Discovery of triazolinone non-nucleoside inhibitors of HIV reverse transcriptase

Article abstract of DOI:10.1016/j.bmcl.2008.06.080

Novel non-nucleoside inhibitors of HIV-RT that contain pyridazinone isosteres were prepared, and a series of triazolinones were found to be potent inhibitors of HIV replication. These compounds were active against several NNRTI-resistant virus strains. Ph

Full text of DOI:10.1016/j.bmcl.2008.06.080

Bioorganic & Medicinal Chemistry Letters 18 (2008) 4348–4351
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery of triazolinone non-nucleoside inhibitors of HIV reverse transcriptase
*
Zachary K. Sweeney , Sahaja Acharya, Andrew Briggs, James P. Dunn, Todd R. Elworthy, Jennifer Fretland,
Anthony M. Giannetti, Gabrielle Heilek, Yu Li, Ann C. Kaiser, Michael Martin, Y. David Saito, Mark Smith,
Judy M. Suh, Steven Swallow, Jeffrey Wu, Julie Q. Hang, Amy S. Zhou, Klaus Klumpp
Roche Palo Alto LLC, 3431 Hillview Avenue, Palo Alto, CA 94304, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Novel non-nucleoside inhibitors of HIV-RT that contain pyridazinone isosteres were prepared, and a ser-
ies of triazolinones were found to be potent inhibitors of HIV replication. These compounds were active
against several NNRTI-resistant virus strains. Pharmacokinetic studies indicated that inhibitor 7e has
good bioavailability in rats. Several fragments of inhibitor 7c were prepared, and the binding of these
compounds to HIV-RT was analyzed by surface plasmon resonance spectroscopy.
Ó 2008 Elsevier Ltd. All rights reserved.
Received 2 April 2008
Revised 20 June 2008
Accepted 24 June 2008
Available online 28 June 2008
Keywords:
Human immunodeficiency virus
HIV-1
Reverse transcriptase
Non-nucleoside
The reverse transcriptase enzyme of the human immunodefi-
ciency virus (HIV-RT) transcribes the RNA viral genome into
DNA. This essential step in the viral replication cycle can be inhib-
ited by a class of drugs known as non-nucleoside reverse transcrip-
tase inhibitors (NNRTIs).1 There are four marketed NNRTIs for the
treatment of HIV infection, and several compounds that inhibit
the replication of commonly observed NNRTI-resistant viruses
are in development.2,3
Table 1
Activity of 1 against wild-type and mutant virusesa
F
NC
O
N
NH
Cl
O
Cl
CH3
1
In a previous letter,4 we described the discovery of a series of
pyridazinones that are potent inhibitors of viral replication in vitro.
Pyridazinone 1, was determined to have excellent activity against
the wild-type virus and several commonly observed resistant
viruses (Table 1). However, low aqueous solubility for a number
of these inhibitors appeared to limit oral bioavailability at higher
doses. In order to further optimize the potency and bioavailability
of this series of NNRTIs, we focused on modification of the pyridaz-
inone functionality. In this letter, we describe diaryl ether NNRTIs
that contain modified pyridazinones and heterocyclic pyridazinone
isosteres. This effort resulted in the discovery of a series of triazoli-
none inhibitors that have good oral bioavailability and excellent
activity against common NNRTI-resistant viruses.
WT
G190A
0.001
K103N/L100I
0.004
K103N/Y181C
0.019
0.001
a
EC50 in lM versus the wild-type virus or the indicated mutant.
tures of the inhibitors themselves revealed that in the solid state
the pyridazinones were associated through extensive aromatic
stacking interactions. Therefore, our design strategy was to retain
the N-NH binding motif while introducing groups that might per-
turb intermolecular stacking interactions and improve the solubil-
ity of the inhibitors.
A simple scaffold that did not contain a fluorinated core ring
was initially used to investigate the activity of the new inhibitors.
