Discovery of BI 99179, a potent and selective inhibitor of type i fatty acid synthase with central exposure

Article abstract of DOI:10.1016/j.bmcl.2011.07.083

Based on a high-throughput screen, cyclopentanecarboxanilides were identified as a new chemotype of non-covalent inhibitors of type I fatty acid synthase (FAS). Starting from initial hits we aimed at generating a tool compound suitable for the in vivo validation of FAS as a therapeutic target. Optimisation yielded BI 99179 which is characterised by high potency, remarkably high selectivity and significant exposure (both peripheral and central) upon oral administration in rats.

Full text of DOI:10.1016/j.bmcl.2011.07.083

Bioorganic & Medicinal Chemistry Letters 21 (2011) 5924–5927
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery of BI 99179, a potent and selective inhibitor of type I fatty acid
synthase with central exposure
Jörg T. Kley a, , Jürgen Mack a, Bradford Hamilton b, Stefan Scheuerer c, Norbert Redemann b
⇑
a Medicinal Chemistry, Boehringer Ingelheim Pharma GmbH & Co. KG, Birkendorfer Strasse 65, D-88397 Biberach, Germany
b CardioMetabolic Diseases Research, Boehringer Ingelheim Pharma GmbH & Co. KG, Birkendorfer Strasse 65, D-88397 Biberach, Germany
c Drug Discovery Support, Boehringer Ingelheim Pharma GmbH & Co. KG, Birkendorfer Strasse 65, D-88397 Biberach, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
Based on a high-throughput screen, cyclopentanecarboxanilides were identified as a new chemotype of
non-covalent inhibitors of type I fatty acid synthase (FAS). Starting from initial hits we aimed at gener-
ating a tool compound suitable for the in vivo validation of FAS as a therapeutic target. Optimisation
yielded BI 99179 which is characterised by high potency, remarkably high selectivity and significant
exposure (both peripheral and central) upon oral administration in rats.
Received 3 June 2011
Revised 21 July 2011
Accepted 22 July 2011
Available online 9 August 2011
Ó 2011 Elsevier Ltd. All rights reserved.
Keywords:
Fatty acid synthase (FAS)
Enzyme inhibitor
Obesity
Acne
CNS exposure
Mammalian type I fatty acid synthase (FAS) is a key enzyme for
lipogenesis and highly expressed in lipogenic tissues. While most
tissues (except liver and adipose tissue) have low levels of FAS
expression and activity, FAS is over expressed in many cancers
and has been suggested as a potential anticancer target.1 More re-
cently, accumulating evidence points towards FAS inhibition as a
potential antiviral principle.2 FAS is also expressed in the brain
and targeted deletion of FAS in hypothalamic neurons produces
hypophagic, lean mice.3 Furthermore central nervous system
administration of the non-specific irreversible FAS inhibitors Ceru-
lenin and C75 in rodents decreases brain levels of orexigenic neu-
ropeptides and produces hypophagia and weight loss, suggesting
central FAS inhibition as a potential way to treat obesity.4 More-
over, FAS is highly expressed in human sebocytes, the lipid produc-
ing cells of the sebaceous glands.5 The pathogenesis of acne, the
most common disorder involving the sebaceous gland, involves li-
pid overproduction by the sebaceous gland,6 and it has been re-
ported that inhibitors of FAS reduce the production of sebum in
sebocytes,7 suggesting topical FAS inhibition as a potential anti-
acne approach.
(c) high target selectivity and (d) suitability for in vivo evaluation
of therapeutic potential, preferably via oral application to rats.
Central exposure, though an ambitious goal, was envisaged as we
were interested in investigations on the role(s) of FAS activity in
the brain.
