Discovery of a new class of glucosylceramide synthase inhibitors

Article abstract of DOI:10.1016/j.bmcl.2011.09.037

A novel series of potent inhibitors of glucosylceramide synthase are described. The optimization of biochemical and cellular potency as well as ADME properties led to compound 23c. Broad tissue distribution was obtained following oral administration to mice. Thus 23c could be another useful tool compound for studying the effects of GCS inhibition in vitro and in vivo.

Full text of DOI:10.1016/j.bmcl.2011.09.037

Bioorganic & Medicinal Chemistry Letters 21 (2011) 6773–6777
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery of a new class of glucosylceramide synthase inhibitors
⇑
Elena Koltun, Steven Richards , Vicky Chan, Jason Nachtigall, Hongwang Du, Kevin Noson, Adam Galan,
Naing Aay, Art Hanel, Amanda Harrison, Jeff Zhang, Kwang-Ai Won, Danny Tam, Fawn Qian, Tao Wang,
Patricia Finn, Kathleen Ogilvie, Jon Rosen, Raju Mohan, Christopher Larson, Peter Lamb, John Nuss,
Patrick Kearney
Exelixis, 210 E. Grand Avenue, South San Francisco, CA 94005, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
A novel series of potent inhibitors of glucosylceramide synthase are described. The optimization of
biochemical and cellular potency as well as ADME properties led to compound 23c. Broad tissue distri-
bution was obtained following oral administration to mice. Thus 23c could be another useful tool
compound for studying the effects of GCS inhibition in vitro and in vivo.
Received 25 August 2011
Revised 8 September 2011
Accepted 12 September 2011
Available online 17 September 2011
Ó 2011 Elsevier Ltd. All rights reserved.
Keywords:
Glucosylceramide synthase inhibitor
GCS
Diabetes
Gaucher disease
Cancer
Inflammation
Glucosylceramide synthase (GCS) is a glucosyltranserase that
processes the sphingolipid ceramide (1) (Fig. 1). The product, gluco-
sylceramide (2), can be further elaborated with a variety of oligosac-
charides to become glycosphingolipids called gangliosides (most
notably GM1 and GM3 (3)). Excess glycosphingolipids have been
correlated with a variety of human diseases.1 The pathologies asso-
ciated with lysosomal storage disorders (Gaucher, Fabry, Tay-Sachs,
and Sandhoff) stem from improper disposal of glucosylceramide and
other glycosphingolipids.2 Excess ganglioside GM3 has been linked
to insulin resistance wherein GM3 in the cell membrane forms lipid
rafts that disrupt the membrane environment around the insulin
receptor. As a result of this perturbation, the insulin receptor does
not signal properly resulting in insulin resistance.3 In addition, dis-
regulated glycosphingolipids have been implicated in multi-drug
resistance, atherosclerosis and inflammation.4–6
The beneficial effects of small molecule inhibition of GCS have
been well studied. There are two ‘‘privileged’’ classes from which
most known GCS inhibitors arise (Fig. 2). The iminosugar class is
well represented by Migustat (4) (ZavescaÒ) and AMP-DNM
(5).7,8 These compounds are also potent inhibitors of intestinal gly-
cosidases.9 Miglustat (4) is available as substrate reduction therapy
(SRT) for patients who are not candidates for enzyme replacement
therapy (ERT). It is also used to treat Niemann-Pick type C disease
and has shown efficacy in patients with cystic fibrosis.10 A large
scale synthesis of AMP-DNM (5) was recently reported to support
pre-clinical development.11
Ceramide analogs are best represented by Genz-123346 (6) and
Genz-112638 (7).12–14 Genz-123346 (6) has been used to prove the
concept that chronic GCS inhibition can ameliorate insulin resis-
tance in animal models of diabetes.12 Genz-112638 (7) has ad-
vanced to phase 3 clinical trials as SRT for type 1 Gaucher disease.
Given the broad therapeutic potential of small molecule GCS
inhibitors, we sought to determine if other compound classes could
modulate GCS activity. As GCS is expressed throughout the body,
the goal was to identify potent inhibitors with systemic exposure
and minimal off-target effects.