Compounds were assessed only in terms of their ability to inhibit
the polymerase activity of wild-type reverse transcriptase, as our
experience with the pyridazinone compounds had suggested that
potency against NNRTI-resistant viruses was likely to be highly
dependent upon the substitution pattern of the phenyl rings. Se-
lected NNRTIs prepared in this initial series are shown in Table 2,
along with the corresponding activity in the polymerase enzyme
assay.5
Crystal structures of inhibitors similar to 1 bound to HIV-RT
demonstrated that the pyridazinones interact with the protein
backbone through a pair of hydrogen bonds between the amide
of the K103 amino acid and the N-NH acceptor–donor motif of
the pyridazinone. Inspection of the small molecule crystal struc-
* Corresponding author. Tel.: +1 650 855 6376; fax: +1 650 852 1311.
E-mail address: zachary.sweeney@roche.com (Z.K. Sweeney).
0960-894X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2008.06.080

Z. K. Sweeney et al. / Bioorg. Med. Chem. Lett. 18 (2008) 4348–4351
4349
Table 2
Potency of a series of NNRTIs
the evaluation of NNRTIs that feature a fluorinated central phenyl
ring and a 3-chloro, 5-cyano substituted terminal phenyl ring so
that the compounds could be easily compared to lead compound
1. The dihydropyridazinone 10 and the dihydrotriazinone 11 were
designed to project the methyl group out of the plane of the het-
erocycle, as it was thought that a similar orientation of the methyl
group of oxadiazinone 5a might be responsible for the potency of
this inhibitor. Selected examples of this group of compounds are
shown in Table 3. These NNRTIs again had excellent potency in
the polymerase assay. The nearly identical activity of these diverse
compounds suggests that the strength of the hydrogen bonding
interactions of the heterocycle with the protein backbone do not
strongly influence the thermodynamics of the protein–inhibitor
interaction. It was later demonstrated that these compounds de-
rive much of their potency from hydrophobic interactions of the
biaryl ether (vide infra). However, excision of the methyl group
from the heterocyclic ring of these inhibitors did result in a loss
of activity.
Br
O
Heterocycle
Cl
Compound
Heterocycle
IC50
N
NH
2a
0.13
O
CH3
N
NH
3a
4a
5a
6a
7a
0.23
0.33
0.05
0.15
0.69
O
N
H3C
N
CH3
The potency of the new set of inhibitors against several NNRTI-
resistant viruses was good, and the EC50 of 5b, 7b, and 11 against
the K103N/Y181C mutant virus was less than 50 nM.6 Unfortu-
nately, compound 5b was determined to be a strong inhibitor of
N
NH
O
CH3
CYP3A4 (IC50 = 0.27 lM). Strong CYP inhibition was also observed
with other 1,3,4-oxadiazin-5-ones, and this group of compounds
was not investigated further. Dihydropyridazinone 10 was less po-
tent than the other inhibitors in the cellular assay. Compound 11
and similar triazinones were rapidly metabolized in the presence
of rat microsomal protein (Cl = 130 lg/mL/mg), and the reduced
stability of compounds of this type was confirmed through in vivo
N
NH
O
O
CH3
N
studies.
NH
Triazolinone 7b showed good potency in the cell-based assay
and was only slowly metabolized in the presence of human liver
N
O
microsomes (Cl = 5
lg/mL/min).
A
set of triazolinones with
CH3
N
Table 3
NH
O
Activity of inhibitors 5b, 7b, 10, and 11a
N
H3C
F
NC
O
Heterocycle
N
Cl
NH
Cl
8a
9a
0.53
5.80
S
O
Compound
Heterocycle
WT IC50
0.013
WT EC50
0.005
K103N/Y181C EC50
0.034
N
R
N
NH
O
NH
O
5b
O
O
CH3
N
IC50 in lM versus the wild-type virus.
