Fluorine-substitution in cholesteryl ester transfer protein inhibitors (CETP-inhibitors) - Biology, chemistry, SAR, and properties

Article abstract of DOI:10.2533/000942904777678082

The inhibition of the cholesteryl ester transfer protein (CETP) provides a method for the elevation of the high density lipoprotein cholesterol (HDL-C) level, i.e. the 'good' cholesterol. The expected anti-atherogenic effect of this approach is independen

Full text of DOI:10.2533/000942904777678082

FLUORINE IN THE LIFE SCIENCE INDUSTRY
123
CHIMIA 2004, 58, No. 3
Chimia 58 (2004) 123–127
© Schweizerische Chemische Gesellschaft
ISSN 0009–4293
Fluorine-Substitution in Cholesteryl
Ester Transfer Protein Inhibitors (CETP-
Inhibitors) – Biology, Chemistry, SAR,
and Properties
Holger Paulsena*, Carsten Schmecka, Arndt Brandesa, Gunter Schmidta, Jürgen Stoltefussa,
Stephan N. Wirtza, Michael Lögersa, Paul Naaba, Klaus-Dieter Bremmb, Hilmar Bischoffb, Delf
Schmidtb, Siegfried Zaissb, and Stefan Antonsc
Abstract: The inhibition of the cholesteryl ester transfer protein (CETP) provides a method for the elevation of the
high density lipoprotein cholesterol (HDL-C) level, i.e. the ‘good’ cholesterol. The expected anti-atherogenic effect
of this approach is independent of the proven benefits of lowering the low density lipoprotein cholesterol (LDL-C)
level, i.e. the ‘bad’ cholesterol. A medicinal chemistry project is presented starting from the first screening hit to the
second generation development candidate. The structure–activity relationship, the syntheses and the role of fluo-
rine during optimization, as well as the biological activities are discussed.
Keywords: Cholesteryl ester transfer protein · Domino reaction · Hantzsch-reaction · HDL-C elevation ·
Oxazaborolidine reduction
1. Introduction
modifiable risk factors for atherosclerosis ample cell walls of coronary arteries, to-
Coronary heart disease (CHD) is the and eventually CHD. Non-modifiable fac- wards the liver. This removal of cholesterol
leading cause of death in the industrialized tors include age, gender, and family or per- from a cell membrane starts with an esteri-
countries. Angina pectoris and myocardial sonal history of CHD. Examples of modifi- fication by action of lecithin cholesterol
infarction are clinical manifestations of able factors are dyslipidemia, high blood acyl transferase (LCAT). Typical choles-
CHD. They are caused by a stenosis or an pressure, obesity, diabetes mellitus, dietary teryl esters are fatty acid esters, e.g. linole-
occlusion of a coronary artery through ath- factors, lack of exercise, thrombogenic fac- ic acid. These very lipophilic compounds
erosclerosis.
tors, excess alcohol consumption, and are carried through the aqueous blood-
Atherosclerosis is the development of smoking. Lipid disorders are key risk fac- stream in the HDL-particles. In the liver
fibro-fatty plaques within the inner walls of tors for atherosclerosis. It is known that cholesterol is stored or metabolized into
arteries, resulting in a narrowing of the ves- high LDL-C (low density lipoprotein cho- bile acids but it is also a site of biosynthesis
sel. There are several non-modifiable and lesterol) levels and low HDL-C (high den- of cholesterol.
sity lipo-protein cholesterol) levels are in-
An increase of the HDL-C level can be
dependent risk factors for the development achieved by inhibition of the cholesteryl es-
of atherosclerosis [1]. Excess cholesterol in ter transfer protein (CETP)[3]. This protein
the cell walls of coronary arteries is in- mediates the exchange of cholesteryl esters
volved in oxidative events which trigger the from HDL to LDL (Fig. 1). The blockade
invasion of macrophages. The macrophages therefore increases the efficiency of the re-
can degenerate and transform themselves verse cholesterol transport.
into foam-cells by intake of too much ox-
LDL-C. As a consequence atherosclerotic
*Correspondence: Dr. H. Paulsena
Tel.: +49 (0)202 368650
Fax: +49 (0)202 364061
E-Mail: holger.paulsen.hp@bayer-ag.de
aBayer AG
plaques are formed. They consist of prolif- 2. Lead Candidate
erated smooth muscle cells, fibrous tissue
and the lipid accumulation. Lowering LDL-
A medicinal chemistry project aiming at
C by blocking cholesterol biosynthesis with the discovery and optimization of CETP in-
so-called ‘statins’ is well-established med- hibitors was started. The screening for ac-
ical practice. Elevation of HDL-C levels of- tive compounds was performed using a
fers a new chance for a promising therapy scintillation proximity assay: HDL loaded
[2], because it is an important player in the with radioactively labelled cholesteryl ester
reverse cholesterol transport, i.e. the trans- is exposed to CETP and tagged LDL. The
port of cholesterol from periphery, for ex- tagged LDL is scavenged by a fluoro-bead
BHC-PH-R-EU-Chemical Research
D-42096 Wuppertal
bBayer AG
BHC-PH-R-EU-Cardiovascular Research
D-42096 Wuppertal
cBayer AG
BHC-PH-OP-ELB-Chemical Development
D-42096 Wuppertal

