Evolution of anti-HIV drug candidates. Part 2: Diaryltriazine (DATA) analogues

Article abstract of DOI:10.1016/S0960-894X(01)00411-5

A synthesis program directed toward improving the stability of imidoyl thiourea based non-nucleoside reverse transcriptase inhibitors (NNRTIs) led to the discovery of diaryltriazines (DATAs), a new class of potent NNRTIs. The synthesis and anti-HIV struct

Full text of DOI:10.1016/S0960-894X(01)00411-5

Bioorganic & Medicinal Chemistry Letters 11 (2001) 2229–2234
Evolution of Anti-HIV Drug Candidates
Part 2: Diaryltriazine (DATA) Analogues
Donald W. Ludovici,a Robert W. Kavash,a,* Michael J. Kukla,a,* Chih Y. Ho,a
Hong Ye,a Bart L. De Corte,a Koen Andries,b Marie-Pierre de Bethune,c Hilde Azijn,c
Rudi Pauwels,c Henry E. L. Moereels,d Jan Heeres,d Lucien M. H. Koymans,d
Marc R. de Jonge,d Koen J. A. Van Aken,d Frederik F. D. Daeyaert,d Paul J. Lewi,d
Kalyan Das,e Edward Arnolde and Paul A. J. Janssend
aJanssen Research Foundation, Welsh and McKean Roads, Spring House, PA 19477, USA
bJanssen Research Foundation, Turnhoutseweg 30, B-2340 Beerse, Belgium
cTIBOTEC, Generaal De Wittelaan L 11 B3, B-2800 Mechelen, Belgium
dJanssen Research Foundation, Center for Molecular Design, Antwerpsesteenweg 37, B-2350 Vosselaar, Belgium
eCenter for Advanced Biotechnology and Medicine and Rutgers University Chemistry Department,
679 Hoes Lane, Piscataway, NJ 08854, USA
Received 5 September 2000; accepted 9 May 2001
Abstract—Asynthesis program directed toward improving the stability of imidoyl thiourea based non-nucleoside reverse tran-
scriptase inhibitors (NNRTIs) led to the discovery of diaryltriazines (DATAs), a new class of potent NNRTIs. The synthesis and
anti-HIV structure–activity relationship (SAR) studies of a series of DATA derivatives are described. # 2001 Elsevier Science Ltd.
All rights reserved.
In the preceding paper, the imidoyl thioureas (ITUs)
were introduced as a new class of potent non-nucleoside
reverse transcriptase inhibitors (NNRTIs) of HIV-1.1
During the course of this work, however, it was quickly
established that ITUs such as 1 (IC50=2.5 nM vs the
LAI strain of HIV-1) were prone to both hydrolytic
lability and oxidative instability. The latter was shown
by the facile conversion to thiadiazole 2, a compound
that was completely devoid of HIV inhibitory activity.
reaction of amidine 3 with N-cyano-O-phenyl isourea 4
resulted in clean conversion to a single product that was
revealed by spectral analysis to be triazine 6, and not the
desired product 5. Presumably, compound 5 had formed
but cyclized spontaneously under the reaction condi-
tions (Fig. 1). Considering the inactivity of 2, we were
surprised to find that 6 was a very potent NNRTI of
HIV-1LAI (IC50=6.3 nM). Indeed, it was nearly as
active as 1, the most potent compound in the ITU ser-
ies. More importantly, 6 was also highly active against a
battery of HIV-1 mutant strains (Table 1). The combi-
nation of potent biological activity and greatly
enhanced structural stability relative to the ITUs
encouraged us to embark on a program to identify and
optimize a second generation of NNRTIs based on the
diaryltriazine (DATA) compound 6. Selection of some
of the modifications made, particularly on the Ar1 ring
and the C-4 position of the triazine ring, was guided by
molecular modeling (3D-SAR) based on the crystal
structure of HIV-1 RT bound with the prototypic
DATA compound 6.3 The molecular modeling showed
the Ar1 ring was positioned in the important Y181,
Y188, W229 region of the enzyme, while the C-4
This suggested to us that compounds in the ITU series
were able to achieve favorable conformations for bind-
ing to HIV-1 reverse transcriptase (RT) that are not
available to the more rigid thiadiazole system and
prompted us to pursue a strategy of attempting to sta-
bilize the ITUs while maintaining their conformational
flexibility. In this context, an attempt was made to pre-
pare compound 5, which incorporated a cyanoguani-
dine moiety, a well-known thiourea bioisostere.2 The
*Corresponding authors. Tel.:+1-215-628-5822; fax:+1-215-628-
5047; e-mail: rkavash@prius.jnj.com. Tel.:+1-215-542-9405; e-mail:
m-rkukla@erols.com
0960-894X/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(01)00411-5

2230
D. W. Ludovici et al. / Bioorg. Med. Chem. Lett. 11 (2001) 2229–2234
position was located at the opening of the binding
pocket. This led us to believe modifications to the Ar1
ring should have an important influence on binding
affinity and the mutant profile, while the C-4 position
should be able to accommodate a large range of
structural changes.