During the last years, several non-covalent inhibitors of type I
FAS have been reported in the literature (Fig. 1): Compound 1 was
described by Rivkin et al. as the most potent compound among a ser-
ies of 3-aryl-4-hydroxyquinolinones,8 and the symmetric urea
derivative 2 was published by Vásquez et al. as GSK837149A.9
Althoughliterature IC50 values for both compounds are in the double
digit nM range, potency in our FAS assay was considerably lower.10
From a library of furanylmethylene-pyrimidinetriones, Richardson
and Smith identified compound 3 as the most potent one with a cal-
culated Ki value of 0.38 lM for the inhibition of the thioesterase do-
main of FAS.11 For none of these compounds pharmacokinetic or
in vivo data have been reported. In contrast, Butlin et al. and Walle-
nius et al. disclosed compound 4 as a type I FAS inhibitor (reported
IC50 for rat FAS inhibition: 350 nM) with oral availability in rats.12
To our knowledge, this is the first non-covalent inhibitor of type I
FAS reported to be suitable for in vivo experiments. However, the
authors state that no significant CNS exposure was measured for
compounds from the series containing 4.
A high throughput screen was set up at Boehringer Ingelheim
with the aim of generating a tool compound suitable for the valida-
tion of FAS as a therapeutic target. Intended properties were: (a) a
non-covalent mode of action, (b) sufficient potency (IC50 <1
l
M),
Here, we wish to report the discovery and characterisation of BI
99179 (compound 13) as a potent and selective inhibitor of human
FAS with central exposure upon oral administration in rats.13
Based on a high throughput screen of our internal compound
collection, we identified the cyclopentanecarboxanilide 5 with
⇑
Corresponding author. Tel./fax: +49 7351 540.
E-mail addresses: joerg.kley@boehringer-ingelheim.com, jtkley@web.de (J.T.
Kley).
0960-894X/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2011.07.083

J. T. Kley et al. / Bioorg. Med. Chem. Lett. 21 (2011) 5924–5927
5925
OH
O
H
N
H
N
N
N
O
S
N
O
O
N
S
N
H
N
Cl
N
H
O
N
O
H O
1
2
O
S
O
O
O
HN
O
N
N
O
N
H
O
4
3
N
Figure 1. Noncovalent inhibitors of type I FAS reported in the literature.
submicromolar IC50 in our FAS assay.10 Interestingly, like the struc-
turally diverse competitors’ compounds 1–4,14 compound 5 and
the other potent cyclopentanecarboxanilides described herein are
quite rigid structures with less than five rotatable bonds.15
The general synthetic approach employed to prepare the
cyclopentanecarboxanilides is outlined in Schemes 1–4.16 N-Meth-
ylation of the respective anilines or benzimidazole/benzoxazole
formation starting from 4-methylamino-benzoic acid furnishes
the required N-methylanilines (Scheme 1). These are acylated
using FMOC-3-aminocyclopentane-carboxylic acids under cou-
pling conditions outlined in Scheme 2. Removal of the FMOC moi-
ety (Scheme 2) is followed by the final diversifying acylation
(Scheme 3). Alternatively, the need for a protecting group can be
avoided by coupling preformed 3-acylamino-cyclopentane-carbox-
ylic acids to N-methylanilines as outlined in Scheme 4.
O
NH2
N
NH
F
(b)
(a)
F
F
R2
R2
R2
29a: R2 = phenyl
29b: R2 = imidazo[1,2-a]pyridin-2-yl
NH
NH
HA
NH2
A first series of R1 variations intended to explore SAR of the
right-hand side acyl moiety yielded the slightly more potent propi-
onyl compound 6 along with less potent analogues some of which
are listed in table 1 (compounds 7–9).
(c)
+
HO
A
O
N
29c: A = NH
Systematic modifications of all other parts of the molecule re-
vealed generally steep SAR except for R1 and R2 (data not shown).