FLUORINE IN THE LIFE SCIENCE INDUSTRY
124
CHIMIA 2004, 58, No. 3
exchange of the p-fluoro-phenyl substituent
against a p-trifluoromethyl phenyl sub-
stituent accompanied by a shortening of the
spacer (3) resulted in another increase of
potency. The introduction of fluorine at the
bis-benzylic bridge (rac-4) led to a further
improvement and the exchange of the iso-
propyl versus a cyclopentyl moiety pro-
duced a highly potent CETP inhibitor. In its
enantiomerically pure form lead candidate
5 already has an inhibitory activity at CETP
of IC50= 13nM.
Fig. 1
3. First Generation CETP-Inhibitors
However, the metabolic stability of this
lead compound was insufficient. During the
later phase of the project the optimization
of this and other parameters was addressed
in order to meet the requirements for a de-
velopment candidate. The metabolism of
the primary alcohol was tackled by rigidifi-
cation in a bicyclic structure (Fig. 3). Al-
though alkyl replacement of the right hand
p-F-phenyl substituent (rac-6) led to a com-
plete loss of activity the rigidification of the
alcohol group through cyclization gave a
promising result, albeit at the expense of in-
troducing another stereogenic center: The
anti-diastereomer displays an IC50 of 100
nM as a racemate (rac-7). Based on this
new bicyclic core structure an intensive
structure–activity screening by chemical
derivatization was performed (Fig. 4). The
para-fluoro substitution at the phenyl-sub-
stituent, commonly introduced to prevent
metabolic para-hydroxylation, improves
potency by a factor of three. The fluorine
substituent in the bis-benzylic bridge leads
to the most potent inhibitors as shown in
Fig. 4. Another unique substituent was the
trifluoromethyl group in p-position of the
phenyl substituent. Probably due to optimal
size and lipophilicity no other substituent
gave a better result regarding potency. A
breakthrough was achieved by methyl-sub-
stitution at the cyclohexyl moiety in 1,3 po-
sition to the hydroxy group (Fig. 4).
To avoid another stereogenic center a
geminal dimethyl compound was prepared
and proved to be equally active. The enan-
tiomerically pure compound 8 was chosen
as the first development candidate for
CETP-inhibition (Fig. 5). The configura-
tion of the alcoholic carbon center is crucial
for high potency as isomer 9 demonstrates.
The inversion of the fluorine bearing center
as in compound ent-9 leads to a more than
twofold decrease in activity.
Fig. 2
via the biotin-streptavidin system. If a successful optimization delivered a highly
transfer of cholesteryl ester to the LDL par- potent lead compound 5. Fluorine substi-
ticle has taken place the radioactivity is lo- tution played a key role in this process
cated in the vicinity of the fluoro-bead. This (Fig. 2).
In order to provide sufficient quantities
of material a technically feasible synthesis
was established for candidate 8 (Scheme 1).
After preparation of enamine 10 the reac-
tion sequence starts with an unsymmetrical
Hantzsch-dihydropyridine synthesis of
compound 11, followed by an oxidation to
results in a light-signal which is detected in
A
significant improvement was
a scintillation counter. The first screening achieved by introduction of a second p-flu-
hit 1 displayed a potency of IC50 = 15 000 oro-phenyl-substituent instead of an iso-
nM. During the early phase of the project a propyl group leading to compound 2. The

