Tetrazole and ester substituted tetrahydoquinoxalines as potent cholesteryl ester transfer protein inhibitors

Article abstract of DOI:10.1016/j.bmcl.2007.01.112

Cholesteryl ester transfer protein is a plasma glycoprotein that transfers cholesterol ester between lipoprotein particles. Inhibition of this protein, in vitro and in vivo, produces an increase in plasma high density lipoprotein cholesterol (HDL-C). This

Full text of DOI:10.1016/j.bmcl.2007.01.112

Bioorganic & Medicinal Chemistry Letters 17 (2007) 2608–2613
Tetrazole and ester substituted tetrahydoquinoxalines
as potent cholesteryl ester transfer protein inhibitors
C. Todd Eary,* Zachary S. Jones, Robert D. Groneberg, Laurence E. Burgess,
David A. Mareska, Mark D. Drew, James F. Blake, Ellen R. Laird, Devan Balachari,
Michael O’Sullivan, Andrew Allen and Vivienne Marsh
Array BioPharma, 3200 Walnut Street Boulder, CO 80301, USA
Received 3 January 2007; revised 30 January 2007; accepted 31 January 2007
Available online 8 February 2007
Abstract—Cholesteryl ester transfer protein is a plasma glycoprotein that transfers cholesterol ester between lipoprotein particles.
Inhibition of this protein, in vitro and in vivo, produces an increase in plasma high density lipoprotein cholesterol (HDL-C). This
communication will describe the SAR and synthesis of a series of substituted tetrahydroquinoxaline CETP inhibitors from early l
lead to advanced enantiomerically pure analogs.
ꢀ 2007 Elsevier Ltd. All rights reserved.
Atherosclerosis is a systemic disease that consists of the
accumulation of lipid-rich plaques within the walls of
large arteries; it contributes to coronary heart disease,
stroke, and peripheral vascular disease. Atherosclerosis
is the leading cause of deaths in industrialized nations
and the standard for treatment is the administration of
statins to decrease low density lipoprotein cholesterol
levels (LDL-C). In addition to reducing the number of
coronary events through the reduction of LDL-C, stat-
ins have been shown to modestly increase (5–10%) high
density lipoprotein cholesterol (HDL-C).1
Cholesteryl ester transfer protein (CETP) is a plasma
glycoprotein that transfers cholesteryl ester between
lipoprotein particles with a balanced exchange of triglyc-
eride (TG). The net effect of CETP is to remove cho-
lesteryl ester from HDL and transfer it to LDL and
VLDL. Populations deficient in CETP do show in-
creased levels of HDL.5 Several research programs are
currently targeting the inhibition of CETP as a method
to raise HDL-C levels and several potent inhibitors have
been disclosed (Fig. 1).6–8 The most advanced programs
are Pfizer’s Torcetrapib and Roche/Japan Tobacco’s
JTT-705, these programs entered into phase III with sig-
nificant elevations of HDL-C observed in phase II clin-
ical trials.9
The anti-atherogenic role of HDL in reverse cholesterol
transport is now well established and the elevation of
HDL has become a therapeutic target. Studies have
shown an inverse relationship between HDL and cardio-
vascular disease, even a modest 1 mg/dL elevation of
HDL may reduce the risk of cardiovascular disease by
2–3 percent.2 At present, niacin and Niaspan, extended
release niacin, are the most effective therapeutic agents
to elevate HDL-C.3 These treatments can elevate
HDL-C from 15% to 35% however; niacin can produce
undesirable side effects such as flushing and
hepatotoxicity.4
We recently became interested in pursuing a research
program on CETP inhibition and, following an evalua-
tion of published work, we chose a substituted tetrahy-
droquinoxaline as a potentially novel lead structure
(Fig. 2). This type of core would provide multiple points
for further synthetic diversification as shown.
The substituted tetrahydroquinoxaline cores were pre-
pared in a three-step sequence beginning with SNAr
reactions on commercially available ortho nitro aryl flu-
orides (Scheme 1). The appended alcohols 1 and 2 were
converted to mesylates under standard conditions. The
tetrahydroquinoxalines 5 and 6 were formed through
reduction of the nitro group followed by intramolecular
SN2 displacement of the mesylate group. Compound 7
Keywords: Cholesteryl ester transfer protein; CETP; HDL; Tetrahydo-
quinoxalines; Cholesterol.
*
Corresponding author. Tel.: +1 303 386 1101; e-mail: teary@
arraybiopharma.com
0960-894X/$ - see front matter ꢀ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.01.112

