Aminopyrrolidineamide inhibitors of site-1 protease

Article abstract of DOI:10.1016/j.bmcl.2007.06.031

The discovery and efficacy of a series of potent aminopyrrolidineamide-based inhibitors of sterol regulatory element binding protein site-1 protease is described.

Full text of DOI:10.1016/j.bmcl.2007.06.031

Bioorganic & Medicinal Chemistry Letters 17 (2007) 4411–4414
Aminopyrrolidineamide inhibitors of site-1 protease
Bruce A. Hay,a,* Barbara Abrams,a Allice Y. Zumbrunn,a James J. Valentine,a
Laurie C. Warren,a Stephen F. Petras,a Lorraine D. Shelly,a Angela Xia,a
Alison H. Varghese,a Julie L. Hawkins,a Jennifer A. Van Camp,b Michael D. Robbins,a
Katherine Landschulza and H. James Harwood, Jr.a
aPfizer Global Research and Development—Groton Laboratories, Groton, CT 06340, USA
bPfizer Global Research and Development—Ann Arbor Laboratories, Ann Arbor, MI 48105, USA
Received 16 May 2007; revised 4 June 2007; accepted 5 June 2007
Available online 10 June 2007
Abstract—The discovery and efficacy of a series of potent aminopyrrolidineamide-based inhibitors of sterol regulatory element bind-
ing protein site-1 protease is described.
Ó 2007 Elsevier Ltd. All rights reserved.
Site-1 protease (S1P) is a unique membrane-bound ser-
ine protease responsible for cholesterol-sensitive cleav-
age of the transcription factor sterol regulatory
element binding protein (SREBP).1–4 The S1P-mediated
SREBP cleavage is the first of two cleavage events re-
quired to release mature transcription-competent
SREBP from the Golgi and enable its passage to the nu-
cleus1–3 where it regulates the gene expression of a wide
variety of lipogenic genes involved in cholesterol synthe-
sis (e.g., HMG-CoA synthase, HMG-CoA reductase,
squalene synthase, etc.) and fatty acid synthesis (e.g.,
fatty acid synthase, acetyl-CoA carboxylase, stearoyl-
CoA desaturase, etc.) as well as the expression of a
variety of other genes involved in nutrient uptake and
processing, such as the LDL receptor.
and the metabolic syndrome.6,7 With these potential
therapeutic benefits in mind, we initiated a program to
look for small molecule inhibitors of S1P. Herein, we de-
scribe the initial results of this effort.
A high throughput screen, using purified human S1P
and a fluorescent assay format that assesses cleavage
of the synthetic cleavage site peptide, Ac-VFRSLK-
MCA (360 nm excitation; 460 nm emission),8 was
employed to search the Pfizer compound file for S1P
inhibitors. Only a handful of compounds showed great-
er than 50% inhibition of S1P at 2.25 lM. One of the
most interesting hits was the aminopyrrolidine deriva-
tive 1a (Fig. 1), with an IC50 of 1.4 lM. This compound
was originally produced as part of a large (3786 com-
pounds) library, and an analysis of the high throughput
screen results revealed that 39 of these library com-
pounds inhibited S1P greater than 25% at 2.25 lM.
Retesting these compounds yielded measurable IC50s
for two close-in derivatives, 1b and 1c (Table 1). All of
the other library compounds had IC50s greater than
30 lM.
Inhibition of S1P activity has the potential to simulta-
neously inhibit both the cholesterolgenic and fatty acid
synthetic pathways and, as a consequence, lower plasma
cholesterol and triglycerides in patients with dislipide-
mia.1–3 A S1P inhibitor would thus combine the benefits
of an HMG-CoA reductase inhibitor5 with those of an
acetyl-CoA carboxylase inhibitor6 and therefore would
be useful not only for patients with dislipidemia, but
also for those with a variety of cardiometabolic risk fac-
tors associated with insulin resistance, diabetes, obesity,
O
N
Cl
O
HN
Keywords: Site-1 protease; SREBP; Sterol regulatory element binding
protein; Metabolic syndrome.
1a
*
Figure 1. High throughput screening hit 1a.
Corresponding author. E-mail: bruce_hay@sbcglobal.net
0960-894X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.06.031