We concluded that the N-methylcarboxanilide scaffold of 5 consti-
tutes an optimum and potency can be improved only through
modifications of R1 and R2. Furthermore, the optical antipode of
5, compound 10 was found to be inactive. Therefore, we next ex-
plored R2 more extensively, also aiming at improved solubility
through more polar and/or more flexible replacements of the phe-
nyl group. While most modifications significantly reduced potency
29d: A = NMe
29e: A = O
Scheme 1. Synthesis of N-methylanilines 29. Reagents and conditions: (a) (1)
(TFA)2O, TEA, DCM, 0 °C to rt; (2) MeI, K2CO3/acetone (for 29a) or NaH/DMF (for
29b), rt; (b) K2CO3, MeOH/H2O, 50 °C; (c) polyphosphoric acid, 200–210 °C (4 h),
then H2O (80 °C to rt) over night.
O
R2 ' =
Method
A, B, or C
NH
N
1
3
NH2
A
N
R2
R2
29
30
Educt
29a
29c
29d
29e
29a
29e
Product
R2
Method applied
Stereochemistry
(1R,3S)
30a
30c
30d
30e
30f
phenyl
A
B
B
C
A
C
R2’; A = NH
R2’; A = NMe
R2’; A = O
phenyl
(1R,3S)
(1R,3S)
(1R,3S)
(1S,3R)
30g
R2’; A = O
(1S,3R)
Scheme 2. Preparation of 3-aminocylopentanecarboxanilides 30. Reagents and conditions: Method A: (1) N-FMOC-3-aminocyclopentane-1-carboxylic acid (respective
diastereomer), TBTU, DIPEA, DMF, rt, over night; (2) diethylamine, DCM, rt, 2 h. Method B: (1) N-FMOC-3-aminocyclopentane-1-carboxylic acid (respective diastereomer),
diphenylphosphinic chloride, TEA, THF (10 min, rt) then 29, 60 °C, over night; (2) piperidine/THF, rt, over night; Method C: (1) N-FMOC-3-aminocyclopentane-1-carboxylic
acid (respective diastereomer), 1-chloro-N,N,2-trimethylpropenyldiamine, DCM, (rt, 2 h) then 29, 2,4,6-collidine, DCM, over night; (2) diethyl amine/THF, 40 °C, 2 h.

5926
J. T. Kley et al. / Bioorg. Med. Chem. Lett. 21 (2011) 5924–5927
O
O
Methods
N
N
D, E, or F
H
N
NH2
R1
O
R2
R2
30a-g
5-28
Scheme 3. Final acylation step. Reagents and conditions: Method D: R1COCl, TEA, DMF, rt; Method E: ethyl isocyanate, TEA, DMF, rt; Method F: 6-dimethylamino hexanoic
acid hydrobromide (for 9, 18, 20, 27) or difluoroacetic acid (for 23), TEA, TBTU, DMF, rt.
O
O
N
Method G
(a)
Method G
(b)
O
H
N
HO
H
N
HO
NH2
O
O
O
N
31
32
13
Scheme 4. Alternative synthesis of propionamides as exemplified for 13. Reagents and conditions (Method G): (a) propionyl chloride added drop wise to 31 in THF/aq NaOH;
(b) (1) 1-chloro-N,N,2-trimethylpropenyldiamine, DCM, rt; (2) addition to 29e, 2,4,6-collidine, DCM, rt.