C. T. Eary et al. / Bioorg. Med. Chem. Lett. 17 (2007) 2608–2613
2609
CF3
CF3
O
O
CF3
O
N
N
F3C
O
NH
CF3
N
S
F3C
N
O
O
O
O
O
Roche & Japan Tobacco
JTT-705 (covalent)
Lilly Example 119
42 nM (Buffer)
Pfizer-Torcetrapib
52 nM (Plasma)
μ
6.0 M( Plasma)
Figure 1. IC50 values for cholesteryl ester transfer protein inhibitors.
under SN2 conditions to provide 13.12 Compound 13
underwent electrophilic aromatic halogenation to give
1413 and 15. The remaining unsubstituted anilines were
capped with the ethyl carbamate functionality to give
analogs 16–18.
O
R2
R1
n
R4
R5
N
N
R3
O
O
Since the SN2 approach utilized to prepare 13 was
unsuccessful for the 6-trifluoromethyl series an alternate
route was undertaken (Scheme 3). Selective reductive
amination with 3,5-bis(trifluoromethyl)benzaldehyde
followed by acylation with ethyl chloroformate gave
compound 21. Selective benzylic lithiation and subse-
quent trapping with methyl chloroformate produced
the desired carbon–carbon bond in 22.
Figure 2. Substituted tetrahydroquinoxalines as CETP inhibitors.
was prepared from selective reductive amination of the
less hindered nitrogen with ethyl oxoacetate and conver-
sion to the methyl ester. The remaining aniline was
capped with an ethyl carbamate group followed by sub-
sequent enolate alkylation to give 9 and 10.10
The preliminary SAR of these series (Table 1) from a
human plasma CETP assay14 indicated that direct
attachment of the aryl ring to the N2 methine was opti-
mal to inhibit CETP. Additionally it was found that
deviation away from an ester to methyl, dimethyl, and
The chain-shortened analogs (n = 0) were prepared from
benzylic bromination of methyl 3,5-bis(trifluorometh-
yl)phenyl acetate (11) (Scheme 2).11 The resulting bro-
mides were treated with tetrahydroquinoxaline core 5
O
O
R2
H3CO
H3CO
R4
NO2
NH
H
N
R4
R4
NO2
F
(+/-)
R4
R4
N
N
N
c
h
a
d, e, f
R3
N
H
N
R
OR
O
O
1 R, R4 = H
R = H, R = CF3
5 R4 = H
2
7 R, R4 = H
4
6 R4 = CF3
9
R4 = H
10 R2, R3 = CF3
b
g
R = Ms R4 = H
3
R = CO2Et R4 = H
4 R = Ms, R4 = CF3
8
Scheme 1. Reagents and conditions: (a) 2-aminobutan-1-ol, K2CO3, DMF rt to 80 ꢁC (75–90%); (b) MsCl, py, DCM (50–70%); (c) H2, Pd/C, NMP,
then TBAI, 80 ꢁC (50–60%); (d) ethyl 2-oxoacetate, NaCNBH3, MeOH, D (50–80%); (e) LiOH, MeOH, 70–85%; (f) TMSCHN2, THF/MeOH (90%);
(g) ClCO2Et, py, DCM, (76%); (h) LDA, HMPA, THF, 3,5-bis CF3 benzyl bromide or benzyl bromide, À78 ꢁC (50–60%).
CF3
CF3
CF3
CF3
O
O
O
O
O
C
b
O
CF3
c
O
CF3
d
F3
R4
R5
N
N
R4
R5
N
N
O
a
CF3
R
N
H
N
H
11
R = H
12 R = Br
O
O
13
14R4 = Br, R5 = H
16 R4 = H, R5 = H
17 R4 = Br, R5 = H
18
4 = Cl, R5 = Cl
15
R
R
4 = Cl, R5 = Cl
Scheme 2. Reagents and condition: (a) NBS, AIBN, CCl4, D (50%); (b) 5, TBAI, K2CO3, DMF (40–60%); (c) 15, NBS, DMF or NCS, DMF
(5–22%); (d) ClCO2Et, py, DCM (44–65%).

