Tricyclic HIV integrase inhibitors: Potent and orally bioavailable C5-aza analogs

Article abstract of DOI:10.1016/j.bmcl.2008.01.018

A series of C5-aza tricyclic HIV integrase inhibitors was prepared. A highly potent and orally bioavailable compound (compound 9) was identified and selected for development.

Full text of DOI:10.1016/j.bmcl.2008.01.018

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry Letters 18 (2008) 1388–1391
Tricyclic HIV integrase inhibitors: Potent and orally
bioavailable C5-aza analogs
Haolun Jin,* Matthew Wright, Richard Pastor, Michael Mish, Sammy Metobo,
Salman Jabri, Rachael Lansdown, Ruby Cai, Peter Pyun, Manuel Tsiang,
Xiaowu Chen and Choung U. Kim
Department of Medicinal Chemistry, Gilead Sciences, Inc., 333 Lakeside Drive, Foster City, CA 94404, USA
Received 3 October 2007; revised 31 December 2007; accepted 3 January 2008
Available online 9 January 2008
Abstract—A series of C5-aza tricyclic HIV integrase inhibitors was prepared. A highly potent and orally bioavailable compound
(compound 9) was identified and selected for development.
ꢀ 2008 Elsevier Ltd. All rights reserved.
Human immunodeficiency virus (HIV), the causative
pathogen of AIDS, replicates utilizing three essential
enzymes encoded in the HIV pol gene, reverse transcrip-
tase (RT), protease (PR), and integrase (IN). While
many licensed anti-HIV agents target either RT or PR,
the recent progress in extensive researches by the phar-
maceutical industry has led to the emergence of promis-
ing IN inhibitors.1 The most advanced lead, MK-0518
(Raltegravir, 1) (Fig. 1), has been approved by the
FDA as the first HIV integrase inhibitor to be marketed
in the US.
Our previous publications described the design and SAR
studies of a few series of novel and highly organized tri-
cyclic HIV IN inhibitors shown as 2 (Fig. 1).2 Those
analogs possessed potent activity against both the HIV
integrase-catalyzed strand transfer activity in the enzy-
matic assay as well as HIV replication in cell culture as-
says. A representative compound (3), for example,
exhibited the good activity in the presence of either
10% FBS or human serum proteins (HSP) (Fig. 1).2a
An important objective for our HIV integrase inhibitor
research program is to identify an orally bioavailable
compound for clinical development. Its pharmacoki-
netic properties will allow for the simple dosing regi-
mens, particularly in the combination therapy, a
routine clinical practice in the AIDS medication. It
was observed that compound 3 showed exposure upon
oral dosing in both rat and dog, but its overall pharma-
cokinetic profiles were sub-optimal.3 In order to im-
prove the key pharmacokinetic parameters such as
half-life (t1/2) and the oral absorption of these potent
IN inhibitors, we extended the SAR studies beyond
those analogs reported previously. For the tricyclic scaf-
fold, a variety of C5 substituents were relatively well tol-
erated in preserving acceptable potency. These
substituents included ether, carbamate, sulfamate,
aryl, and carboxamide functionalities. Therefore, we
envisaged that further exploitation of effects by other
C5 substituents on both potency and pharmacokinetics
was possible because of the direct electronic and steric
influences of these substituents on the tricyclic
pharmacophore. In this communication, we report the
N
N
2
O
O
X
X=
F
HN
H
OMe
N
F
N
OCONMe2
OSO2NMe2
2,6-F-Ph
N
N
H
N
O
O
H
O
O
O
H
O
1 (MK-0518, Raltegravir))
3
( X=
)
N
O
IC50
EC50(10%FBS)
EC50(HSP)
27nM
50 nM
3.4nM
54 nM
6 nM
16nM
Figure 1. The structures of 1 (K-0518, Raltegravir), 2 and 3.
Keywords: AIDS; HIV; Integrase inhibitor; Tricyclic.
*
Corresponding author. Tel.: +1 650 522 5460; fax: +1 650 522
5169; e-mail: hjin@gilead.com
0960-894X/$ - see front matter ꢀ 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2008.01.018