4412
B. A. Hay et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4411–4414
Table 1. S1P IC50 values for compounds 1a–n
R3
R4
R2
O
( )n
N
R1
R5
HN
Compound
n
R1
R2
R3
R4
R5
S1P IC50 (nM)
1a
1b
1c
1d
1e
1f
1
1
1
1
1
1
1
0
1
1
1
1
1
Cl
F
H
H
Cl
H
H
H
H
H
H
H
Cl
H
H
H
H
H
H
H
H
H
H
Cl
H
H
H
H
H
H
H
H
H
H
H
H
H
Cl
H
H
H
OiPr
1400
9000
9600
540
OiPr
H
OiPr
OMe
OMe
Cl
CH2N(Et)2
OBn
310
840
CH2N(Et)2
OBn
OiPr
1g
1h
1i
Cl
Cl
590
>5000
430
Cl
H
CH2N(Et)2
CH2N(Et)2
OBn
1j
3700
510
1k
1m
1n
H
Me
OEt
CH2N(Et)2
CH2N(Et)2
1600
16,000
Follow-up libraries around compound 1a were initiated.
The first library was a full combinatorial cross of 11
amines by 32 acids, with monomers chosen based on
similarity to the original hit and some of the SAR
gleaned from the initial high throughput screen. The
core template remained static. The library was prepared
as shown in Scheme 1, involving reductive amination of
the BOC-pyrrolidinone followed by amide coupling and
removal of the BOC protecting group. A total of 243 out
of 352 (69%) compounds were returned after synthesis
and HPLC autopurification. Potencies of selected com-
pounds are listed in Table 1. Several compounds were
significantly more potent than the initial lead, including
the dibasic amines 1d and 1f, and the more lipophilic
compounds 1e and 1g. The 2-substituted phenethyl-
amine monomer was important for S1P activity, 2-Cl
and 2-methoxyphenethylamine derived compounds had
comparable potency, with the 2-methylphenethylamine
(1m) slightly less potent.
Increasing the size of the 2-substituent was detrimental,
exemplified by the ethoxy compound 1n. In general,
2-substituted phenethylamines led to equal or better
potency relative to 3-substituted phenethylamines (1a
vs 1c and 1g vs 1k) while 4-substituted phenethylamines
led to significantly less active compounds (1j vs 1f).
Additional substitutions on either phenyl ring were det-
rimental, the lone exception being the 2,6-dichlorophen-
ethylamine monomer in 1i. The 2-carbon chain in the
phenethylamine was also critical, as the benzylamine
derivative 1h had no measurable activity. Compounds
with amides other than 4-substituted benzamides also
had no measurable activity with any of the phenethyl-
amine monomers. Based on the data from this initial
library we felt that there might be an opportunity for
further potency improvements by additional modifica-
tion of the benzamide substituent at the 4-position.
A second library was prepared using the route shown in
Scheme 1 in which the 4-isopropoxybenzamide and 2-
Cl-phenethylamine monomers were kept constant while
the core was varied. No measurable S1P inhibition was
observed with any of these compounds. Selected data
are shown in Table 2, including the tetrahydrothiophene
R1
O
HN
a
( )n
( )n
X
X
3
Table 2. S1P IC50 values for compounds 1o–1t
2
O
b
R2
N
R2
Cl
R1
( )n
c
X
X =
NBOC
Compound
X
n
R1
IC50 (nM)
R1
R1
NH
O
N
O
N
1a
1o
1p
1q
1r
1s
1t
NH
NCH3
S
1
1
1
1
2
0
1
OiPr
OiPr
OiPr
OiPr
OiPr
OiPr
1400
>30,000
>30,000
>30,000
>30,000
>30,000
( )n
( )n
X
X
1
4
NCH@O
NH
NH
Scheme 1. Synthesis of S1P inhibitors. Reagents: (a) i—R1NH2,
EtOH, Toluene; ii—NaBH4, EtOH; (b) substituted benzoic acid,
HBTU, DIPEA, DCE; (c) HCl, DCE.
O
Methyl(1-piperidine) >30,000

