Structure-activity relationships of diamine inhibitors of cytochrome P450 (CYP) 3A as novel pharmacoenhancers. Part II: P2/P3 region and discovery of cobicistat (GS-9350)

Article abstract of DOI:10.1016/j.bmcl.2013.12.057

The HIV protease inhibitor (PI) ritonavir (RTV) has been widely used as a pharmacoenhancer for other PIs, which are substrates of cytochrome P450 3A (CYP3A). However the potent anti-HIV activity of ritonavir may limit its use as a pharmacoenhancer with other classes of anti-HIV agents. Ritonavir is also associated with limitations such as poor physicochemical properties. To address these issues a series of compounds with replacements at the P2 and/or P3 region was designed and evaluated as novel CYP3A inhibitors. Through these efforts, a potent and selective inhibitor of CYP3A, GS-9350 (cobicistat) with improved physiochemical properties was discovered.

Full text of DOI:10.1016/j.bmcl.2013.12.057

Bioorganic & Medicinal Chemistry Letters xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Structure–activity relationships of diamine inhibitors of cytochrome
P450 (CYP) 3A as novel pharmacoenhancers. Part II: P2/P3 region and
discovery of cobicistat (GS-9350)
a
a,
Lianhong Xu a, , Hongtao Liu , Allen Hong , Randy Vivian a, Bernard P. Murray b, Christian Callebaut c,
You-Chul Choi a, , Melody S. Lee b, , Jennifer Chau b, Luong K. Tsai c, Kirsten M. Stray c, Robert G. Strickley d,
Jianhong Wang b, Leah Tong b, Swami Swaminathan e, Gerry R. Rhodes b, Manoj C. Desai a
⇑
a Department of Medicinal Chemistry, Gilead Sciences, 333 Lakeside Drive, Foster City, CA 94303, USA
b Department of Drug Metabolism, Gilead Sciences, 333 Lakeside Drive, Foster City, CA 94303, USA
c Department of Biology, Gilead Sciences, 333 Lakeside Drive, Foster City, CA 94303, USA
d Department of Formulation, Gilead Sciences, 333 Lakeside Drive, Foster City, CA 94303, USA
e Department of Structural Chemistry, Gilead Sciences, 333 Lakeside Drive, Foster City, CA 94303, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Available online xxxx
The HIV protease inhibitor (PI) ritonavir (RTV) has been widely used as a pharmacoenhancer for other PIs,
which are substrates of cytochrome P450 3A (CYP3A). However the potent anti-HIV activity of ritonavir
may limit its use as a pharmacoenhancer with other classes of anti-HIV agents. Ritonavir is also associ-
ated with limitations such as poor physicochemical properties. To address these issues a series of com-
pounds with replacements at the P2 and/or P3 region was designed and evaluated as novel CYP3A
inhibitors. Through these efforts, a potent and selective inhibitor of CYP3A, GS-9350 (cobicistat) with
improved physiochemical properties was discovered.
Keywords:
CYP3A inhibitors
1,4-Diamines
Pharmacoenhancer
Selectivity against different CYP enzymes
Ó 2013 Elsevier Ltd. All rights reserved.
Cytochrome P450 (CYP) 3A4 is the most abundant human CYP
enzyme in the liver and plays a key role in both the detoxification
of xenobiotics and the metabolism of signaling molecules.1–4 Most
importantly, it metabolizes >50% of all approved drugs and, in
many cases, limits their ability to achieve sufficient steady-state
minimum drug concentration required for durable efficacy. In par-
ticular, all approved HIV protease inhibitors (PIs) are metabolized
primarily by CYP3A4 and the related enzyme, CYP3A5, in the liver
and intestine and thus, have unfavorable pharmacokinetic profiles,
including poor and/or variable oral bioavailability and relatively
short plasma elimination half-lives.5 As a result, early PI-contain-
ing regimens were complex—they required frequent dosing, had
high pill burdens, and were associated with significant side effects
and undesirable toxicities. Utilization of ritonavir (RTV, 14), an ap-
proved PI (antiviral therapeutic dose 600 mg twice daily) and also a
potent mechanism-based inhibitor of CYP3A, as a pharmacoenh-
ancer at subtherapeutic doses (100 or 200 mg daily) results in im-
proved pharmacokinetics (PK) of concomitantly dosed PIs, thereby
improving effectiveness while simplifying the regimens. This
unique clinical practice that reduces the pill burden and enhances
the adherence of the coadministered PIs is a cornerstone of PI-con-
core
P2
R3
Ph
O
O
O
O
H
N
R1
S
N
O
O
N
N
H
N
H
R2
O
P3
P3
Ph
1a
P2
R3
core
Ph
H
N
R1
S
N
N
N
H
N
H
R2
O
Ph
1b
Figure 1. Structure of compound 1a and compound 1b.