Table 1
Structure and in vitro FAS inhibitory properties for compounds 5–28
O
N
H
N
A
R2 ' =
1
3
N
R1
O
R2
R1
R2
Absolute stereochemistry
IC50 (lM)
b
Compound
Final acylation step
Starting material
Method applieda
5
6
7
8
30a
30a
30a
30a
30a
30f
D
D
D
E
Methyl
Ethyl
2-Propyl
NH-Ethyl
–(CH2)5–NMe2
Methyl
Ethyl
Ethyl
Ethyl
Ethyl
Methyl
2-Propyl
NH-Ethyl
–(CH2)5–NMe2
NH-Ethyl
–(CH2)5–NMe2
Methyl
n-Propyl
CHF2
–CH2–O–CH3
NH-Ethyl
NMe2
Phenyl
Phenyl
Phenyl
Phenyl
Phenyl
Phenyl
(1R,3S)
(1R,3S)
(1R,3S)
(1R,3S)
(1R,3S)
(1S,3R)
(1R,3S)
(1R,3S)
(1R,3S)e
(1R,3S)
(1R,3S)
(1R,3S)
(1R,3S)
(1R,3S)
(1R,3S)
(1R,3S)
(1R,3S)
(1R,3S)
(1R,3S)
(1R,3S)
(1R,3S)
(1R,3S)
(1R,3S)
(1S,3R)e
0.48
0.30
0.42c
1.5
9
F
2.0d
>10d
0.19
0.31
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
D
D
D
D
G
D
D
E
30c
30d
30e
29bf
30c
30c
30c
30c
30d
30d
30e
30e
30e
30e
30e
30e
30e
30g
R20; A = NH
R20; A = NMe
R20; A = O
0.079
0.29
1.3d
Imidazo[1,2-a]pyridin-2-yl
R20; A = NH
R20; A = NH
R20; A = NH
R20; A = NH
R20; A = NMe
R20; A = NMe
R20; A = O
1.5d
0.40d
0.70
1.4d
F
E
F
2.4
D
D
F
D
E
D
F
D
0.20d
0.42d
0.27
0.91d
0.53
0.28d
0.58
>3
R20; A = O
R20; A = O
R20; A = O
R20; A = O
R20; A = O
–(CH2)5–NMe2
Ethyl
R20; A = O
R20; A = O
a
b
c
d
e
f
For a description of methods D, E, F, and G see Schemes 3 and 4.
Unless stated otherwise, median values of at least three independent IC50 determinations are given. For a description of the assay see Ref. 10.
Data point from a single IC50 determination.
Median from two independent IC50 determinations.
Specific rotation for 13: ½a D20
ꢀ
(c 0.3, MeOH) = +23; for 28: ½a D20
(c 0.3, MeOH) = ꢁ24).
ꢀ
Compound 14 was prepared by method G (Scheme 4) replacing 29e by 29b as starting material.

J. T. Kley et al. / Bioorg. Med. Chem. Lett. 21 (2011) 5924–5927
5927
(data not shown), we found that a range of heterobicyclic aromatic
systems could replace the phenyl moiety in 5 (compounds 11–14).
Gratifyingly the benzoxazole 13 which proved to be the most po-
tent compound has a significantly lower c log P value as compared
to 5 (c log P = 3.1 vs 3.6, respectively).17 We then further explored
the R1/R2 matrix including some close analogues of compound 13
(compounds 15–28). Data from this series indicate linear SAR
(i.e., the influence of R1 on potency was found to be independent
on the nature of R2 and vice versa) and confirmed the inactivity
of the optical antipodes (compound 28). With respect to both po-
tency and polarity, 13 remained the most interesting compound.
Cellular activity (i.e., inhibition of 14C-acetate incorporation) of
our compounds was determined in the mouse hypothalamic N-42
cell line.18 To confirm that reduced 14C-acetate incorporation is not
due to any cytotoxic effect, we determined cytotoxicity in a stan-
dard LDH-release assay.19 Again, compound 13 proved to be the
discovery of BI 99179 (13) which is characterised by high potency,
remarkably high selectivity and significant exposure (both
peripheral and central) upon oral administration in rats. We are
confident that selective tool compounds suitable for in vivo studies
like BI 99179 will help to better understand the role of FAS in phys-
iological as well as diseased conditions and to evaluate the poten-
tial of selective non-covalent FAS inhibitors as therapeutic agents.
Results from first acute and subchronic in vivo pharmacology stud-
ies with BI 99179 in rodents will be published elsewhere.