2610
C. T. Eary et al. / Bioorg. Med. Chem. Lett. 17 (2007) 2608–2613
CF3
CF3
O
H
CF3
O
CF3
F3C
N
(+/-)
a
c
F3C
N
F3C
N
N
(+/-)
N
H
N
R
6
O
O
20 R = H
b
22
21
R = CO2Et
Scheme 3. Reagents and conditions: (a) 3,5-bis CF3 benzaldehyde, NaCNBH3, MeOH rt to 80 ꢁC (50%); (b) ClCO2Et, Cs2CO3, DCM (50%); (c) sec-
BuLi, HMPA, THF, ClCO2Me, À78 ꢁC (10–35%).
Table 1. Activity of preliminary analogs in the plasma CETP assay
O
R2
R1
n
R4
R5
N
N
R3
(+/-)
O
O
Compound
R1
R2
R3
R4
R5
n
IC50 (lM)
D1a
D2a
9a,9b
OCH3
OCH3
OCH3
OCH3
OCH3
OCH3
H
H
H
H
H
H
H
Cl
H
1
1
0
0
0
0
>100
>100
5.0
1.9b
92.5
0.5
>100
>100
8.6
10a,10b
16a,16b
17
CF3
CF3
CF3
CF3
CF3
CF3
CF3
CF3
CF3
CF3
H
H
Br
Cl
CF3
18a,18b
22a,22b
93.4
0.56
a The D1 and D2 nomenclature denotes the first and second eluting diastereomers undergoing chromatography. This convention is used throughout
this work.
b Diastereomers were not separated (ꢀ1:1 mixture).
CF3
CF3
CF3
R
N
CF3
N
N
N
N
CN
c
N
b
F3C
F3C
e, f
CN
N
N
F3C
F3C
F3C
F3C
F3C
a
N
(+, -)
R
N
N
N
23
R = H
24 R = Br
N
H
N
H
25
26
27
O
O
R = H
d
R = Me
28a(D1) & 28b(D2)
Scheme 4. Reagents and condition: (a) NBS, AIBN, CCl4, reflux (25–38%); (b) racemic or R or S 6, DIEA, DMP, argon (74%); (c) NaN3, NH4Cl,
DMF (65%); (d) TMSCHN2, THF/MeOH (80–95%); (e) SiO2; (f) EtCO2Cl, py, DCM (40–70%).
CF3
Methyl Tetrazole
Methyl Ester
CF3
N
N
N
O
N
F3C
F3C
O
F3C
F3C
(+/-)
N
N
(+/-)
N
N
O
O
O
O
22a(D1)
500 nM
22b(D2)
560 nM
28a(D1)
870 nM
28b(D2)
230 nM
Figure 3. Separated methyl tetrazole diastereomers differ in potency.

C. T. Eary et al. / Bioorg. Med. Chem. Lett. 17 (2007) 2608–2613
2611
hydroxyethyl amides resulted in a complete loss of activ-
ity (data not shown). The introduction of aryl substitu-
ents (R4 and R5) increased potency, except for chloro
disubstitution. The R4 trifluoromethyl group was opti-
mal with both diastereomers providing an IC50 of about
500 nM, a 10-fold increase in potency over compounds
16a, 16b.
place of the methyl ester.15 The tetrazole group was pre-
pared in a five-step sequence which began with the rad-
ical bromination of 3,5-bis-trifluoromethyl benzonitrile
(23) (Scheme 4). Alkylation of the more reactive aniline
with 24 produced the desired derivative for cycloaddi-
tion with sodium azide to the free tetrazole. Subsequent
selective16 N2 methylation with TMS-diazomethane
produced diastereomers that could be separated via sili-
ca gel chromatography.17 Finally, the free aniline was
capped as the ethyl carbamate group. The methyl tetra-
zole substitution did produce a boost in potency over
the ester series for one of the diastereomers, D2. The
second eluting diastereomer 28b (D2) was found to be
To suppress potential metabolic liabilities we became
interested in substituting a methyl tetrazole group in
Table 2. Activity of Nl derivatives in the plasma CETP assay
CF3
N
N
N
N
F3C
F3C
Table 3. Activity of tetrazole analogs in the plasma CETP assay
(+, -)
N
N
CF3
N
N
N
N
F3C
F3C
Compound
Methoda
A
IC50 (nM)
D2
N
N
D1
(+, -)
28a,28b
870
1110
230
O
O
O
O
29
A
A
O
O
Compound
IC50 (nM)
D2
D1
870
30
4400
232
O
O
28a,28b
230
31
A
A
B
B
B
B
B
C
C
C
O
O
O
O
O
43a
32,500
4420
O
O
32
62,590
O
O
44a
OH
33a,33b
34
100,000
4570
O
O
N
H
O
45a
14,580
46,380
N
H
OH
46a,46bb
50a,50ba
51a
2265
5220
2225
35
100,000
100,000
36,900
854
O
O
O
N
N
N
O
100,000
H2N
N
36
7815
54,570
3420
37
O
52a
O
HO
NC
38
53a
39a,39b
40
1300
5700
54a
4555
a Analog prepared by converting the aniline of compound 27 to the
ethyl carbamate (EtCO2Cl, pyridine, DCM) followed by tetrazole
formation (N3, NH4Cl, DMF, D). This material was then alkylated16
with the corresponding alkyl halide utilizing DIEA or NaH as base.
b This analog was prepared through the reduction of compound 44
with LiBH4.
O
2915
OH
a Prepared from compound 26 utilizing method (A) RCO2Cl, py,
DCM; (B) Cl2CO, DIEA, DCM then ROH or RNH (R or H); (C)
RCHO, NaCNBH3, MeOH, D.