taining highly active antiretroviral therapy (HAART).5 Ritonavir
therefore has served a critical function in the development of PI-
based HAART regimens and in the chronic management of HIV
infection. However, as discussed in the preceding Letter,6 RTV
has many limitations when used as a pharmacoenhancer, so we
set out to identify a novel pharmacoenhancer that can maintain
⇑
Corresponding author. Tel.: +1 650 522 5828; fax: +1 650 522 5899.
E-mail address: lianhong.xu@gilead.com (L. Xu).
effectiveness as
limitations.
a CYP3A inhibitor while minimizing RTV’s
These authors are no longer with Gilead Sciences.
0960-894X/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmcl.2013.12.057
Please cite this article in press as: Xu, L.; et al. Bioorg. Med. Chem. Lett. (2014), http://dx.doi.org/10.1016/j.bmcl.2013.12.057

2
L. Xu et al. / Bioorg. Med. Chem. Lett. xxx (2014) xxx–xxx
R3
Ph
O
Ph
O
H
N
b
S
N
H2N
Ph
S
N
O
O
O
N
H
BocHN
COOH
N
H
R3
O
2a
Ph
3a
BocHN
O
4
R3
R3
c
O
R1
OH
a
COOMe
b
N
N
H
6
H2N
R2
O
O
R3
Ph
5
O
H
N
R1
S
N
O
N
N
H
N
H
R2
O
Ph
1a
Scheme 1. Synthesis of compound 1a. Reagents and conditions: (a) (1) CDI/iPr2NEt; (2) R1R2NH; 3. NaOH/CH3OH; (b) EDCI, HOBt, iPr2NEt; (c) (1) TFA or HCl/dioxane; (2) CDI,
iPr2NEt; (3) R1R2NH.
In the preceding letter, we identified R,R- and S,S-1,4-diamine
cores (exemplified by 1a and 1b Fig. 1) as our optimal central
pieces. We herein report the synthesis and exploration of P2 and
P3 regions of the diamine CYP3A inhibitors; these investigations
led to the discovery of a novel pharmacoenhancer cobicistat
(compound 35, GS-9350).
simultaneously studied the SAR at the P3 and P2 portions of the
lead compound 15.
To establish SAR on P2/P3, our initial goal was to use modifica-
tions at P3 and/or P2 portion to dissociate the anti-HIV activity
from the inhibitory potency against CYP3A possessed by the lead
compound. Combination of the optimized P2 and P3 moieties with
the best core would then allow us to identify an analog that is ac-
tive against CYP3A but devoid of anti-HIV activity. The results are
summarized in Table 1.
A general sequence for the synthesis of compound 1a analogs is
outlined in Scheme 1. Compounds with formula 1b were prepared
using the same procedure except replacing the R,R-core of 2a with
the corresponding S,S-core. Compound 1a can be constructed using
two different routes. Amine 2a6 was coupled with Boc-protected
amino acid 4 to give amide 3a. Removal of the Boc-group yielded
the corresponding amine, which was then acetylated with one
equivalent CDI and subsequently reacted with amine R1R2NH to af-
ford 1a. Alternatively, compound 1a can be prepared through cou-
pling of acid 6, which already contains a urea moiety, with amine
2a. Acid 6 was prepared by treatment of protected amino acid ester
5 with CDI followed by R1R2NH. Subsequent ester hydrolysis gen-
erates acid 6 which is then coupled to 2a using standard condi-
tions. The amino acid 4 or 5 with various R3 side chains were
either commercially available in protected forms suitable for use
in the described sequences, or prepared following conventional
methods for synthesis of protected amino acids.