To that end, we embarked on an HTS campaign in which our li-
brary of over 4 million compounds was tested in a coupled assay
format. GCS activity was measured as the amount of UDP-glucose
consumed during the synthase-catalyzed reaction (Fig. 3). Upon
quenching, remaining UDP-glucose was processed by UDP-glucose
dehydrogenase, producing NADH. The NADH formed then partici-
pated in a diaphorase mediated reduction of resazurin (8) to fluo-
rescent resorufin (9). This assay was sufficiently robust to also
serve as the primary assay for our hit-to-lead research efforts.
The screen yielded a variety of potent hits. One of which was
the racemic dipeptide shown in Figure 4. After both enantiomers
were synthesized, the R isomer (10, derived from D serine) was
found to be about 80 fold more potent than the S isomer (11). This
⇑
Corresponding author. Tel.: +1 650 837 7000; fax: +1 650 837 8177.
E-mail address: srichard@exelixis.com (S. Richards).
0960-894X/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2011.09.037

6774
E. Koltun et al. / Bioorg. Med. Chem. Lett. 21 (2011) 6773–6777
O
Reaction
HN
16
UDP-glucose
HO
1, ceramide
UDP-glucose
dehydrogenase
11
UDP-D glucuronate
OH
1, ceramide
2 H+
NAD+
+ 2 NADH +
GCS
UDP
UDP-glucose
GCS
UDP
2, glucosylceramide
O
Readout
OH
O
HN
16
HO
HO
NADH
O
HO
O
O
HO
O
N
O
H+
11
OH
OH
+
2, glucosylceramide
diaphorase
N
O-
1) Lactosylceramide synthase
2) GM3 synthase
8
9
Figure 3. GCS biochemical screen.
HO
OH
O
CO2H
O
OH
O
OH
O
H
N
OH
HN
O
H
16
O
O
O
N
N
HO
HO
O
11
N
OH
OH
OH
OH
3, GM3
N
O
O
O
Cl
Figure 1. GCS mediated glycosphingolipid synthesis.
10 - (R) GCS IC50 = 32 nM
OH
OH
N
11
- (S) GCS IC50 = 2,700 nM
HO
HO
HO
HO
Figure 4. HTS hit.
N
O
Substituted serine derivatives (18) were prepared as shown in
Schemes 2 and 3. The dianion of boc-serine was generated using
NaH in DMF and trapped with the appropriately substituted alkyl
bromide (Scheme 2).15 The resulting acid was coupled with
N-methyl piperazine using HATU to generate boc amide 17. Com-
pound 17 was then deprotected under acidic conditions and
coupled with acid 15 to yield serine analogs (18).
Boc-aziridine methyl ester 20 was prepared in three steps from
boc-serine methyl ester (19) in an analogous manner to what has
been reported (Scheme 3).16 Activation of 19 as the mesylate fol-
lowed by boc-deprotection, cyclization and re-protection pro-
ceeded smoothly. The aziridine (20) was then subjected to Lewis
acid mediated opening with alcohols.17 Acidic boc deprotection
and HATU coupling of the resulting amine with acid 15 provided
ester 21. This ester was hydrolyzed and coupled with N-methyl
piperazine to furnish serine analogs 18f and 18g.
OH
OH
4, Miglustat
5, AMP-DNM
GCS IC50 = 90 nM
GCS IC50 = 22,000 nM
A549 Cell IC50 = 16,500 nM
OH
OH
O
O
N
N
HN
HN
O
O
O
O
7
6
6, Genz-123346
7, Genz-112638
GCS IC50 = 9 nM
GCS IC50 = 16 nM
A549 Cell IC50 = 2 nM
A549 Cell IC50 = 2 nM
Figure 2. Representative GCS inhibitors.22
Analogs of the western amide portion of the scaffold were syn-
thesized as shown in Scheme 4. Boc-benzyl serine (12) was treated
with methanolic HCl to both deprotect the boc group and form the
methyl ester.18 The resulting amino ester was coupled with the
appropriately substituted acid (15) and hydrolyzed to form acid
22. HATU coupling with amines followed to provide the desired
analogs (23). In the case of compounds 23c–d, boc amines were
coupled, necessitating a subsequent acidic deprotection step.
A summary of our efforts to develop aryl ether SAR is shown in
Table 1. Early work on this portion of the scaffold revealed that the
aryl ether substituent appended to the pyridine was required for
activity. Substitution of that aryl ring was tolerated. The para chlo-
rophenyl substituent of lead 10 could be replaced by a methyl
group (16a) without losing potency. However, deletion of the
4-chlorophenyl substituent resulted in analog 16b which was 30
stereochemical preference was confirmed by the preparation of
analogous enantiomer pairs (data not shown). Thus we decided
to begin developing SAR around active enantiomer 10.