B. A. Hay et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4411–4414
4413
Table 3. IC50s for selected second iteration compounds 7a–h
1p, N-methylpyrrolidine 1o, N-formylpyrrolidine 1q, 3-
piperidine 1r, and azetidine 1s. A tetrahydrofuran ana-
log 1t was synthesized at a later date and also had no
measurable inhibition of S1P.
O
N
R2
R1
HN
R2
CH3
A third library was then synthesized to follow up on the
activity of the diethylamine derivatives 1d and 1f. The
template formation involved reductive amination of
BOC-pyrrolidinone with 2-(2-methoxyphenyl)ethyl-
amine or 2-(2-chlorophenyl)ethylamine and amide for-
mation with 4-carboxybenzaldehyde (Scheme 1),
followed by parallel chemistry (reductive amination
and BOC protecting group removal, Scheme 2) for the
conversion to the final diamines. Many of these com-
pounds showed significant potency improvements (Ta-
ble 3), 7b was the most potent compound from this
library with an IC50 of 8 nM.
Compound
R1
IC50 (nM)
N
N
7a
OMe
37
CH3
7b
7c
7d
7e
7f
Cl
8
29
OMe
OMe
OMe
OMe
OMe
OMe
N
Compound 1d was selected for further evaluation be-
cause it had reasonable potency and lower molecular
weight and lipophilicity relative to the other potent com-
pounds. The enantiomers were separated by preparative
HPLC,9 and the more potent 1u (PF-429242, Table 4)
was assigned the R stereochemistry after direct synthesis
from the chiral 8 (Scheme 3), followed by amide cou-
pling and removal of the BOC protecting group as in
Scheme 1.
95
N
OH
H
N
100
250
140
160
N
N
O
7g
7h
CH3
PF-429242 showed a high degree of selectivity for S1P
inhibition relative to other serine proteases, exhibiting
an IC50 for S1P inhibition of 170 nM and showing no
significant inhibition of trypsin, elastase, proteinase K,
plasmin, kallikren, factor XIa, thrombin, or furin at
concentrations up to 100 lM and only modest inhibition
of urokinase (IC50 = 50 lM) and factor Xa
(IC50 = 100 lM). In cultured human liver (Hep-G2)
cells, PF-429242 prevented proteolytic processing and
nuclear translocation of SREBP10 (complete inhibition
at a dose of 10 lM), reduced the expression of key genes
involved in cholesterol synthesis (e.g., HMG-CoA syn-
thase; EC50 = 0.3 lM) and fatty acid synthesis (e.g.,
fatty acid synthase; EC50 = 2 lM),11 and inhibited both
pathways (cholesterol synthesis, EC50 = 600 nM; fatty
acid synthesis, 43% inhibition at 10 lM)12 in the absence
of any cytotoxicity (LDH release assay). LDL receptor
gene expression and LDL receptor-mediated LDL inter-
nalization13 by Hep-G2 cells were also inhibited by PF-
429242, but to a very small degree relative to inhibition
H2N
N
O
Table 4. IC50s and enantiomeric configurations of 1u and 1v
Compound
Structure
IC50 (nM)
O
N
1u (PF-429242)
170
N
O
O
HN
O
N
1v
970
N
HN
NH2
*
HN
R4
N
a
R4
N
*
O
O
O
H
R3
R3
N
BOC
N
BOC
b
a
8
9
R1
R1
R1
Scheme 3. Synthesis of chiral 3-pyrrolidineamides. Reagents: (a) 1-(2-
bromoethy)-2-methoxybenzene, Et3N.
O
N
N
O
N
HN
N
BOC
N
BOC
of cholesterol synthesis. For example, LDL receptor-
mediated LDL internalization was not altered by PF-
429242 at concentrations that produced 65% inhibition
of cholesterol synthesis, and was only inhibited by
5
6
7
Scheme 2. Synthesis of dibasic S1P inhibitors. Reagents: (a) R3R4NH,
N(CH3)4BH(OAc)3, DCE; (b) HCl, DCE.