Alimitation of this approach was that ring formation
resulted in placement of an NH2 group at the C-4 posi-
tion of the triazine. The NH2 group was derivatized by a
variety of acylating reagents under standard conditions;
however, in order to further probe the SAR of the C-4
position a second synthetic route was required. The 2,6-
dichlorophenyl moiety, which was associated with very
good HIV inhibitory activity, was chosen as Ar1. Use of
this Ar1 group was compatible with the synthetic route
shown in Scheme 2. The readily formed Grignard
reagent of 2,6-dichlorobenzyl bromide 12, when reacted
with one equivalent of cyanuric chloride (15), afforded a
dichloro intermediate,6 which could be treated with 4-
aminobenzonitrile to form 13 in an efficient one-pot
procedure. Chlorotriazine 13 was a useful, stable inter-
mediate that reacted readily with a wide variety of
The chemistry that led to the formation of 6 was readily
adapted to the preparation of triazine analogues 11 in
which both Ar1 and Ar2 could be varied as shown in
Scheme 1. The requisite amidines 8 were readily pre-
pared from the corresponding phenylacetonitriles 7 by
the procedure of Garigipati,4 while the isourea partners
10 were obtained in modest yields by the reaction of
diphenyl cyanocarbonimidate 95 with the appropriate
aniline.
Figure 1.
Table 1. Inhibition of HIV-1 (IC50, mM)9
Compd
R
X
LAI
100I
103N
181C
188L
6
2,6-diCl
H
2-Cl
CH2
CH2
CH2
CH2
CH2
CH2
CH2
CH2
CH2
CH2
CH2
NH
S
O
O
O
O
O
O
O
O
0.0063
0.285
0.130
0.398
8.07
2.65
0.040
8.47
0.200
>100
2.57
0.791
0.286
0.062
0.225
1.57
0.316
nd
7.37
11a
11b
11c
11d
11e
11f
11g
11h
11i
0.400
0.274
0.063
0.016
0.042
0.273
5.08
2,4-diCl
0.014
2.10
2.32
2,3,6-triCl
2,4,6-triCl
2,6-diMe
2,5-diMe
3,5-diMe
2,3,5,6-tetraMe
2,4,6-triMe
2,4,6-triMe
2,4,6-triMe
2,4.6-triMe
2,6-diMe
2,6-diMeO
2,6-diCl
2,4,6-triCl
2,4,6-triBr
2,4,6-triF
2,6-diCl-4-F
2Cl,4Br,6-Me
2,6-diBr-4-Me
2,6-diMe-4-Br
2,6-diMe-4-Cl
2,6-diMe-4-I
2,6-diMe-4-NO2
2,6-diMe-4-NH2
0.0030
0.0014
0.0012
0.0031
0.091
0.0046
0.0008
0.0010
0.0029
0.0010
0.0031
0.031
0.0031
0.0024
0.0054
0.051
0.0073
0.0021
0.0020
0.0015
0.0035
0.030
0.416
0.221
1.46
5.86
77.9
0.265
0.161
0.105
1.03
>100
0.367
0.045
0.124
0.036
0.182
0.327
nd
6.37
1.19
6.70
0.085
0.0030
0.0020
0.0060
0.0020
0.017
0.526
0.082
0.012
0.017
0.087
0.065
0.0060
0.0050
0.0080
0.011
0.010
0.0080
>100
11j
0.082
0.037
0.064
0.075
0.851
nd
0.384
0.087
0.131
0.216
0.706
0.059
0.019
0.133
0.171
0.204
0.330
nd
0.015
0.011
0.019
0.013
0.333
1.56
0.279
0.062
0.061
3.65
0.259
0.035
0.017
0.042
0.055
0.068
0.057
>100
21a
21b
21c
21d
21e
21f
21g
21h
21i
21j
21k
21l
21m
21n
2o
21p
21q
4.57
0.502
0.619
8.95
1.02
O
O
O
O
O
O
O
0.103
0.415
0.056
0.121
0.075
0.123
nd
0.0020
7.83
nd: not determined.