H. Jin et al. / Bioorg. Med. Chem. Lett. 18 (2008) 1388–1391
1389
preparation of a series of C5-amino derivatives as potent
inhibitors of both HIV integrase-catalyzed strand
transfer activity and HIV infection in cell cultures. The
pharmacokinetic evaluation of a subset of this SAR
series is also described and has led to the identification
of a compound for development.
To examine the cyclization effect of the C5-aza group on
the tricyclic pharmacophore, the preparation of com-
pound 15 was undertaken and its synthesis is shown in
Scheme 2. Intermediate 5 was reacted with 3-chloro-
propansulfonyl chloride, followed by removal of the
TEOC protecting group. Upon treating with base to
form the 5-membered sultam and removal of DPM, 15
was obtained in good overall yield.
The preparation of these analogs is illustrated in
Schemes 1 and 2. Compound 4, described in our previ-
ous publication,2a was converted to TEOC- protected
C5 amine 5 via the Curtius rearrangement protocol cou-
pled with trapping isocyanate with TMS ethanol. Upon
methylation and removal of TEOC, analogs 8–12 were
obtained by treating C5 methylamine 7 with appropriate
reagents and deprotecting diphenylmethyl ether (DPM)
of the phenol.
All analogs prepared were tested for their activity in
both HIV integrase strand transfer assay as well as
anti-HIV infection in a cell culture assay. The data are
summarized in Table 1.
In general all the compounds showed good to excellent
potency in both the enzyme and the cell culture-based
assays. C5 Sulfonamide 9, sulfonylurea 11, and sultam
15 exhibited improved activity in the cell assay over 3.
Further, the superior potency of 9 and 11 was main-
tained when human serum proteins, serum albumin
and a-1 acidic glycoproteins, were added to reach phys-
iologic concentrations in the cell culture based assay.
Interestingly, the simple amide 8, urea 10, and oxalyla-
mide 12 were found to be less potent although 10 was
the most soluble among those tested. The potency of
C5 sultam analog 15 was more affected by the serum
proteins than either 9 or 11. The results that we observed
in this work are consistent with those reported by our
laboratory previously. Both C5 carbamate and sulfa-
mate groups on the same tricyclic pharmacophore were
more potent than other C5 substituents on the tricyclic
scaffold.2a,b However, C5 amine derived analogs are
more attractive due to their greater stability toward
hydrolysis as compared to the analogous carbamates
or sulfamates.
O
CO2H
R
TMS
N
O
N
c
a
N
N
N
O
OCH(Ph)2
O
OCH(Ph)2
F
F
4
R
5
6
H
Me
b
R
Ac
Ms
CONM
SO2NMe2
8
9
10
R
N
e
2
NH
11
d,e
N
N
N
O
N
O
OH
12
O
OCH(Ph)2
NMe2
O
F
F
The two most potent compounds, 9 and 11, along with
compound 15, were chosen for pharmacokinetic evalua-
tions in both rat and dog and the major pharmacoki-
netic parameters are shown in Table 2.
7
Scheme 1. Ac, acetyl; Ms, methansulfonyl. Reagents and conditions:
(a) phosphorazidic acid diphenyl ester, TEA, rt, 5 h, then TMS
ethanol, 60 ꢁC, 26 h, 70%; (b) NaH, MeI, DMF, 0 ꢁC–rt, 1 h, 80%; (c)
TBAF, rt, overnight, 84%; (d) for 8: NaH, DMF, acetyl chloride, 0 ꢁC,
2 h, 86%; for 9: pyridine, methansulfonyl chloride, rt, overnight, 85%;
for 10: ClCONMe2, TEA, DMAP, dichloromethane, rt, 12 h, 50%; for
11: 3-dimethylsulfamoyl-1-methyl-3H-imidazol-1-ium triflate, 120 ꢁC,
microwave, 90 min, 80%; for 11: 1—TEA, N,N-dimethylsulfamoyl
chloride, for 12: 1—TEA, methyl chlorooxoacetate, DCM, rt, 30 min;
2—dimethylamine, rt, 10 min, 76%; (e) TFA, triethylsilane, rt, 50–93%.
Both 9 and 11 were found to be orally bioavailable in
these two species. Notably higher oral bioavailability
Table 1. Integration strand transfer inhibition and anti-HIV prolifer-
ation assay results for compounds 8–12, 15a
Compound
IC50
EC50
(10% FBS)c
EC50
(HSP)d
Solubilitye
b
(lM)
8
9
100
28
47
nd
nd
11
Cl
1.7
10
11.4
nd
O
10
11
12
15
71
62
844
4.5
nd
2.8
S
O
N
O
S
R
3.1
31
8.4
nd
c
O
N
a
265
13
5
N
1.4
49
N
N
a Values are means of at least two experiments, given in nM. nd, not
N
O
OH
O
OCH(Ph)2
determined.
F
4a
F
b Ref.
c Ref.
.
.
R
4b
15
13 TEOC
d HSP, human serum proteins adjusted EC50, obtained by assaying
compounds in the presence of physiological concentrations of human
b
H
14
4c
serum albumin and AAG; see Ref. for details.
Scheme 2. Reagents and conditions: (a) NaH, DMF, 3-chloro-
propansulfonyl chloride, 0 ꢁC–rt, 0.5 h, 60%; (b) TBAF, THF, rt,
90%; (c) 1—NaH, DMF; 2—triethylsilane, TFA, rt, 15 min, 50%.
e Measured by dissolving the solid of testing compound in pH 7.3
aqueous media and determining the concentration by HPLC.

