malkonen,
M.; Manttari, M.; Norola, S.; Pasternack, A.; Pikkarai-
nen, J.; Romo, M.; Sjoblom, T.; Nikkila, E. A. N. Engl. J.
Med. 1987, 317, 1237.
6. Manninen, V.; Elo, M. O.; Frick, M. H.; Haapa, K.;
Heinonen, O. P.; Heinsalmi, P.; Helo, P.; Huttunen, J. K.;
Kaitaniemi, P.; Koskinen, P.; Ma¨enpa¨a¨, H.; Ma¨lko¨nen,
M.; Ma¨ntta¨ri, M.; Norola, S.; Pasternack, A.; Pikkarai-
nen, J.; Romo, M.; Sjo¨blom, T.; Nikkila¨, E. A. J. Am.
Med. Assoc. (JAMA) 1988, 260, 641.
7. Rubins, H. B.; Robins, S. J.; Collins, D.; Fye, C. L.;
Anderson, J. W.; Elam, M. B.; Faas, F. H.; Linares, E.;
Schaefer, E. J.; Schectman, G.; Wilt, T. J.; Wittes, J. N.
Engl. J. Med. 1999, 341, 410.
8. Picardo, M.; Massey, J. B.; Kuhn, D. E.; Gotto, A. M.;
Gianturco, S. H.; Pownall, H. J. Arteriosclerosis 1986, 6,
434.
In conclusion, 1-alkyl-3-phenylthioureas have been
shown to be effective HDL and Apo A-I-elevating
agents which also exhibit TG-lowering properties.
Some analogues produced larger, more buoyant
HDL particles. The optimal analogue 8d was effective
at raising HDL cholesterol in the CF and NF rat, CF
and NF hamster, and NF dog and CF monkey
models.
9. Miyazaki, A.; Rahim, A. T. M. A.; Ohta, T.; Morino, Y.;
Horiuchi, S. Biochim. Biophys. Acta 1992, 1126, 73.
10. Fielding, C. J.; Fielding, P. E. J. Lipid Res. 1995, 36, 211.
11. Rubin, E. M.; Krauss, R. M.; Spangler, E. A.; Verstuyft,
J. G.; Clift, S. M. Nature 1991, 353, 265.
References and notes
1. Gordon, D. J.; Rifkind, B. M. N. Engl. J. Med. 1989, 321,
1311.
´
12. Paszty, C.; Maeda, N.; Verstuyft, J.; Rubin, E. M. J. Clin.
Invest. 1994, 94, 899.
2. Gordon, D. J.; Probstfield, J. L.; Garrison, R. J.; Neaton,
J. D.; Castelli, W. P.; Knoke, J. D.; Jacobs, D. R.;
Bangdiwala, S.; Tyroler, H. A. Circulation 1989, 79, 8.
3. Barter, P. J.; Rye, K.-A. Atherosclerosis 1996, 121, 1.
13. Plump, A. S.; Scott, C. J.; Breslow, J. L. Proc. Natl. Acad.
Sci. U.S.A. 1994, 91, 9607.
14. Coppola, G. M.; Damon, R. E. European Patent Appli-
cation EP0528146, 1992.

G. M. Coppola et al. / Bioorg. Med. Chem. Lett. 16 (2006) 113–117
117
15. Coppola, G. M.; Damon, R. E. Canadian Patent
Application CA 2072704, 1993; Chem. Abstr., 119,
95134.
16. Coppola, G. M.; Damon, R. E.; Eskesen, J. B.; France, D.
S., Jr.; Paterniti, J. R., Jr. Bioorg. Med. Chem. Lett. 2002,
12, 2439.
17. Coppola, G. M.; Damon, R. E.; Eskesen, J. B.; France, D.
S., Jr.; Paterniti, J. R., Jr. Bioorg. Med. Chem. Lett. 2005,
15, 809.
18. Reductions in body weight gain are highly correlated to
food consumption and it cannot be ruled out that a
portion of the HDL effects may be attributable to the
diminished food consumption.
21. Clifton, P. M.; MacKinnon, A. M.; Barter, P. J.
J. Chromatogr. 1987, 414, 25.
22. France, D. S.; Quinby, R. E.; Babiak, J.; Lapen, D. C.;
Paterniti, J. R.; Weinstein, D. B. Lab. Rob. Autom. 1990,
2, 155.
23. France, D. S.; Weinstein, D. B.; Quinby, R. E.; Babiak, J.;
Lapen, D. C.; Murdoch, M. K.; Paterniti, J. R., Jr. Adv.
Lab. Autom. Rob. 1991, 7, 645.
24. Havel, R. J.; Eder, H. A.; Bragdon, J. H. J. Clin. Invest.
1955, 34, 1345.
25. Elokdah, H.; Sulkowski, T. S.; Abou-Gharbia, M.;
Butera, J. A.; Chai, S.; McFarlane, G. R.; McKean, M.;
Babiak, J. L.; Adelman, S. J.; Quinet, E. M. J. Med. Chem.
2004, 47, 681.
26. HDL cholesterol was determined by FPLC for the alcohol
4 and its enantiomer (7n and 7o in Ref. 17) and gave values
of 184% and 296% of controls, respectively.
19. Craig, J. C.; Pinder, A. R. J. Org. Chem. 1971, 36, 3648.
20. France, D. S.; Hughes, T. E.; Miserendino, R.; Spirito, J.
A.; Babiak, J.; Eskesen, J. B.; Tapparelli, C.; Paterniti, J.
R., Jr. J. Lipid Res. 1989, 30, 1997.