Reducing the size or removing the 20-isopropyl group of the P3
portion 4-thiazolyl moiety (17, 18 vs 15), in attempts to minimize
interaction of the analogs with the HIV protease, only marginally
decreased the binding with HIV protease and antiviral activity.
Similarly, replacing the N-methyl with N-cPr (16 vs 15), aiming
to interrupt the interactions of the compound with HIV protease,
resulted in only minor reduction of anti-HIV activity. Compounds
16–18 with modified P3 possess comparable activity towards
CYP3A. Combining these two modifications provided compound
19 which had a ꢀ14-fold reduction of inhibitory activity against
HIV yet maintained potency against CYP3A. This represents an in-
crease in selectivity of approximately 80-fold compared to RTV.
Compound 19 now has clinically marginal anti-HIV EC50 of 4 lM,
but we desired a compound inactive against HIV. Using 5-thiazole
instead 4-thiazole at P3 (20 vs 18) did not offer any advantage in
terms of selectivity, but simplified synthesis, as it can be obtained
from the same intermediate as the right-side (P20) thiazole.
We then turned our attention to the P2 portion of the molecule.
Replacing valine with b-alanine at P2 (21) provided ꢀ8-fold reduc-
tion of anti-HIV activity to give a compound with micromolar
activity against HIV. Replacement of the valine at the 2-position
with a serine, containing a more polar moiety, offered 5-fold reduc-
tion in activity against HIV. With serine at P2, the SAR trend at the
P3 portion was consistent with that observed in the P2 valine ser-
ies (compounds 22–25 vs 15–19). Compound 25 has an EC50 of
As shown in Scheme 2, 40- or 50-thiazolyl methyl amines
(R1R2NH) 7 or 8 were synthesized from 4- or 5-chloromethyl
thiazoles 11 or 13, respectively, by treatment with primary amines.
2-Substituted 4-chloromethyl thiazole 11 was prepared from con-
densation of thioamide 9 with 1,3-dichloroacetone 10.7 5-Chloro-
methyl thiazole 13 was obtained from commercial available
5-hydroxymethyl thiazole 12 by reacting with methanesulfonyl
chloride.
As discussed in the previous communication,6 we identified
desoxy-RTV (15) as
a lead compound for structure-activity
relationship (SAR) studies. Although the antiviral potency of des-
oxy-RTV is greatly reduced compared to RTV, it still possesses
modest anti-HIV activity with an EC50 of 290 nM (Table 1). While
building our understanding of the SAR for the diamine core, we
20 lM and shows over 400-fold improvement of selectivity com-
pared to RTV. Further exploration at the P2 site showed that polar
and bulky moieties can be tolerated by CYP3A but are less well tol-
erated by HIV protease (compounds 27–29). The fact that that in-
creased steric bulk at P2 can significantly reduce inhibitory
activity against HIV protease and virus replication is exemplified
by comparing compound 27 to 26. The selectivity and activity of
compound 27 met the minimum criterion for a selective CYP3A
inhibitor we set at the outset of the program. However the physi-
cochemical properties of compound 27, especially its aqueous sol-
ubility, were not desirable.
Cl
R2
Cl
Cl
S
S
N
H
R'
NH2
+
a
c
O
N
N
S
S
R'
R'
Cl
9
10
11
7
8
R2
N
H
c
N
OH
N
N
b
S
S
12
After understanding the SAR for disassociating the HIV activity
from CYP inhibitory potency, our next goal was to improve other
properties including physicochemical properties. We thus incorpo-
rated moieties with potential to provide analogs with desirable
13
Scheme 2. Synthesis of compounds 7 and 8. Reagents and conditions: (a) MgSO4/
acetone/reflux. (b) MsCl/Et3N/MeCN; (c) R2NH2.