The synthesis of aryl ether analogs (16) is shown in Scheme 1.
Benzyl protected, N-boc serine (12) was coupled with N-methyl
piperazine under standard conditions. Following acid mediated
boc deprotection, the resulting amine (13) was ready for coupling
with acid 15. The acid (15) was prepared in a two step procedure
starting from the appropriately substituted ethyl-2-chloronicoti-
nates (14). The SNAr displacement of the chloride with substituted
phenols required heating to 120 °C in a sealed tube under basic
conditions. The resulting ester was hydrolyzed and the two part-
ners (15 and 13) were coupled to provide the desired aryl ether
analogs (16).

E. Koltun et al. / Bioorg. Med. Chem. Lett. 21 (2011) 6773–6777
6775
O
O
O
O
O
O
H
N
H
N
NH2
.
O
O
a, b, c
d, e, f
a, b
O
O
N
HO
HCl
N
O
N
O
HO
O
O
19
20
O
12
13
H
N
N
g, h
R1
O
Cl
18
R2
O
O
O
O
O
R
H
N
N
c
N
R3
CF3
N
O
O
R4
R5
21
R7
Scheme 3. Reagents and conditions: (a) MsCl, Et3N, THF, 0 °C to rt, 18 h; (b) 4 M HCl
in dioxane, Et2O, rt, 18 h; (c) (i) Hünig’s base, MeCN, 60 °C, 8 h. (ii) (Boc)2O, Hünig’s
base, MeCN, rt, 18 h; (d) alcohol, BF3.OEt2, CHCl3, rt, 18 h; (e) 4 M HCl in dioxane, rt,
18 h; (f) HATU, DCE, 15, rt, 18 h; (g) LiOH, THF, H2O, rt, 1 h; (h) EDC, HOBt, CHCl2, N-
methylpiperazine, rt, 8 h.
R6
16
R1
R1
O
H
R2
R2
R3
N
N
a, b, c
HO
Cl
d, e
12
O
N
HO
N
O
O
O
O
Cl
O
O
R4
R5
CF3
R7
R6
22
14
15
O
O
Scheme 1. Reagents and conditions: (a) EDC, HOBt, CHCl2, N-methylpiperazine, rt,
8 h; (b) 4 M HCl in dioxane, MeOH, 55 °C, 3 h; (c) HATU, MeCN, Hünig’s base, 15, rt,
6 h; (d) Cs2CO3, DMA, 120 °C, 48 h; (e) LiOH, THF, H2O, 100 °C, 4 h.
H
N
N
d, e
R
Cl
O
O
CF3
O
O
H
N
H
N
O
O
a, b
HO
HO
N
O
N
O
23
O
R
Scheme 4. Reagents and conditions: (a) TMSCl, MeOH, rt, 18 h; (b) HATU, DCE, 15,
rt, 18 h; (c) LiOH, THF, H2O, rt, 1 h; (d) amine, HATU, DCE, rt, 18 h; (e) 4 M HCl in
dioxane, MeOH, 55 °C, 3 h (only for compounds 23c–d).
16
17
compounds 16h and 16i were both marginally less active than lead
10. When the 3-fluro pyridyl substituent was combined with
2,4-dichloro phenyl ether, the resulting compound, 16j, was 10
fold more potent that lead 10. This increased potency was found
to be driven entirely by the aryl ether substitution, as evidenced
by the 2,4-dichlorophenyl analog 16k. The 2-methyl, 4-chloro phe-
nyl analog 16m was similarly potent. Analog 16n with its 2-Chloro,
4-trifluromethyl phenyl substitution was 30 fold more potent than
lead compound 10. Thus the 2,4-disubstitued aryl ether motif was
shown to provide the most significant potency improvement and
was conserved for subsequent SAR development.
O
H
N
N
c, d
N
Cl
N
O
O
O
R
CF3
18
Scheme 2. Reagents and conditions: (a) NaH, DMF,0 °C, bromide, 18 h; (b) EDC,
HOBt, CHCl2, N-methylpiperazine, rt, 8 h; (c) 4 M HCl in dioxane, MeOH, 55 °C, 3 h;
(d) HATU, DCE, 15, rt, 18 h.
fold less potent. Replacing the 4-chloro substituent with a fluoro
group (16c) also resulted in a less potent compound, losing 10 fold
activity. Activity was not improved by the addition of an ortho flu-
oro substituent (16d) but this substitution pattern was tolerated.