D. W. Ludovici et al. / Bioorg. Med. Chem. Lett. 11 (2001) 2229–2234
2231
nucleophiles to form numerous analogues of 6 with dif-
ferent substituents at the triazine C-4 position. More-
over, the DATA lead, compound 6, could be prepared
by this route in high yield by the treatment of 13 with a
solution of ammonia in dioxane–isopropanol at ele-
vated temperature in a sealed tube. Compound 14f was
prepared in two steps by the reaction of intermediate 13
with O-(trimethylsilyl)hydroxylamine and subsequent
hydrolysis of the TMS group under aqueous base con-
ditions. This prevented the formation of hydroxylamine
linked bis-adducts.
exhibited a general cytotoxic effect.9 Promising com-
pounds were subsequently tested against a number of
known mutant strains, of which some are clinically sig-
nificant due to resistance to NNRTIs that have
previously been approved for the treatment of AIDS.
Apreference for the NH linkage between Ar
and the
2
triazine ring was rapidly established since an O-linked
analogue of 6 was found to be completely inactive. The
nature of the Ar2 portion of the DATA series was then
investigated by preparing a small series of Ar2 analo-
gues of 6. For example, the 3-cyano and the 4-chloro
analogues had activities (IC50, mM) against HIV-1LAI of
0.251 and 0.016, respectively. These results led us to
conclude that for Ar2, the SAR of the ITU series1 was
applicable and that a 4-cyanoanilino substituent was
optimal. Thus, the 4-cyanoanilino moiety was maintained
for the remainder of our DATA analogue investigations.
Heteroatom linkages connecting Ar1 to the central tria-
zine ring could be incorporated as shown in Scheme 3.
Starting with commercially available cyanuric chloride
(15), the preferable route was to pass through inter-
mediates 17; however, weaker nucleophiles of type
Ar1XH, especially anilines (X=NH) with strong elec-
tron withdrawing substituents, occasionally necessitated
following the route through intermediates 18. The
chlorotriazines 19 could then be reacted with either
ammonia or a variety of nucleophiles, as was described
previously in the case of 13, to provide DATA analo-
gues 21. This approach was amenable to another struc-
tural variation of interest in that dichlorotriazine 16,
prepared by the method of Harris,7 could be used for
entry into analogues 20 in which Z is hydrogen.8
A SAR profile was initially established for Ar1 by
exploration of DATA compounds 11, which possess
both a methylene linkage to the triazine and an NH2
group at the C-4 position. Later in the project, DATA
compounds 21 (Y=NH2), in which Ar1 is linked to the
triazine by a heteroatom, were in hand and the results
reinforced the earlier observations for the SAR of Ar1.
Furthermore, the linkage between Ar1 and the triazine
did not have a significant effect on the potency of
DATA compounds as can be seen from 11j, 21a, 21b,
and 21c, all of which contain 2,4,6-trimethylphenyl as
the Ar1 group. Numerous compounds were prepared;
however, the entries in Table 1 illustrate the trends that
were observed. In general, halogens and methyl sub-
stituents on Ar1 provided similar activity in inhibiting
HIV-1LAI, with the methyl containing derivatives being
slightly more effective. At least one ortho substituent
appears to be required for a significant level of potency
while 2,6-substitution is generally the most active of the
disubstituted Ar1 patterns (6>11cꢀ11b>11a and
11f>11gꢀ11h). As long as a halogen or methyl 2,6-
disubstitution pattern is present, additional halogens or
methyls in the C-3 and/or C-5 positions tend to be close
in potency towards HIV-1LAI. The range in potency for
Ar1 with these multiple substituent patterns is broader
among the mutant strains but without clear trends (i.e.,
The primary screen for evaluating the DATA com-
pounds was potency in inhibiting replication of wild-
type HIV-1 (LAI strain). Also, a ‘safety index’ (or
therapeutic index, expressed as a ratio of toxic con-
centration to inhibitory concentration) was determined
and used as a guide to avoid selecting compounds that
Scheme 1. (a) NH4Cl, AlMe3/PhMe, 0 ꢁC then 80 ꢁC; (b) 1,4-dioxane/
DMF, 70–80 ꢁC; (c) DMF, 70–80 ꢁC.