2612
C. T. Eary et al. / Bioorg. Med. Chem. Lett. 17 (2007) 2608–2613
CF3
CF3
CF3
CF3
N
N
N
N
N
N
N
N
N
N
N
N
Rat PK
N
N
N
N
F3C
F3C
F3C
F3C
F3C
F3C
F3C
F3C
IV 2 mg/kg (EtOH, cremo/saline)
Cl = 13 mL/min/Kg
N
N
N
N
N
N
N
N
Vd = 2600 mL/Kg
(
R
)
(
S
)
(R
)
(S)
T
1/2 = 2.4
Physical Properties
Solubility pH 7.4 < 10 ng/mL
(amorphous)
O
O
O
O
O
O
O
O
55a (D1)
55c (D2)
55d (D2)
55b (D1)
440 nM
1500 nM
143 nM
47500 nM
The D2 R-diastereomer
is the most potent!
Figure 4. Preparation of all four stereoisomers of analog 28 reveals a preferred enantiomer 55d.
ꢀ4-fold more potent as compared to the first eluting
diasteromer 28a (D1) (Fig. 3).
cally active synthesis. Preparation of all four stereoiso-
mers of 28 (55a–d)18 revealed that the C2-R-second
eluting diastereomer (55d) was the most potent at
143 nM (Fig. 4). The stereochemistry at the benzylic po-
sition was not assigned. Analog 55d was found to have
low clearance following iv dosing in rats, however, the
solubility of 55d was extremely poor at <10 ng/mL.
Since the diastereomerically pure tetrazole analog was
more potent than the corresponding methyl ester and of-
fered esterase stability we conducted a more thorough
exploration of the SAR. Our strategy was to explore
the N1 position of the tetrahydroquinoxaline core as
well as substitution of the tetrazole ring. The synthetic
routes were amenable to these substitutions since both
sites are elaborated late in the synthesis.
In summary, we have utilized a tetrahydroquinoxaline
core to produce potent CETP inhibitors. A key obser-
vation in this process was that a shortened headpiece
is optimal for the tetrahydroquinoxaline core. The
replacement of the methyl ester with the methyl tetra-
zole moiety was also an important advance for the
program. The methyl tetrazole offered improved
potency and potential metabolic stability. Efforts to
improve on tetrazole substitution and N1 aniline sub-
stitution were not successful in producing superior
compounds to 28. The synthesis of all four stereoiso-
mers 55a–d revealed that compound 55d was by far
the most potent with an IC50 = 143 nM in a human
plasma CETP assay. This compound, 55d, displays a
rat PK profile with low clearance, moderate volume
of distribution, and low solubility.
In addition to the ethyl carbamate moiety at N1, several
other carbamate analogs were prepared and the com-
pounds were tested as single diastereomers or ꢀ1:1 mix-
tures of diastereomers (Table 2). The most potent of the
analogs was the isobutyl carbamate 31 with an IC50 of
232 nM as a D1/D2 mixture. This analog was much pre-
ferred over the benzyl, butyl, and 2-ethyl hexyl analogs.
To determine the effect of replacing the carbamate oxy-
gen with nitrogen a series of ureas were prepared. Both
mono- and disubstituted ureas produced a loss in poten-
cy. The D2 diastereomer 33b of the hydroxy ethyl urea
was much more potent than the D1 diastereomer
(33a). Interestingly the direct nitrogen analog of 28
(compound 34) was >40· less potent. We also investi-
gated direct alkyl substituted analogs with the cyclopen-
tyl methyl diastereomers 38 being the most potent at
854 nM. Increasing the bulk of the appended cycloal-
kane produced a negative effect on potency since ana-
logs 39a, 39b, and 40 were in the single digit lM range.
Acknowledgment
The authors thank T.K. Pope and Gregory Poch for
performing the PK analysis and Michelle Livingston
for the solubility determination of 55d.
The SAR of the tetrazole group was also investigated
(Table 3). Replacing the methyl group with methyl
and tert-butyl acetoxy groups 43 and 45 produced a
large decrease in potency. The homologated methyl
acetoxy group (44) retained some potency at 4.4 lM.
Substitution with the hydroxy ethyl group produced
analogs 46a and 46b, which displayed modest potency.
The amide and acid functionalities 50 and 52 produced
a large loss of potency, while the basic amine 51, nitrile
53, and cyclopropyl group 54 displayed moderate activ-
ities. None of the analogs were superior to the 2-methyl
tetrazole moiety in analog 28.
References and notes
1. Vrecer, M.; Turk, S.; Drinovec, J.; Mrhar, A. Int. J. Clin.
Pharmacol. Ther. 2003, 41, 567.
2. Brewer, H. B., Jr. N. Engl. J. Med. 2004, 350, 1491.
3. Nicholls, S. J.; Nissen, S. E. Eur. Heart J. 2005, 26, 853.
4. Ito, M. K. Am. J. Manag. Care 2002, 8, S315.
5. Inazu, A.; Brown, M. L.; Hesler, C. B.; Agellon, L. B.;
Koizumi, J.; Takata, K.; Maruhama, Y.; Mabuchi, H.;
Tall, A. R. N. Engl. J. Med. 1990, 323, 1234.
6. Clark, R. W.; Sutfin, T. A.; Ruggeri, R. B.; Willauer, A.
T.; Sugarman, E. D.; Magnus-Aryitey, G.; Cosgrove, P.
G.; Sand, T. M.; Wester, R. T.; Williams, J. A.; Perlman,
M. E.; Bamberger, M. J. Arterioscler. Thromb. Vasc. Biol.
2004, 24, 490.
The discovery that one of the tetrazole diastereomers
was much more potent prompted us to pursue an opti-