4414
B. A. Hay et al. / Bioorg. Med. Chem. Lett. 17 (2007) 4411–4414
35% at concentrations that inhibited cholesterol synthe-
sis by >90%.
References and notes
1. Brown, M. S.; Goldstein, J. L. Cell. 1997, 89, 331.
2. Brown, M. S.; Goldstein, J. L. Proc. Natl. Acad. Sci.
U.S.A. 1999, 96, 11041.
3. Yang, J.; Goldstein, J. L.; Hammer, R. E.; Moon, Y. A.;
Brown, M. S.; Horton, J. D. Proc. Natl. Acad. Sci. U.S.A.
2001, 98, 13607.
4. Bodvard, K.; Mohlin, J.; Knecht, W. Protein. Expr. Purif.
2007, 51, 308.
5. Harwood, H. J., Jr.; Hamanaka, E. S. Emerg. Drugs 1998,
3, 147.
PK data for PF-429242 showed that it had rapid clear-
ance (CL = 75 ml/min/kg) and poor oral bioavailability
(5%) in rats. The high clearance of PF-429242 was not
predicted in vitro by human microsomes. Based on
pharmacokinetic modeling, it was determined that
administration of PF-429242 intraperitoneally every
6 h for 24 h at a dose of 30 mg/kg/dose would allow suf-
ficient exposure to assess acute (24 h) in vivo efficacy.
6. Harwood, H. J., Jr. Expert Opin. Ther. Targets 2005, 9,
267.
7. Kotzka, J.; Muller-Wieland, D. Expert Opin. Ther. Tar-
gets 2004, 8, 141.
The antilipogenic activity of PF-429242 was then eval-
uated in male CD1 mice (n = 5 per group) at doses of
10 and 30 mg/kg/dose i.p. every 6 h for 24 h. This re-
sulted in greater than 80% reductions of HMG-CoA
synthase gene expression11 at both doses, dose related
reductions in fatty acid synthase gene expression11
(50% and 75%, respectively), and inhibition of the
cholesterol and fatty acid synthetic pathways12 (50%
inhibition of both pathways at 10 mg/kg/dose; 80%
for both pathways at 30 mg/kg/dose). LDL receptor
gene expression11 was not substantially altered at
10 mg/kg/dose but was 35% reduced at 30 mg/kg/dose,
consistent with the smaller reduction of LDL receptor
activity relative to reductions in cholesterol synthesis
in cultured cells.
8. The enzymatic activity and inhibition of S1P were
measured fluorometrically using the MCA-conjugated
peptidyl substrate, Ac-VFRSLK-MCA, essentially as
described in Cheng, D.; Espenshade, P. J.; Slaughter, C.
A.; Jaen, J. C.; Brown, M. S.; Goldstein, J. L. J. Biol.
Chem. 1999, 274, 22805, with the following modifications:
The reaction was conducted in 96-well plate format with
an assay volume of 40 lL, the fluorogenic peptide
substrate concentration was 20 lM, the purified His-
tagged human S1P enzyme (secreted into serum-free pH
8.0 medium from stably transfected CHO-K1 cells and
purified by nickel column affinity chromatography) con-
centration was 25 lg/mL (1 lg/well), the reaction was
conducted for 4 h at 37 °C, compounds dissolved in
DMSO were added such that the final DMSO concentra-
tion in the assay was 2.5%. Under these conditions, the
assay exhibited a S/N ratio of 8 and a coefficient of
variation of 15%. Confirmed S1P inhibitors did not exhibit
fluorescence at either 360 or 460 nm nor did they quench
fluorescence from a control well-containing MCA.
9. Enantiomers were separated and analyzed on preparative
and analytical chiralpak AD columns, respectively, using a
heptane/ethanol gradient with 0.2% DEA as mobile phase.
10. Proteolytic processing and nuclear translocation of
SREBP were assessed as described by Cheng et al.
(1999) op cit (Ref. 8).
11. HMG-CoA synthase, fatty acid synthase, and LDL
receptor gene expression in cultured cells and in hepatic
tissue samples was measured as outlined in Hawkins, J. L.;
Gafvels, M.; Olin, M.; Lund, E. G.; Andersson, U.;
Schuster, G.; Bjorkhem, I.; Russell, D. W.; Eggertsen, G.
J. Clin. Invest. 2002, 110, 1191.
12. Cholesterol and fatty acid synthesis in cultured cells and in
experimental animals were measured as outlined in Har-
wood, H. J., Jr.; Petras, S. F.; Shelly, L. D.; Zaccaro, L.
M.; Perry, D. A.; Makowski, M. R.; Hargrove, D. M.;
Martin, K. A.; Tracey, W. R.; Chapman, J. G.; Magee, W.
P.; Dalvie, D. K.; Soliman, V. F.; Martin, W. H.;
Mularski, C. J.; Eisenbeis, S. A. J. Biol. Chem. 2003,
278, 37099.
In summary, we have identified and characterized the
biological efficacy of the first examples of site-1 prote-
ase inhibitors. Compounds with IC50s of less than
10 nM were identified after 2 library iterations, an in-
crease of greater than 200-fold relative to the initial
high throughput screening hit. Characterization of
the efficacy of PF-429242, a less potent compound
with better physical properties, revealed that the S1P
inhibitor prevented cleavage and nuclear translocation
of the transcription factor SREBP, reduced expression
of a variety of cholesterolgenic and fatty acid syn-
thetic genes, and reduced cholesterol and fatty acid
synthesis in cultured cells and in experimental animals.
S1P inhibition is therefore an attractive therapeutic
target for treating dislipidemia and a variety of car-
diometabolic risk factors associated with altered cho-
lesterol and fatty acid metabolism.
Acknowledgments
The authors wish to thank Robert Corr, Neal Falcone,
Anthony Provatas, Jack Oscarson, and the Pfizer Gro-
ton automated synthesis and registration laboratory
for preparation of the initial S1P library and Kosea
Frederick and Kazuko Sagawa for pharmacokinetic
modeling of PF-429242 in rats.
13. LDL receptor mediated LDL internalization was mea-
sured in cultured cells as outlined in: Harwood, H. J., Jr.;
Pellarin, L. D. Biochem. J 1997, 323, 649, using DiI-LDL
as the probe.