Scheme 2. (a) Mg, ether; (b) cyanuric chloride, benzene; (c) 4-aminobenzonitrile, DIEA, 1,4-dioxane.
Scheme 3. (a) 4-Aminobenzonitrile, DIEA, 1,4-dioxane, 25 ꢁC or reflux; (b) Ar1NH2, DIEAor Ar 1OH, NaH; 1,4-dioxane, 25 ꢁC or reflux.

2232
D. W. Ludovici et al. / Bioorg. Med. Chem. Lett. 11 (2001) 2229–2234
6 vs 11d and 11f vs 11i). Avariety of small substituents
were tolerated at the C-4 position when combined with
halogen or methyl in a 2,6-disubstituted Ar1. With the
exception of the inactive 4-amino 21q, DATA deriva-
tives with 2,4,6-trisubstitution were generally the most
potent against HIV-1LAI by a slight margin but were
significantly more active against all of the HIV-1 mutant
strains.
14n and hydroxyl 14p derivatives were much less active.
The sulfoxide 14q was inactive, as was the fluoride 14x.
Thus, for activity against HIV-1LAI, a variety of small
functional groups were tolerated at this position on the
triazine ring. However, as indicated previously, the
activity against mutant strains was an important filter in
evaluating the compounds. From this perspective, the
hydroxylamine 14f was of great interest, as it possessed
a superior inhibition profile relative to 6 against a
number of the mutant strains, especially 181C and
188L, which was a breakthrough. Unfortunately, in
preclinical studies, 14f was found to be very rapidly
metabolized affording 6 and products resulting from
hydroxylation of the methylene linkage and the C-4
position of Ar1. Furthermore, 14f and its metabolite 6
were readily glucuronidated at the hydroxylamine/
amine functionality and excreted.
Investigation of substitutions at the triazine C-4 posi-
tion of compounds 14 indicated that the original NH2
group was indeed favorable for inhibitory potency
whereas the chloride intermediate 13 was inactive (Table
2). The monomethylamine analogue 14a was also quite
potent while the dimethylamine analogue 14b was con-
siderably less active. Monoalkyl amines were active as
long as the unfunctionalized alkyl group was not too
bulky. Interestingly, a basic dimethylamine group (14e
and 14r) on the end of the chain did allow for good
activity against HIV-1LAI and some of the mutants. The
size of the alkyl chain was also shown to be a factor in
the hydroxylamine series 14f–14j with the parent
hydroxylamine being the most potent. Anumber of
other small amine derivatives had good activity includ-
ing acetamide 14t, N,N-dimethylformamidine 14u,
ethylcarbamate 14v, and methylamidine 14w. The azido
group 14k provided good potency while the hydrazine
14l had surprisingly low relative activity. Some non-
amine functional groups were also prepared. The S-
methyl 14m and O-methyl 14o were active while, unlike
the amine and hydroxylamine, the corresponding thiol
We reached a point in the program where potency was
not the main SAR criterion for moving forward. We
were confident that we had obtained compounds with
adequate potency and needed to improve other features
of the compounds such as solubility, bioavailability, and
pharmacokinetics. The SAR of the DATA series to this
point suggested that there might be some tolerance for