C. T. Eary et al. / Bioorg. Med. Chem. Lett. 17 (2007) 2608–2613
2613
7. Shinkai, H.; Maeda, K.; Yamasaki, T.; Okamoto, H.;
Uchida, I. J. Med. Chem. 2000, 43, 3566.
DMSO concentration 4% (v/v) are incubated on a 96-
well round-bottom polypropylene plate for 6 h at 37 ꢁC.
The reaction is stopped by adding 10 U of heparin in
0.3 M MnCl2 (final concentration) to precipitate the
LDL. Plates are centrifuged at 1000g (2500 cpm, GH 3.8
rotor, Beckman Allegra 6 centrifuge). Volume of super-
natant equal to 2/3 of assay volume is removed for
8. Cao, G.; Escribano, A. M.; Fernandez, M. C.; Fields, T.
Gernert, D. L.; Cioffi, C. L.; Herr, R. J.; Mantlo, N. B.;
Martin De La Nava, E. M.; Mateo, Herranz, A. I.;
Mayhugh, D. R.; Wang, X., WO05037796, 2005.
9. Sikorski, J. A. J. Med. Chem. 2006, 49, 1.
3
10. Gray, B. D.; Jeffs, P. W. J. Chem. Soc., Chem. Commun.
1987, 1329.
counting in a Topcount to quantitate amount of H-CE-
HDL remaining.
11. Prepared from the esterification of commercially available
3,5-bis(trifluoromethyl) phenylacetic acid.
15. Herr, R. J. Bioorg. Med. Chem. 2002, 10, 3379.
16. Alkylation of the tetrazole moiety produced a general
trend with the N2 isomer as major accompanied with
minor amounts of the N1 isomer (5–15%). The regio-
chemistry of compound 27 was determined through
gHMBC NMR.
17. Still, W. C.; Kahn, M.; Mitra, A. J. Org. Chem. 1978, 43,
2923.
18. These analogs were prepared as described in Scheme 4
utilizing optically active tetrahydroquinoxaline 6. The
R and S tetrahydroquinoxaline cores were prepared
12. Aliphatic regio-substitution at N-1 versus N-4 was deter-
mined via NMR 1-D (one dimensional) NOESY (nuclear
Overhauser enhancement spectroscopy) via soft shaped
pulse excitation or 2-D gHMBC (gradient Heteronuclear
Multiple Bond Correlation).
13. The structural assignment for 14 was based upon gHMBC
NMR.
14. Plasma CETP (Cholesteryl Ester Transfer Protein) activ-
ity is determined with a radioactive assay that measures
the transfer of 3H-CE from HDL to LDL. Inhibitor,
human plasma (which contributes CETP and LDL),
3H-CE-HDL (diluted to 50 lg/mL cholesterol), and final
as described in Scheme
1
utilizing commercially
available (R)-(À)-2-amino-1-butanol and (S)-(+)-2-ami-
no-1-butanol.