Acknowledgements
We gratefully appreciate the technical assistance of Regina Balk,
Chris Cantow, Elke Fischer, Walter Hudler, Stefanie Hugger, Oliver
Wachter, Angela Weiler, and Marine Willemse.
most potent derivative with an IC50 of 0.6
cellular assay. Notably, 13 showed no significant LDH release in
the cytotoxicity assay up to 30 M compound concentration. This
clearly indicates that (i) the cellular activity is not due to a cyto-
toxic effect and (ii) compound 13 (as other cyclopentanecarboxani-
lides tested) potently inhibit mouse FAS in addition to the human
enzyme applied in our biochemical assay.
lM in the mouse N-42
Supplementary data
l
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmcl.2011.07.083.
References and notes
1. For recent reviews see: (a) Flavin, R. et al Future Oncol. 2010, 6, 551; (b) Kridel,
S. J.; Lowther, W. T.; Pemble, C. W. Expert Opin. Invest. Drugs 2007, 16, 1817.
2. (a) Heaton, N. S. et al Proc. Natl. Acad. Sci. U.S.A. 2010, 107, 17345; (b) Munger, J.
et al Nat. Biotechnol. 2008, 26, 1179; (c) Yang, W. et al Hepatology 2008, 48,
1396.
3. Chakravarthy, M. V. et al J. Clin. Invest. 2007, 117, 2539.
4. Loftus, T. M. et al Science 2000, 288, 2379.
5. Kusakabe, T. et al J. Histochem. Cytochem. 2000, 48, 613.
6. Smith, K. R.; Thiboutot, D. M. J. Lipid Res. 2008, 49, 271.
7. D’Arcangelis, A. D., et al. U.S. Patent 53631, 2005.
To further evaluate the usefulness of 13 as a tool compound for
the validation of FAS as a potential therapeutic target, it was subject
to extensive selectivity testing: When screened against a panel of
30 diverse targets, 13 showed less than 20% inhibition at a concen-
tration of 10
10
M.16 Also for the major human P450 isoenzymes (CYP1A2,
CYP2C9, CYP2C19, CYP2D6, and CYP3A4) IC50 values were deter-
mined to be higher than 10 M. Compound 13 has been incorpo-
lM in all cases, indicating IC50 values far above
l
l
8. Compound ‘10v’ from: Rivkin, A. et al Bioorg. Med. Chem. Lett. 2006, 16, 4620.
9. Vásquez, M. J. et al FEBS J. 2008, 275, 1556.
rated into Boehringer Ingelheim’s HTS compound collection and
results available to date from a double digit number of screens
against diverse targets confirm the very clean selectivity profile.
Favourable in vitro PK parameters of 13 (CaCo-2 permeability:
94 ꢂ 10ꢁ6 cm/s with no indication for the involvement of efflux
transporters; high metabolic stability in rat and human
microsomes: <27% and <26% of the respective liver-blood-flow
QH; rat plasma protein binding: 97.6 0.5%) appear to translate
into a low clearance (CL = 8.2 ml/min/kg) and a moderate volume
of distribution (Vss = 1.6 L/kg) following intravenous administra-
tion to male Han/Wistar rats. Furthermore, and crucial for the
envisaged in vivo studies, the compound proved to be well suitable
for oral application as seen from the data shown in table 2. Notably
similarly high compound levels in plasma and brain are seen upon
oral administration (the ratio CCSF: Cbrain of approx. 4% can be ratio-
nalized by protein binding which was determined to be 97.6% with
rat plasma).
10. Human FAS enzyme assay: Human type I fatty acid synthase is a large multiunit
protein that harbours seven enzymatic activities catalysing the reductive
synthesis of long chain fatty acids from acetyl CoA and malonyl CoA. In the
reductive steps NADPH is consumed forming NADP, measured by analysing the
decrease in fluorescence.