Please cite this article in press as: Xu, L.; et al. Bioorg. Med. Chem. Lett. (2014), http://dx.doi.org/10.1016/j.bmcl.2013.12.057

L. Xu et al. / Bioorg. Med. Chem. Lett. xxx (2014) xxx–xxx
3
Table 1
Evaluation of pharmacoenhancers
core
core
P2
R3
P2
R3
Ph
Ph
O
O
O
O
H
N
H
N
R1
R1
S
N
S
N
O
O
N
N
H
N
H
N
N
H
N
H
R2
O
R2
O
P3
P3
Ph
R3
Ph
R1
N
Compound
Core
HIV protease IC50
(
lM)
HIV replication EC50
(lM)
CYP3Aa IC50
(lM)
R2
Ph
O
O
H
N
S
N
N
S
14 Ritonavir
0.0006
0.05
0.11
N
N
H
N
H
O
O
OH
Ph
N
N
15
16
17
18
19
R,R
R,R
R,R
R,R
R,R
0.005
0.55
0.29
0.75
0.4
0.8
4
0.13
0.21
0.1
S
N
N
S
N
S
N
0.035
0.242
3.59
N
S
N
N
N
0.14
<0.1
N
S
S
N
20
21
22
R,R
R,R
R,R
0.386
ND
0.7
2.5
1.5
0.08
0.13
0.15
OH
N
N
0.05
S
N
N
23
24
25
26
27
R,R
R,R
R,R
R,R
R,R
1.41
4.3
2.5
12
0.11
0.17
0.12
0.15
0.17
S
S
N
N
N
S
N
19.3
ND
20
OCH3
OBn
N
N
N
0.65
>30
S
N
11.8
S
28
R,R
0.01
10
0.50
OH
O
O
29
R,R
R,R
0.143
ND
1
3
0.13
0.14
NH2
30b
N
(continued on next page)
Please cite this article in press as: Xu, L.; et al. Bioorg. Med. Chem. Lett. (2014), http://dx.doi.org/10.1016/j.bmcl.2013.12.057

4
L. Xu et al. / Bioorg. Med. Chem. Lett. xxx (2014) xxx–xxx
Table 1 (continued)
R3
R1
N
Compound
Core
HIV protease IC50
(
lM)
HIV replication EC50
(lM)
CYP3Aa IC50
(lM)
R2
31
32
33
R,R
R,R
R,R
0.039
>30
2.7
0.14
0.21
0.13
OH
>30
>30
NH2
>30
NH
N
34
35
36
R,R
R,R
R,R
>30
>30
>30
20
0.14
0.15
0.15
>30
>30
N
O
N
O
NH2
37
38
R,R
R,R
>30
2
>30
18
0.12
0.15
NHAc
NHMs
39
40
R,R
R,R
8
3
>30
>10
0.12
0.12
NH
N
N
N
41
42
R,R
R,R
>30
>30
5
0.11
0.10
0.03
N
N
OH
N
N
43
44
45
S,S
S,S
S,S
>30
>30
>30
>30
>30
>30
0.14
0.12
0.12
S
NHAc
N
S
N
N
OH
S
N
46
47
S,S
S,S
>30
>30
>30
>30
0.08
0.09
N
N
S
NH
N
ND = not determined.
a
Compounds are mechanism-based CYP3A inhibitors so IC50 values are representative and were all determined under the same conditions; Inhibitory potency against
midazolam 10-hydroxylase activity catalyzed by CYP3A.9
b
Single compound but the absolute stereochemistry at P2 was not assigned.
physicochemical properties (compounds 30–34). We were grati-
fied to discover that introduction of basic moieties at the P2 site
(compound 32 vs 31) can significantly reduce the binding with
HIV protease while maintaining activity against CYP and conferring
greatly improved aqueous solubility. However, further detailed
profiling of compound 32 indicated that it has poor permeability
resulting in minimal oral bioavailability. Modification of the P2 of
32 to a secondary or tertiary amine moiety, such as methyl amine
(33), dimethyl amine (34) or morpholine (35) can simultaneously
offer much improved permeability while maintaining the desired
selectivity profile. Compound 35 possesses both appropriate basi-
city and bulkiness at P2, therefore has the best overall selectivity
against HIV. In addition, it maintained the favorable physicochem-
ical properties of compound 32. A broader group of analogs with
Please cite this article in press as: Xu, L.; et al. Bioorg. Med. Chem. Lett. (2014), http://dx.doi.org/10.1016/j.bmcl.2013.12.057

L. Xu et al. / Bioorg. Med. Chem. Lett. xxx (2014) xxx–xxx
5
Table 2
Compound 35 was advanced to clinical studies. It was investi-
gated in humans compared to RTV as a pharmacoenhancer for
drugs metabolized by CYP3A to improve their systemic exposure.