The 2,6-dichlorophenyl analog (16e) was also tolerated but re-
mained 10 fold less potent than lead 10. Meta substitution with
a polar dimethyl amino group (16f) was not well tolerated. Simi-
larly, 3,4-dimethyl substitution resulted in analog 16g that was
poorly active.
With the optimal aryl ether substitution established, the scope
of serine substitution was explored (Table 2). To better differenti-
ate this set of compounds, cellular potency and certain ADME char-
acteristics were determined. The cellular readout used was the
quantitation of ganglioside GM1 on the surface of A549 cells.19 De-
creased glucosylceramide synthesis would lead to decreased GM1
synthesis. CYP3A4 inhibition was also measured by incubating
compounds with human liver microsomes and observing their
ability to impede oxidative metabolism of testosterone.20 Incuba-
tion with mouse liver microsomes was performed to gauge com-
pound susceptibility to oxidative degradation.21
Modification of the pyridine ring could be accomplished
without a significant negative impact on potency. For example,

6776
E. Koltun et al. / Bioorg. Med. Chem. Lett. 21 (2011) 6773–6777
Table 1
Table 2
Aryl ether analogs (16)
Serine analogs (18)
R1
O
H
N
R2
N
O
O
N
Cl
H
N
N
N
O
O
N
R3
O
R
N
O
O
R4
R5
CF3
R7
Compound
R
GCS
IC50
(nM)
A549
CYP3A4
IC50
(nM)
Mouse liver
microsome
stability (%
conversion)
a
R6
Cell
IC50
b,c
a
Compound
R1
R2
R3
R4
R5
R6
R7
GCS IC50 (nM)
(nM)
10
H
H
H
H
H
H
H
H
H
Br
F
H
H
H
H
H
H
H
H
Me
H
H
H
H
H
H
H
H
H
H
F
Cl
H
H
H
H
H
Cl
Cl
Me
Cl
H
H
H
H
H
H
NMe2
Me
H
H
H
H
H
Cl
Me
H
F
F
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
Cl
H
H
H
H
H
H
H
H
H
32
53
16a
16b
16c
16d
16e
16f
16g
16h
16i
16j
16k
16l
16m
16n
16n
18a
1
2
13,500
3,300
50
85
1,150
300
230
340
1,075
1,080
90
110
180
3
H
H
1
7
0.3
13
Me
Cl
Cl
Cl
Cl
Cl
Cl
CF3
O
F
O
F
H
H
H
2
2
1
18b
6,300
75
H
H
a
Average value of duplicate experiments reported.
O
18c
18d
20
50
na
>20,000
2,000
60
60
N
Benzyl serine 16n was found to be as potent in cells as it was
biochemically. It had a minimal CYP3A4 liability and was modestly
stable. Dihydrobenzodioxine analog 18a maintained biochemical
and cellular potency relative to 16n but was less stable to micro-
somal oxidation. Compound 18a was also a more potent CYP3A4
inhibitor. 4-Fluorobenzyl analog 18b was about half a log less po-
tent than 16n and had a less desirable ADME profile.
275
N
N
18e
50
na
11,000
75
In an effort to find benzyl group replacements, heterocycles
were investigated. This is because earlier work in this area estab-
lished the fact that the serine had to be alkylated in order to main-
tain activity. While the 2-methyl oxazole derivative (18c) was
tolerated, it was significantly less potent than 16n. It had no mea-
surable CYP3A4 liability but was only modestly stable in mouse
microsomes. 4-pyridyl derivative 18d was almost 300 fold less po-
tent than 16n with a significant CYP3A4 liability and comparable
metabolic stability. Interestingly, the addition of a 3-chloro group
(18e) significantly improved potency and decreased the CYP3A4
liability without increasing metabolic stability. Of the aliphatic ser-
ine analogs prepared, tert-butyl ether 18f was also significantly
less potent than lead 16n with a similar ADME profile. Cyclopro-
pylmethyl ether 18g was not a considerable improvement over
18f. Thus it was determined that the unsubstituted benzyl serine
moiety provided the best balance of GCS inhibitory potency and
desirable ADME properties.