R
NH
7b
10
11
0.011
0.017
0.011
0.008
0.032
0.008
0.040
N
Almost all of the new compounds retained good potency in the
enzyme assay. Substitution of the pyridazinone methyl group with
the considerably larger dimethylamino group resulted in only a
moderate loss in activity. Introduction of an additional heteroatom
adjacent to the methylene bridge was also accommodated (i.e., tri-
azinone 4a). The non-aromatic oxadiazine 5a and 1,2,4-triazine 6a
were equipotent to pyridazinone 2a in this assay. Finally, good
inhibition of polymerase activity was observed with compounds
containing 5-membered heterocycles. For this series of inhibitors,
triazolinone 7a and thiadiazinone 8a were somewhat more potent
than oxadiazinone 9a.
H3C
R
O
N
NH
na
O
CH3
N
R
NH
HN
0.072
O
CH3
In order to further investigate compounds that had promising
activity in the initial screen, inhibitors that contained more highly
substituted biaryl ethers were prepared. These efforts focused on
a
IC50 or EC50 in
lM versus the wild-type virus and the K103N/Y181C mutant
virus.

4350
Z. K. Sweeney et al. / Bioorg. Med. Chem. Lett. 18 (2008) 4348–4351
Table 4
Table 6
Activity of triazolinone inhibitors 7a
Activity of fragments 8 and inhibitor 7c
F
N
NH
O
NC
O
R
N
R
N
NH
N
H3C
O
Cl
7
Compound
R
WT IC50
Kd
7h
7i
7j
7k
7l
7m
7c
H
Ph
>100
>100
>100
>10
34
18
NB
NB
388
102
4
Compound
R
WT
K103N/Y181C
7b
7c
7e
7f
Cl
Br
CH3
Et
c-Pro
0.011
0.002
0.002
0.002
0.002
0.039
0.009
0.026
0.004
0.011
4-Br-Ph
2-F,4-Br-Ph
2-F,3-OMe,-4-Br-Ph
3-PhO-Ph
3
7g
See Table 4
0.005
0.004
a
EC50 in lM versus the wild-type virus and the K103N/Y181C mutant virus.
ing constant extracted from the SPR experiments. The smallest
fragments did not show any binding within the limits of detec-
tion of the SPR assay (ꢂ2 mM). Very weak binding was observed
with halogenated benzyl triazolinone compounds 7j and 7k in
the SPR assay, although this interaction still could not be de-
tected in the enzymatic assay. Only larger compounds 7l and
7m showed micromolar inhibition of the polymerase activity.
These molecules have more heavy atoms than the small com-
pounds commonly included in fragment screening libraries.
The results from these experiments confirm the utility of SPR
fragment screening for lead discovery. Fragment 7j, which has a
ligand efficiency (LE) of 0.3 kcal/mol/non-hydrogen atom, could
have been identified as a useful starting point for the design of
more potent triazolinone NNRTIs in a fragment screening ap-
proach to lead identification. It should be noted, however, that
this LE value had previously been suggested to be a minimum
value for a useful fragment lead.12 Figure 1 plots the free energy
different substituents on the central phenyl ring was prepared in
an effort to further improve the potency of these compounds
against NNRTI-resistant viruses (Table 4). From the group of
triazolinones prepared, inhibitor 7c was selected for further
characterization.