In a randomized, placebo-controlled, double-blind, multicenter,
48-week Phase 3 study, cobicistat (35) demonstrated comparable
efficacy to RTV as a pharmacoenhancer for atazanavir. The study
found that an HIV regimen containing a cobicistat-boosted PI
was non-inferior to a regimen containing a ritonavir-boosted PI
at 48 weeks of therapy.10 Currently, cobicistat is a component of
an approved fixed-dose combination StribildÒ in which it acts as
a pharmacoenhancer for the HIV integrase strand transfer inhibi-
tor, elvitegravir. It is also under regulatory review in USA as an
individual agent. If approved, cobicistat may be an effective option
for boosting the pharmacokinetics in HIV regimens that are based
on PIs.
In summary, using desoxy-RTV as the lead, extensive studies on
core and P2/P3 regions lead to the discovery of cobicistat (35, GS-
9350). Cobicistat is a novel, selective and potent CYP3A inhibitor
without any detectable antiviral activity and with desirable phys-
icochemical properties. It is expected cobicistat will have broader
applications in HIV patients compared to RTV once it is approved
by regulatory agencies. In addition, cobicistat is expected to have
broader utility compared to RTV as a pharmacoenhancer in treat-
ment of other life-threatening diseases, such as anti-HCV, given
its more desirable profile.
Evaluation of selectivity of pharmacoenhancers
Compound
CYP enzyme IC50
(
l
M)
PXR activation %9
3A
1A2
2C19
2C9
2D6
RTV
32
35
36
41
43
44
47
0.11
0.11
0.15
0.15
0.11
0.16
0.12
0.09
>25
>25
>25
>25
>25
>5
12.7
>25
>25
1.3
4.9
>25
>25
8
4.7
>25
15.5
4.7
2.3
0.8
9.2
3.1
0.6
1.3
0.8
0.4
51
5
10
32
15
23
15
15
2.2
>25
>25
2.2
>25
15
neutral and weakly basic to basic P2 moieties were also evaluated.
Consistent with our earlier findings, basic and/or bulky P2 groups
have significantly reduced activity against HIV (compounds 36–
41), while P2 groups that are weakly basic and/or less sterically
demanding retain some residual anti-HIV activity (compound
42). Incorporation of moieties possessing both basicity and bulki-
ness can generate analogs devoid of anti-HIV activity (35–37,
41). As discussed in our previous Letter,6 incorporation of a S,S-
diamine core can afford favorable selectivity between anti-HIV
potency and CYP potency. Such S,S-analogs containing P2 modifica-
tions with desirable CYP activity identified in R,R-series were pre-
pared (compounds 43–47). Consistent with previous observations,
these analogs maintain CYP3A inhibitory potency and have no
anti-HIV inhibitory activity.
Acknowledgements
Analogs that are devoid of anti-HIV activity and possess desired
inhibitory activity against CYP3A were further profiled for their
selectivity against the most important human drug metabolizing
CYP enzymes and for their potential to activate PXR (Table 2).8
As discussed in the preceding Letter, minimal PXR activation is
desirable to avoid additional potential for drug-drug interactions
and reducing the effectiveness of the pharmacoenhancer in vivo.
Compounds 32, 43 and 44 demonstrated generally favorable
CYP selectivity and less induction potential than RTV but had
The authors would like to thank contributions of the work and
manuscript from Chris Yang, Yujing Wang, Kelly Wang, Jaclyn
Hayes, Eugene Eisenberg, Quynh Iwata, Carina Cannizzaro, Ke-Yu
Wang and Tomas Cihlar.