The final focus of SAR development on this scaffold was the wes-
tern amide portion. A summary of this work is shown in Table 3.
There was very little flexibility in this portion of the scaffold. Non
basic amides were generally not tolerated (data not shown). Many
active amides which did contain a basic amine had undesirable
ADME profiles. Compounds 23a and 23b are examples of this phe-
nomenon. While the amino pyrrolidide 23b was 3 fold more potent
than amino azetidide 23a, both were metabolically unstable. We
speculated that the low metabolic stability of these compounds
was due to N-dealkylation. Screening of the mono N-methyl amino
pyrrolidide diastereomers, 23c and 23d, supported our hypothesis
in that both were significantly more metabolically stable. Fortu-
nately, 23c also had excellent cell potency and only a modest
CYP3A4 liability. Further efforts to identify basic amides which
Cl
18f
120
50
na
na
12,000
50
50
18g
>20,000
a
b
c
Average value of duplicate experiments reported.
Average value of triplicate experiments reported.
Na = data not available.
might be metabolically stable led us to amino-quinuclidide diaste-
reomers 23e and 23f. Both of these compounds were potent
biochemically and in cells, inactive against CYP3A4 and had modest
metabolic stability. The identification of compound 23c demon-
strated that it was possible to achieve a desirable compound profile
by tuning the western amide portion of the scaffold.
During the course of this work, 23c emerged as a compound
that combined potent in vitro GCS inhibition with acceptable
ADME parameters. When 23c was dosed in mice, significant levels
of the compound were measured in plasma, liver, fat and muscle
4 h post a 30 mg/kg oral dose (Table 4).
In summary, HTS efforts identified a unique GCS inhibitor (10).
That hit was elaborated into lead 23c which had improved cellular
potency and ADME properties. Furthermore, 23c has a comparable
biochemical profile to Genz-112638. The distribution to metaboli-
cally active tissues, especially liver, was desirable given our inter-
est in diabetes therapy. As such, 23c represents a novel GCS
inhibitor that might be an additional valuable tool for measuring
the effects of GCS inhibition in vitro and in vivo.

E. Koltun et al. / Bioorg. Med. Chem. Lett. 21 (2011) 6773–6777
6777
Table 3
Amide analogs (23)
our colleagues in the compound repository, Eliana Bustamante,
Dorothy Trent, Melanie Dimapasoc, Yelena Zherebina, Shaun
Nguyen, and Josh Rulloda, for their support of this project.
O
O
H
N
N
R
Cl
References and notes
O
O
1. For an excellent recent review of glycosphingolipid biology, see: Wennekes, T.;
van den Berg, R.; Boot, R.; van der Marel, G.; Overkleeft, H.; Aerts, J. Angew.
Chem., Int. Ed. 2009, 48, 8848.
CF3
2. Butters, T.; Mellor, H.; Narita, K.; Dwek, R.; Platt, F. Phil. Trans. R. Soc. Lond. B
2003, 358, 927.
3. Tagami, S.; Inokuchi, J.; Kabayama, K.; Yoshimura, H.; Kitamura, F.; Uemura, S.;
Ogawa, C.; Ishii, A.; Saito, M.; Ohtsuka, Y.; Sakaue, S.; Igarashi, Y. J. Biol. Chem.
2002, 277, 3085.
Compound
R
GCS
IC50
(nM) IC50
A549
Cell
CYP3A4
IC50
(nM)
Mouse liver
microsome
stability (%
conversion)
a
b,c
(nM)
4. (a) Bleicher, R.; Cabot, M. Biochim. Biophys. Acta 2002, 1585, 172; (b) Kok, J.;
Sietsma, H. Curr. Drug Targets 2004, 5, 375.
5. Summers, S. Prog. Lipid Res. 2006, 45, 42.
N
6. Karman, J.; Tedstone, J.; Gumlaw, N.; Zhu, Y.; Yew, N.; Siegel, C.; Guo, S.;
Siwkowski, A.; Ruzek, M.; Jiang, C.; Cheng, S. Int. Immunol. 2010, 22, 593.
7. Giraldo, P.; Alfonso, P.; Atutxa, K.; Fernández-Galán, M.; Barez, A.; Franco, R.;
Alonso, D.; Martin, A.; Latre, P.; Pocovi, M. Haematologica 2009, 94, 1771.