1390
H. Jin et al. / Bioorg. Med. Chem. Lett. 18 (2008) 1388–1391
Table 2. Pharmacokinetics of 9, 11, and 15 in rat and doga
Species
Rat
T1/2 (h)
Dog
T1/2 (h)
F (%)
CL (L/h/kg)
F (%)
CL (L/h/kg)
9
11
15
15
13
4
1.1
0.9
0.6
0.28
0.23
1.53
45
16
8.3
4.9
4.5
0.8
0.4
0.34
0.3
F (%), fraction absorbed upon oral dosing testing compounds as compared to iv dosing, calculated based on AUC from iv and po groups, expressed
as %; CL, total body clearance obtained from iv dosing groups.
a All compounds were dosed as free parent in a solution form (EtOH, PG, PEG400; and citric acid; pH 3.3 for iv and pH 2.2 for po); and t1/2 was
generated from the iv dosing group. The values were means of data obtained from samples of three animals in each study.
(45%) and longer half-life (4.9 h) than 11 and 15 in dog
were observed for compound 9. It is interesting to note
that although compound 15 exhibited similar clearance
in dog as both 9 and 11, its half-life was much shorter.
Acknowledgments
The authors are grateful to Xubin Zheng and Bing Lu
for analyzing samples in PK studies; Gregg Jones and
Fang Yu for carrying out biological assays and Vahid
Zia and co-workers in the formulation group for mea-
suring the solubility. The authors are also indebted to
Professor Arnold L. Rheingold and his co-workers of
Department of Chemistry and Biochemistry, University
of California at San Diego for determining the X-ray
crystal structure of compound 9.
We were intrigued by the fact that compound 9 is more
potent than MK-0518 in the presence of serum proteins.
The X-ray crystal structure of 9 was determined and is
shown in Figure 2.5 It is interesting to note that N of
the C5 methanesulfonamido group is an sp2 hybridized
center. N-methyl and methansulfonyl moieties are ori-
ented perpendicular to the tri-cyclic core plane. The
hydrophobic p-F-phenyl moiety is also in a similar
orientation versus the plane of the tri-cyclic core. The
observed fixed conformation is the same as the one
shown in our initial molecular modeling studies2a and
could be a preferred conformation at the active site
during the strand transfer of HIV integration process.
References and notes
1. Merck & Co., Press Release, September 5, 2007; for a recent
review: Makhija, M. T., Curr. Med. Chem. 2006, 13, 2429;
For recent presentations of clinical trials of leading HIV
integrase inhibitors: (a) Cooper, D., Steigbigel, R. et al.,
Results of BENCHMRK-1/2, a Phase III Study Evaluating
the Efficacy and Safety of MK-0518, a Novel HIV-1
Integrase Inhibitor, in Patients with Triple-class Resistant
Virus, Abstracts of Papers, # 105 a&bLB, 14th Conference
on Retroviruses and Opportunistic Infections, February
25–28, 2007; Los Angeles, CA; (b) Zolopa, A. et al., The
HIV Integrase Inhibitor GS-9137 Demonstrates Potent
ARV Activity in Treatment-experienced Patients, Abstracts
of Papers, # 143LB, 14th Conference on Retroviruses and
Opportunistic Infections, February 25–28, 2007; Los Ange-
les, CA.
2. (a) Jin, H.; Cai, R. Z.; Schacherer, L.; Jabri, S.; Tsiang, M.;
Fardis, M.; Chen, X.; Chen, J. M.; Kim, C. U. . Bioorg.
Med. Chem. Lett. 2006, 16, 3989, and the references cited
within; (b) Fardis, M.; Jin, H.; Jabri, S.; Cai, R. Z.; Mish,
M.; Tsiang, M.; Kim, C. U. Bioorg. Med. Chem. Lett. 2006,
16, 4031; (c) Metobo, S.; Jin, H.; Tsiang, M.; Kim, C. U.
Bioorg. Med. Chem. Lett. 2006, 16, 3985.
In conclusion, by building upon our previously pub-
lished SAR studies and by exploring a variety of C5-
amine derivatives, we have successfully identified even
more highly active HIV integrase inhibitors with good
oral bioavailability in pre-clinical PK studies. These
results further demonstrated that a design strategy that
rigidifies the pharmacophore of the IN inhibitor is
critical to the observed activity against both IN strand
transfer and HIV replication in cell culture assays.
Because of its favorable potency and PK properties,
compound 9 was selected for further work in pre-clinical
and clinical development.6
3. Eisenberg, G. unpublished communication.
4. (a) Strand transfer assay modified from a previous report
(Hazuda et al., Nucleic Acid Res. 1994, 22, 1121). Biotin-
ylated donor DNA was bound to Reacti-Bind High
Binding Capacity Streptavidin coated white plates. DIG-
tagged target DNA with anti-DIG antibody-conjugated
horseradish peroxidase detection was used.; (b) For anti-
viral assay, 50 ll of 2· test concentration of fivefold serially
diluted drug in culture medium was added to each well of a
96-well plate (9 concentrations) in triplicate. MT-2 cells
were infected with HIV-1 IIIB at an m.o.i. of 0.01 for 3 h.
Fifty microliters of infected cell suspension in culture
medium (ꢀ1.5 · 104 cells) was then added to each well
containing the drug dilutions. The plates are incubated at
37 ꢁC for 5 days. One hundred microliters of CellTiter-
GloTM Reagent (catalog # G7571, Promega Biosciences,
Inc., Madison, WI) was then added to each well. Cell lysis
Figure 2. The ORTEP representation of X-ray 3D crystal structure of
9. Nitrogen atoms are in purple, oxygen atoms in red, the sulfur in
yellow, and the fluorine in green.