FLUORINE IN THE LIFE SCIENCE INDUSTRY
125
CHIMIA 2004, 58, No. 3
the pyridine 12. Diketone 12 undergoes a
completely regioselective CBS-type reduc-
tion [4] which in addition is highly enan-
tioselective (95% ee). After TBS-protection
of the resultant hydroxy function (13), the
remaining keto group is reduced diastereo-
selectively in favour of the desired diaste-
reomer 14 (75% ds) with RedAl®. The de-
sired diastereomerically and enantiomeri-
cally pure compound 14 is isolated by
crystallization in 60% yield. The mixture of
diastereomers remaining in the mother
liquor can be oxidatively recycled to the
silyloxy ketone 13. The mechanistic basis
for the diastereoselectivity remains to be
clarified. The inducing stereogenic center
in this fairly rigid system is located at sig-
nificant distance to the newly generated
center. The role of the p-F-phenyl-sub-
stituent in the conveyance of stereo-infor-
mation has to be elucidated. The direct and
stereoselective replacement of the un-
protected hydroxyl group was achieved by
Fig. 3.
using
diethylaminosulfur
trifluoride
(DAST) or the Ishikawa reagent (diethyl-
1,1,2,3,3,3-hexa-fluoropropylamine) [5]
as fluorinating reagent. In contrast to the
usual SN2 pathway of DAST-reactions, at
this sterically very encumbered site, the
reaction proceeds with retention of config-
uration [6]. This was proven by X-ray
structures of candidate 8 and a close ana-
log of compound 14. The mechanism for
this reaction is under investigation and
will be published subsequently. Final
TBAF or acidic silyl-deprotection provid-
ed the desired compound 8 in high overall
yield.
Fig. 4.
With this compound in hand, a signifi-
cant in vivo activity was demonstrated after
oral administration. An increase in HDL-C
by 35% and 50% was obtained in trans-
genic human-CETP mice after three days of
treatment with 5 mg/kg p.o. and 10 mg/kg
p.o., respectively. The anti-atherogenic ef-
fect was proven in New Zealand white rab-
bits under high fat diet. After three months
of treatment with compound 8 in food, a
daily dose of 50 mg/kg p.o. reduced the ath-
erosclerotic plaque area by 40%, a daily
dose of 150 mg/kg p.o. reduced the plaque
area by 70%. These results pave the way for
a new approach towards treatment of ather-
osclerosis.
4. Second Generation
CETP-Inhibitors
During the optimization program, it was
revealed that carba-analogs of the pyridine
compounds were even more potent CETP-
inhibitors. A direct comparison of both
classes is shown in Fig. 6. The second gen-
eration candidate 17 was therefore selected
from this structural class of CETP-in-
hibitors.
Fig. 5.