8. Aerts, J.; Ottenhoff, R.; Powlson, A.; Grefhorst, A.; van Eijk, M.; Dubbelhuis, P.;
Aten, J.; Kuipers, F.; Serlie, M.; Wennekes, T.; Sethi, J.; O’Rahilly, S.; Overkleeft,
H. Diabetes 2007, 56, 1341.
9. Wennekes, T.; Meijer, A.; Groen, A.; Boot, R.; Groener, J.; van Eijk, M.; Ottenhoff,
R.; Bijl, N.; Ghauharali, K.; Song, H.; O’Shea, T.; Liu, H.; Yew, N.; Copeland, D.;
van den Berg, R.; van der Marel, G.; Overkleeft, H.; Aerts, J. J. Med. Chem. 2010,
53, 689.
23a
40
na
5,400
90
N
N
23b
23c
23d
N
10
16
65
na
5
7,200
12,000
9,000
90
20
40
N
HN
HN
N
na
10. Actelion Smadar trial registry. http://trials.actelion.com/asp/Trial_Registry/
ListStudy.asp?PN=ACT-149071 (accessed September 2011).
11. Wennekes, T.; Lang, B.; Leeman, M.; van der Marel, G.; Smits, E.; Weber, M.; van
Wiltenburg, J.; Wolberg, M.; Aerts, J.; Overkleeft, H. Org. Process Res. Dev. 2008,
12, 414.
23e
23f
5
7
4
>20,000
>20,000
40
50
N
H
N
N
12. Zhao, H.; Przybylska, M.; Wu, I.; Zhang, J.; Siegel, C.; Komarnitsky, S.; Yew, N.;
Cheng, S. Diabetes 2007, 56, 1210.
12
N
H
13. Peterschmitt, M.; Burke, A.; Blankstein, L.; Smith, S.; Puga, A.; Kramer, W.;
Harris, J.; Mathews, D.; Bonate, P. J. Clin. Pharmacol. 2011, 51, 695.
14. McEachern, K.; Fung, J.; Komarnitsky, S.; Siegel, C.; Chuang, W.; Hutto, E.;
Shayman, J.; Grabowski, G.; Aerts, J.; Cheng, S.; Copeland, D.; Marshall, J. Mol.
Genet. Metab. 2007, 91, 259.
a
b
c
Average value of duplicate experiments reported.
Average value of triplicate experiments reported.
Na = data not available
15. Sugano, H.; Miyoshi, M. J. Org. Chem 1976, 41, 2352.
16. (a) Xu, J. Tetrahedron: Asymmetry 2002, 13, 1129; (b) Giles, P.; Rogers-Evans,
M.; Soukup, M.; Knight, J. Org. Process Res. Dev. 2003, 7, 22.
17. Lee, T.; Jones, B. Terahedron 1995, 51, 7331.
18. For an example of this transformation using AcCl in MeOH, see Nudelman, A.;
Bechor, Y.; Falb, E.; Fischer, B.; Wexler, B.; Nudelman, A. Synthetic Commun.
1998, 28, 471.
19. A549 cells were incubated in the presence of compounds for 72 h. The cells
were then stained with fluorescent cholera toxin subunit B (FL-CTB) that binds
to ganglioside GM1 on the cell membrane. After fixing the cells with
formaldehyde and Hoechst stain, they were analyzed using an automated
fluorescent microscope to quantify the amount of GM1 remaining on the cell
membrane.
Table 4
Mouse tissue exposure for 23c
Compound
Plasmaa
2
(lM)
Muscle (
l
M)
Fat (
lM)
Liver (
20
lM)
23c
7
2
a
C57BL6 mouse (n = 3), 30 mpk, PO, data collected 4 h post dose.
20. Compounds were incubated for 45 min with human liver microsomes
Acknowledgments
(0.25 mg/mL) in the presence of NADPH (1 mmol) and testosterone (20
Substrate conversion was monitored by MS.
lM).
21. Compounds (15
lM) were incubated for 30 min with mouse liver microsomes
The authors are grateful to Sean Wu, Lory Tan, and John Bui for
collecting the reported ADME data. We would also like to thank
Ron Aoyama, Michael Yakes, Les Bornheim and Artur Plonowski for
obtaining the mouse PK data. Finally, we would like to acknowledge
(0.5 mg/mL). Compound depletion was monitored by MS.
22. Biochemical and cell data reported here were obtained in-house according to
the text and note 19. The compounds tested were prepared at Exelixis or
partner CRO.