As indicated by the data in Table 5, 7c was able to effectively re-
duce the replication of viruses containing mutations at positions
103, 181, 100, and 190 of the reverse transcriptase (Table 5). Sur-
face plasmon resonance (SPR) experiments7 showed that 7c bound
very tightly to an immobilized wild-type reverse transcriptase
enzyme, although binding was readily reversible. Under these
conditions, the dissociation constant (Kd) was calculated to be
4 ꢀ 10ꢁ9 Mꢁ1. Finally, 7e did not strongly inhibit the activity of
major cytochrome p450 enzymes (e.g., CYP3A4 IC50 = 12 l
M).8
The solubility of the triazolinone inhibitors was significantly
improved relative to the analogous pyridazinone compounds. For
example, although the solubility of compound 7c (solubility in
of binding (DG) calculated from the SPR binding data versus the
pH 6.5 buffer = 2.4 lg/mL) is low, this inhibitor is more than 10-
number of heavy atoms for compounds 7j–m and 7c. As has
been observed in earlier work, there is a roughly linear correla-
tion between binding energy and mass.13 The slope of the plot
indicates that each heavy atom contributes, on average,
0.59 kcal/mol to the binding energy of the molecule. This num-
ber is substantially greater than the slopes observed in most ear-
lier deconstruction analyses for compounds that have between
15 and 25 heavy atoms, and might be explained by the hydro-
phobic nature of the NNRTI binding site. It is possible that a
molecule such as 7c, which has an excellent LE (0.43), may have
been discovered starting from a fragment hit with a substantially
worse binding efficiency. Such a finding may encourage optimi-
zation of even modestly efficient leads when the compound
binding site is already known to be highly druggable.
fold more soluble than the pyridazinone lead 1.9 The permeability
of 7c as measured in CACO-2 experiments was high
(Papp = 15.1 cm2/s), and no indication of transporter-mediated ef-
flux was observed. Administration of an aqueous suspension of this
inhibitor to dogs (0.5 mg/kg iv, 2 mg/kg po) demonstrated that the
7c was metabolically stable and well absorbed (Cl = 1.5 mL/min/kg,
Cmax = 1.3 lM, AUC0–1 = 38.4 lM h, F% = 74).
Inhibitor deconstruction and fragment screening: Fragment
screening has emerged as a useful method for the discovery of
new templates for medicinal chemistry lead optimization pro-
grams.10 However, there have been few reports of the successful
application of fragment screening for the discovery of inhibitors
that bind to induced pockets such as the HIV-RT NNRTI binding
site. In order to gain insight into the size of fragments that would
have been required to identify triazolinones as templates for fur-
ther optimization, we synthesized various fragments of 7c and
studied their binding to immobilized HIV-RT using surface Plas-
mon resonance (SPR).11 These smaller compounds were also as-
sayed as inhibitors of the polymerase activity of HIV-RT. The Kd
determined in the SPR study and the IC50 calculated from the poly-
merase assay are listed in Table 6.
12
10
8
In general, there was good agreement between the inhibitory
IC50 derived from the functional polymerase assay and the bind-
6
4
Table 5
2
Activity of 7c against a set of NNRTI-resistant virusesa
0
Virus
WT
Y181C
0.005
G190A
0.001
L100I/K103N
0.002
12
14
16
18
20
22
24
26
28
EC50
0.002
# Heavy Atoms
a
EC50 in lM against the indicated virus.
Figure 1.
DG binding versus number of heavy atoms for 7j–m and 7c.

Z. K. Sweeney et al. / Bioorg. Med. Chem. Lett. 18 (2008) 4348–4351
4351
Acknowledgment
F
OH
NC
O
N
a
CO2Et
We thank Grace Lam, Witold Woroniecki, and Jessica Brilliant
for their analytical expertise.
O2
NC
Cl
Cl
References and notes
b, c, d
1. Spence, R. A.; Kati, W. M.; Anderson, K. S.; Johnson, K. A. Science 1995,
267, 988.
2. Balzarini, J. Curr. Top. Med. Chem. 2004, 4, 921.
3. Sorbera, L. A.; Serradell, N.; Bolos, E.; Rosa, J. Drugs Future 2007, 32, 1046.
4. Sweeney, Z. K.; Dunn, J. P.; Li, Y.; Heilek, G.; Dunten, P.; Elworthy, T. R.; Han, X.;
Harris, S. F.; Hirschfeld, D. R.; Hogg, J. H.; Huber, W.; Kaiser, A.; Kertesz, D. J.;
Kim, W.; Roepel, T.; Mirzadegan, M. G.; Saito, Y. D.; Silva, T. M. P. C.; Swallow,
S.; Tracy, J. L.; Villasenor, A.; Vora, H.; Zhou, A. S.; Klumpp, K. Bioorg. Med. Chem.