making modifications with these goals in mind while
maintaining highly active compounds. Consequently,
DATA analogues with heteroatom Ar1 triazine linkages
and substituents at the C-4 position of Ar1 were first
synthesized. Indeed, compound 21a (Table 1) possessed
an excellent potency profile and was determined to have
Table 2. Inhibition of HIV-1 (IC50, mM)9
Compd
Y
LAI
100I
103N
181C
188L
6
13
NH2
Cl
NHMe
0.0063
1.08
0.398
nd
1.16
32.2
nd
0.040
10.0
0.137
4.69
nd
0.200
10.0
0.440
13.4
nd
0.316
nd
1.02
nd
nd
nd
0.297
0.026
0.804
nd
nd
nd
nd
7.32
nd
nd
1.55
nd
nd
0.315
nd
0.158
0.251
1.73
0.358
nd
14a
14b
14c
14d
14e
14f
14g
14h
14i
0.0070
0.166
0.849
0.194
0.0050
0.0024
0.0062
0.0246
0.100
0.086
0.014
0.610
0.018
0.094
0.012
0.160
1.55
0.0055
0.379
0.016
0.0025
0.012
0.0027
5.26
NMe2
NHnPropyl
N-Morpholino
NHCH2CH2N(Me)2
NHOH
nd
nd
nd
1.26
0.107
1.38
nd
nd
nd
nd
5.83
nd
0.021
0.013
0.082
nd
nd
nd
0.224
1.74
0.336
0.723
0.231
0.607
15.5
0.037
7.55
0.040
0.016
0.548
0.040
52.6
0.379
0.038
0.628
nd
nd
nd
1.14
5.83
2.35
1.61
1.45
2.12
68.5
1.39
43.8
100
NHOMe
NHOEt
NHOPropyl
NMeOMe
N3
NHNH2
SMe
SH
OMe
OH
SOMe
14j
14k
14l
14m
14n
14o
14p
14q
14r
14s
14t
14u
14v
14w
14x
nd
1.15
nd
nd
OCH2CH2N(Me)2
NHCONHiPr
NHCOMe
N¼CN(Me)2
NHCO2Et
NH(C¼NH)Me
F
1.08
35.1
100
0.631
1.64
0.611
nd
0.316
1.32
0.210
100
nd: not determined.

D. W. Ludovici et al. / Bioorg. Med. Chem. Lett. 11 (2001) 2229–2234
Table 3. Inhibition of HIV-1 (IC50, mM)9
2233
Cp
R
X
LAI
100I
103N
181C
188L
100I +103N
103N +181C
20a
20b
20c
20d
20e
20f
20g
20h
2,4,6-triMe
2,6-diBr-4-Me
2,6-diMe-4-Br
2,6-diMe-4-CN
2,4,6-triMe
2,6-di-Br-4-Me
2,6-diMe-4-CN
2,4,6-triMe
NH
NH
NH
NH
O
O
O
S
0.0003
0.0005
0.0006
0.0010
0.0006
0.0013
0.0040
0.0020
0.013
0.0030
0.079
0.251
0.020
0.016
1.000
0.032
0.003
0.0030
0.0060
0.0080
0.0030
0.0060
0.020
0.0080
0.0030
0.016
0.050
0.020
0.040
0.501
0.013
0.040
0.079
0.063
0.050
0.063
0.398
0.251
0.020
1.26
>10.0
3.16
0.050
nd
0.126
0.126
nd
nd
0.398
nd
2.51
>10.0
nd
3.98
0.0030
>10.0
increased metabolic stability versus 14f, although a site
for glucuronidation was still present in 21a.
newly characterized double mutant of HIV that was
potentially important. These considerations led to the
development of the diaminopyrimidine (DAPY) analo-
gues, which are the subject of the next paper in this series.
Consideration of these nonpotency issues led us to syn-
thesize DATA compounds 20 with hydrogen at the C-4
position of the triazine. We were very pleased to see that
compounds 20a–h were very potent. In fact, 20a was
the most potent of all the DATA compounds against
HIV-1LAI and had a very good profile against all of the
HIV-1 mutants that we had been testing up to that time
(Table 3). Compound 20a also had good relative meta-
bolic stability in addition to not having a site for glu-
curonidation. However, new clinically relevant
mutations were continuously being identified and we
became aware of additional important HIV-1 strains
with double mutations in the RT. Despite very good
potency against the 100I and 103N single mutants,
compounds 20 were effectively inactive against the double
mutant 100I/103N.