FLUORINE IN THE LIFE SCIENCE INDUSTRY
126
CHIMIA 2004, 58, No. 3
The seemingly small chemical change
in the core structure resulted in a consider-
able synthetic challenge. Several pathways
were examined including Diels-Alder ap-
proaches, a carba-Hantzsch [7] pathway,
and other strategies. But most successful
was a new domino cyclization process [8]
(Scheme 2). The introduction of the stereo-
genic centers was expected to be achieved
analogous to the pyridine case. Therefore,
retrosynthesis focused on the correspond-
ing 1,5-diketone 25.
Since the Mukaiyama-Michael addition
is recognized as a very reliable method for
the synthesis of 1,5-diketones, the follow-
ing domino reaction was established for the
one-step construction of the complete core
structure: the enolate 20, formed in the
Mukaiyama-Michael addition of silyl enol
ether 18 to cyclohexenone 19, is submitted
in situ to an intermolecular Michael addi-
tion to chalcone 21. The newly generated
enolate finally undergoes an intramolecular
aldol addition to form a six-membered ring
(22), thus generating the carbocyclic core in
a one-pot procedure from readily accessible
starting materials. The cyclohexanol 22 is
isolated readily in good yield via crystal-
lization as a single diastereomer (all C-sub-
stituents equatorial). All stereogenic cen-
ters are destroyed in the following aroma-
tization. This is performed in a double
halogenation–elimination sequence with
the cyclohexadiene intermediate being oxi-
dized in situ during the second elimination.
The establishment of the stereogenic cen-
ters proceeds in analogy to the sequence de-
veloped for pyridine 12, with minor modi-
fications. This synthesis is performed in a
kg-scale and is carried out without chro-
matographic purification. The final product
17 is isolated as a pure diastereomer and
enantiomer (>99% ds, >99% ee after crys-
tallization).
Scheme 1.
Fig. 6.

FLUORINE IN THE LIFE SCIENCE INDUSTRY
127
CHIMIA 2004, 58, No. 3
Scheme 2.
Acknowledgements
[5] A. Takaoka, H. Iwakiri, N. Ishikawa,
Bull. Chem. Soc. Jpn. 1979, 52,
3377–3380.
Significant HDL-C increase in hamsters
and hCETP mice at doses of <1 mg/kg p.o.
are achieved with this compound. A seven
day treatment with 0.6 mg/kg of compound
17 administered in the food resulted in a
57% increase in HDL-C in human-CETP
mice, proving the impressive in vivo poten-
cy of the second generation candidate.
We thank Dr. Axel Göhrt and Dr. J. Benet-
Buchholz for the X-ray analyses and the col-
leagues of PH-R-Strukturchemie for providing
the analytical data.
[6] Y. Mori, N. Morishima, Bull. Chem. Soc.
Jpn. 1994, 67, 236–241; S. Castillon, A.
Dessinges, R. Faghih, G. Lukacs, A.
Olesker, T.T. Thang, J. Org. Chem. 1985,
50, 4913–4917.
Received: December 18, 2003
[7] An example of a carba-Hantzsch reaction
is given in: K. Urbahns, D. Kieck, A.
Goehrt, Synthesis 1998, 1807–1811.
[8] H. Paulsen, S. Antons, A. Brandes, M.
Lögers, S.N. Müller, P. Naab, C. Schmeck,
S. Schneider, J. Stoltefuss, Angew. Chem.
Int. Ed. 1999, 38, 3373–3375.
[1] G. Assmann, H. Schulte, A. von Eckard-
stein,Y. Huang, Atherosclerosis 1996, 124
(Suppl), 11–20; CAPLUS 1996: 395970.
[2] G. Assmann, J.-R. Nofer, Annu. Rev. Med.
2003, 54, 321–341.
5. Conclusion
In summary, CETP inhibition is a prom-
ising principle for the elevation of HDL-C
and treatment of atherosclerosis. Starting
from weak hits, extremely potent CETP in-
hibitors were discovered with the use of flu-
orine substitution. Significant synthetic
challenges were overcome by newly invent-
ed synthetic methods.
[3] A.R. Tall, J. Lipid Res. 1993, 34,
1255–1274.
[4] Use of aminoindanol derived oxazaboro-
lidines: a) B. Di Simone, D. Savoia, E.
Tagliavini, A. Umani-Ronchi, Tetrahe-
dron: Asymmetry 1995, 6, 301–306;
b) Review: A. K. Ghosh, S. Fidanze, C. H.
Senanayake, Synthesis 1998, 937–961.