Lett. 2008, 18, 4352.
F
e, f
NHNH2
NC
O
Br
7c
O
Cl
Scheme 1. Reagents: (a) KOtBu, ethyl (2, 3-difluoro-4-nitrophenyl)acetate, THF,
5. Enzymatic assay: IC50 values were obtained from a scintillation counting assay
using poly rA primer and tritiated dTTP.
t
90%; (b) Fe, NH4Cl, EtOH, H2O, 91%; (c) BuONO, CuBr2, LiBr, CH3CN, 50%; (d) N2H4,
EtOH; (e) OCNCH3, THF; (f) KOtBu, tBuOH, 74% for three steps.
6. Replication assay: EC50 values were obtained following incubation of
compound with infected MT4 cells in the presence of 10% fetal bovine
serum. The assay assesses the reduction in the cytopathic effect of HIV-1 on
MT4 cells in the presence of HIV-RT inhibitors. Each value represents the
average of at least two independent assays.
7. Surface plasmon resonance experiments were performed using HIV-RT protein
immobilized to the CM5 sensor chip via an amide-coupling reaction. The
binding experiments were performed using Hepes buffer (50 mM Hepes (pH
8.0), 150 mM NaCl, 10 mM MgCl2, 0.01% Tween 20, 5% DMSO) as the running
buffer. For characterization of other NNRTIs using SPR see: Geitmann, M.;
Danielson, U. H. Bioorg. Med. Chem. 2007, 15, 7344.
8. CYP3A4 inhibition was assessed using 7-benzyloxy-4-(trifluoromethyl)-
coumarin as a fluorogenic substrate.
9. Interestingly, the differences in solubility between these inhibitors cannot
simply be attributed to differences in intermolecular interactions in the solid
state. The melting temperature of crystals of 1 (181–182 °C) is nearly identical
to that of 7c (185–186 °C).
The synthesis of inhibitor 7c is outlined in Scheme 1.14 An aro-
matic substitution reaction of 3-cyano, 5-chlorophenol with a
difluoronitrophenyl acetate provided the diaryl ether in excellent
yield. Reduction of the nitro group with iron, followed by a Sand-
meyer reaction gave the chlorinated product. Sequential reaction
of the ester with hydrazine and methyl isocyanate provided a
semicarbazide that could be dehydrated with a catalytic amount
of KOtBu in BuOH to form the triazolinone NNRTI.
t
In summary, a number of non-nucleoside inhibitors of HIV re-
verse transcriptase were prepared. Evaluation of the potency and
physical properties of these compounds led to the identification
of a series of triazolinones that strongly inhibited wild-type and
NNRTI-resistant viruses in vitro. These compounds had signifi-
cantly improved physical properties relative to their pyridazinone
analogs, and inhibitor 7c was determined to have excellent bio-
availability in animal studies. A deconstruction exercise was used
to determine the suitability of fragment-based approaches for lead
identification for the NNRTI binding site.
10. Abad-Zapatero, C. Expert Opin. Drug Disc. 2007, 2, 469.
11. For an early example of such
a ‘‘inhibitor deconstruction” exercise see:
Babaoglu, K.; Shoichet, B. K. Nat. Chem. Biol. 2006, 2, 720.
12. Hopkins, A. L.; Groom, C. R.; Alex, A. Drug Discovery Today 2004, 9, 430.
13. Hajduk, P. J. J. Med. Chem. 2006, 49, 6972.
14. For complete details of the synthesis of inhibitor 7c and other NNRTIs
described in this communication see: Saito, Y. D.; Smith, M.; Sweeney, Z. K. US
2007/078128 and Dunn, J. P.; Elworthy, T. R.; Stefanidis, D.; Sweeney, Z. K. WO
2006/010545.