Compounds diluted in 0.1% DMSO, 326 lM NADPH, 60 lM acetyl CoA (from
Sigma, cat.-no. A-2181) in phosphate buffer (100 mM KH2PO4/K2HPO4, 1 mM
EDTA, 1 mM DTT, pH 7.0), and FAS enzyme (prepared from HeLa cells) diluted
in enzyme buffer (20 mM KH2PO4/K2HPO4, 1 mM EDTA, 1 mM DTT, 5% glycerol,
pH 7.4) were incubated for 60 min at 37 °C. The reaction was then started by
the addition of 200 lM malonyl CoA. The optical density was determined at
340 nm over 10 minutes and the slope (i.e., Vmax) was calculated.
The IC50 value of the known covalent FAS inhibitor Cerulenin was determined
to be 5.8 lM in this assay.
11. Richardson, R. D.; Smith, J. W. Mol. Cancer Ther. 2007, 6, 2120.
12. Butlin, R. J., et al. Poster presented at the 15th RSC-SCI Medicinal Chemistry
Symposium, Cambridge, UK, September 2009; Wallenius, K., et al.; Poster
presented at the American Diabetes Association: 68th Scientific Sessions, San
Francisco, CA, Poster 58-LB, June 2008; Compound 4 is also found as example
117 in: Butlin, R. J., et al. PCT Int. Appl. WO 2008/075070, 2008.
13. BI 99179 and other cyclopentanecarboxanilides are claimed in: Kley, J., et al.
PCT Int. Appl. WO 2011/048018, 2011.
14. In addition to structural diversity, the compounds may also bind to different
domains of FAS. From preliminary data (not shown) we conclude that the
cyclopentanecarboxanilides disclosed herein probably bind to the ketoacyl
reductase (KAR) domain of FAS.
Clearly, the relatively low aqueous solubility of 13 (5 lg/ml @
pH 7.4) does not compromise the compound’s good PK properties.
Nevertheless, for experiments where high aqueous solubility is key
rather than cellular activity and favourable PK properties, com-
pounds 18 and 27 with submicromolar potency in the enzymatic
assay, bearing a basic moiety designed to improve aqueous solubil-
ity may be the tools or choice.
In conclusion, cyclopentanecarboxanilides were identified from
a high throughput screen of Boehringer Ingelheim’s compound col-
lection for FAS inhibitors. Optimisation of the initial hits lead to the
15. The number of rotatable bonds was calculated using an in-house algorithm.
Introduction of the flexible dimethylaminopentyl side chain (compounds 9, 18,
20, and 27 with a higher number of rotatable bonds) significantly decreases
potency as shown in Table 1.
16. For details see Supplementary data.
17. c log P values were calculated using Program CLOGP, Version 4.83, Daylight
Chemical Information Systems, Inc. Los Altos, CA. (http://www.daylight.com)
Chou, J. T.; Jurs, P. C. J. Chem. Inf. Comput. Sci. 1979, 19, 172.
18. 14C-Acetate incorporation in N-42 cells: N-42 cells are immortalized clonal
neuronal mouse hypothalamic cells from CELLutions Biosystems Inc. N-42 cells
were incubated first with the compound for 60 minutes at 37 °C, then with 14C-
acetate dilution (106 dpm) for 4 h at 37 °C. After chloroform/methanol
extraction, the lower fraction was vaporized and incorporated radioactivity
was quantified by scintillation counting.
Table 2
Pharmacokinetic parameters of BI 99179 (13) in male Han/Wistar rats (fasted) upon
oral application of 4 mg/kg
t1/2
(h)
tmax
(h)
Cmax
(nM)
AUC0-inf
(nM h)
F
(%)
Cbrain,2h
(nM)
CCSF,2h
(nM)
19. Cytotoxicity assay: Measurement of LDH release from U937 cells after 20 h of
compound incubation (CytoTox-ONE™ Homogeneous Membrane Integrity
Assay, Promega, G7890).
3.0
0.5
2110
9350
46
1300
50