Acknowledgements
Anumber of people have been instrumental in this
work. Thanks go to Diane Gautier and Greg Leo for
help with NMR spectral analyses; to Theo Thielemans
for continuously updating extremely well organized
data presentations; to Bill Leister, Kim Strauss, and
Brian Rafferty for numerous analyses and sample pur-
ifications; to Inge Dierynck and Kurt Hertogs (TIBO-
TEC) for the construction of HIV-1 mutated strains;
and to Art Clark and others at CABM and Rutgers
University and Stephen Hughes and his colleagues at
NIH-NCI for their help with the structural studies.
Other promising DATA compounds such as 21a were
tested against the 100I/103N double mutant and were
found to also be completely inactive. In the following
paper in this series10 we report how replacement of the
triazine ring with a pyrimidine ring system allowed for
the identification of a structural feature in which high
potency in inhibiting even this double mutant could be
achieved.
References and Notes
1. Ludovici, D. W.; Kukla, M. J.; Grous, P. G.; Krishnan, S.;
Andries, K.; de Bethune, M.-P.; Azijn, H.; Pauwels, R.; De
Clercq, E.; Arnold, E.; Janssen, P. A. J. Bioorg. Med. Chem.
Lett. 2001, 11, 2225 (first paper in series).
2. Durant, G. J.; Emmett, J. C.; Ganellin, C. R.; Miles, P. D.;
Parsons, M. E.; Prain, H. D.; White, G. R. J. Med. Chem.
1977, 20, 901.
Starting from the active but unstable ITU series of
HIV-1 NNRTIs, we discovered the new class of DATA
compounds which maintained high potency and safety
levels while solving the stability problem. Asystematic
SAR program, supplemented with 3D-SAR via molecular
modeling and structural studies, identified features of
the DATA series that led to subnanomolar potency
versus the wild-type HIV-1 (LAI strain) and low nano-
molar potency against a battery of clinically important
HIV-1 mutants. As the project matured, nonpotency
issues including solubility, bioavailability, and pharma-
cokinetics became the focus of the SAR development
program. These efforts culminated in the identification
of DATA compounds 20a and 21a, which met the goals
of the project very well. However, HIV is a ‘moving
target’ owing to its high mutation rate, and compounds
in the DATA series had suboptimal activity against a
3. Detailed results will be described in future publications.
4. Garigipati, R. S. Tetrahedron Lett. 1990, 31, 1969
5. Webb, R. L.; Labaw, C. S. J. Heterocycl. Chem. 1982, 19,
1205.
6. Hirt, R.; Nidecker, H.; Berchtold, R. Helv. Chim. Acta
1950, 33, 1365.
7. Harris, R. L. N. Synthesis 1981, 907.
8. Attempts to prepare a methylene-linked analogue of 6 via a
Grignard reaction with 16 were unsuccessful.
9. (a) All compounds were tested for potency (IC50, mM) in
terms of concentrations required to achieve 50% protection of
MT-4 cells from HIV-1 cytopathicity as determined by the
MTT method (Pauwels, R.; Balzarini, J.; Baba, M.; Snoek, R.;
Schols, D.; Herdewijn, P.; Desmyter, J.; De Clercq, E. J. Virol.
Methods 1988, 20, 309). Unless noted otherwise, the LAI
strain of HIV-1 was the infecting virus. Other infecting mutant
strains of HIV-1 (with changes in the RT) are characterized in
the tables by the mutated amino acid position and the one

2234
D. W. Ludovici et al. / Bioorg. Med. Chem. Lett. 11 (2001) 2229–2234
letter codes. For instance, 181C refers to replacement of tyr-
osine at position 181 with cysteine. All determinations are the
result of multiple tests. (b) nd, not determined.
10. Ludovici, D. W.; De Corte, B. L.; Kukla, M. J.; Ye, H.;
Ho, C. Y.; Lichtenstein, M. A.; Kavash, R. W.; Andries, K.;
de Bethune, M.-P.; Azijn, H.; Pauwels, R.; Lewi, P. J.; Heeres,
J.; Koymans, L. M. H.; de Jonge, M. R.; Van Aken, K. J. A.;
Daeyaert, F. F. D.; Das, K.; Arnold, E.; Janssen, P. A. J.
Bioorg. Med. Chem. Lett. 2001, 11, 2235.