References and notes
1. Nelson, D. R.; Koymans, L.; Kamataki, T., et al Pharmacogenetics 1996, 6, 1.
2. Rendic, S.; Di Carlo, F. J. Drug Metab. Rev 1997, 29, 413.
3. Wrighton, S. A. et al Drug Metab. Rev. 2000, 32, 339.
IC50 values for CYP2D6 of ꢀ1
lM. Overall, compound 35 exhibits
4. Rendic, S. Drug Metab. Rev. 2002, 34, 83.
the best selectivity compared with RTV (Table 2). It has no signif-
icant activity against the tested CYP enzymes at concentrations
likely to be achieved clinically, and also has minimal induction po-
tential, indicating that it may have a lower potential for causing
undesired drug-drug interactions compared to RTV.
5. For reviews on boosted PI see following reviews and references therein: (a)
Gerber, J. G. Clin. Infect. Dis. 2000, 30, S123; (b) Moyle, G. J.; Back, D. HIV Med.
2001, 2, 105; (c) Cooper, C. L.; van Heeswijk, R. P. G.; Gallicano, K.; Cameron, D.
W. Clin. Infect. Dis. 2003, 36, 1585; (d) Becker, S. L. Expert Opin. Invest. Drugs
2003, 12, 401; (e) Gallant, J. E. AIDS Rev. 2004, 6, 226; (f) Youle, M. J. Antimicrob.
Chemother. 2007, 60, 1195; (g) Busse, K. H.; Penzak, S. R. Expert Rev. Clin.
Pharmacol. 2008, 1, 533; (h) Xu, L.; Desai, M. C. Curr. Opin. Invest. Drugs 2009,
10, 775.
6. Liu, H.; Xu, L.; Hui, H.; Vivian, R.; Callebaut, C.; Murray, B. P.; Hong, A.; Lee, M.
S.; Tsai, L. K.; Chau, J.; Stray, K.; Cannizzaro, C.; Choi, Y.-C.; Rhodes, G. R.; Desai,
M. C. Bioorg. Med. Chem. Lett. 2013. http://dx.doi.org/10.1016/j.bmcl.
2013.12.058.
7. Kempf, D. J.; Sham, H. L.; Marsh, K. C.; Flentge, C. A.; Betebenner, D., et al J. Med.
Chem. 1998, 41, 602.
8. (a) Kharasch, E. D.; Mitchell, D.; Coles, R.; Blanco, R. Antimicrob. Agents
Chemother. 2008, 52, 1663; (b) Foisy, M. M.; Yakiwchuk, E. M.; Hughes, C. A.
Ann. Pharmacother. 2008, 42, 1048.
Compound 35 (GS-9350, cobicistat) was further profiled in vivo.
Results indicated that the absorption of 35 is above 50% in all non-
clinical species, comparable to that of RTV.9 Compound 35 was
therefore expected to have high absorption potential in humans.
Extensive in vitro studies9 comparing compound 35 and RTV
have shown that 35 is a potent and selective CYP3A inhibitor that
lacks significant anti-HIV activity. Compound 35 shares RTV’s
property of potent inhibition of the metabolism of a wide range
of CYP3A substrates and retains the characteristic of mechanism-
based inhibition of CYP3A. In addition, compound 35 also has
favorable physicochemical properties, especially high aqueous sol-
ubility, allowing it to be formulated as a tablet and co-formulated
with other drugs.9
9. Xu, L.; Liu, H.; Murray, B. P.; Callebaut, C.; Lee, M. S.; Hong, A.; Strickley, R. G.;
Tsai, L. K.; Stray, K. M.; Wang, Y.; Rhodes, G. R.; Desai, M. C. ACS Med. Chem. Lett.
2010, 1, 209.
10. DeJesus, E.; Rockstroh, J.; Henry, K.; Molina, J.-M.; Gathe, J.; Ramanathan, S.;
Wei, X.; Yale, K.; Szwarcberg, J.; White, K.; Cheng, A.; Kearney, B., et al Lancet
2012, 379, 2429.
Please cite this article in press as: Xu, L.; et al. Bioorg. Med. Chem. Lett. (2014), http://dx.doi.org/10.1016/j.bmcl.2013.12.057