H. Jin et al. / Bioorg. Med. Chem. Lett. 18 (2008) 1388–1391
1391
was allowed to complete by incubating at room tempera-
ture for 10 min. Chemiluminescence was then read. For the
cytotoxicity assay, the protocol is identical to that of the
antiviral assay, except that uninfected cells and a threefold
serial dilution of drugs were used.; (c) The effect of
compounds binding to serum protein components was
evaluated by determining the antiviral EC50 in MT-2 cells in
10% FBS in the presence or absence of serum concentra-
tions of HSA (35 mg/ml) or a1-AGP (1.5 mg/ml). From the
EC50 data in the presence of each individual protein, the
EC50 resulting from the combined effect of both proteins (as
in serum) can be calculated. The derivation of the appro-
priate equation for this calculation can be made through
competitive binding assumptions.
5. Crystallographic data for 9 were collected at the University
of California at San Diego and have been deposited in the
Cambridge Crystallographic Data Centre as supplementary
publication number 662530.
6. Other physicochemical characterizations of compound 9:
1HNMR (300 MHz) (DMSO-d6) d: 10.91 (br s, 1H), 8.95
(s, 1H), 8.43 (s, 1H), 7.78 (s, 1H), 7.39–7.33 (m, 2H),
7.21–7.13 (m, 2H), 4.71 (s, 2H), 4.53 (s, 2H), 3.25 (s,
3H), 3.18 (s, 3H); 19F NMR d: À115.87; MS: 416.1
(M+H).