Design, synthesis and SAR study of bridged tricyclic pyrimidinone carboxamides as HIV-1 integrase inhibitors

Article abstract of DOI:10.1016/j.bmc.2020.115541

The design, synthesis and structure-activity relationships associated with a series of bridged tricyclic pyrimidinone carboxamides as potent inhibitors of HIV-1 integrase strand transfer are described. Structural modifications to these molecules were made in order to examine the effect on potency towards wild-type and clinically-relevant resistant viruses. The [3.2.2]-bridged tricyclic system was identified as an advantageous chemotype, with representatives exhibiting excellent antiviral activity against both wild-type viruses and the G140S/Q148H resistant virus that arises in response to therapy with raltegravir and elvitegravir.

Full text of DOI:10.1016/j.bmc.2020.115541

Bioorganic&MedicinalChemistryxxx(xxxx)xxxx
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Design, synthesis and SAR study of bridged tricyclic pyrimidinone
carboxamides as HIV-1 integrase inhibitors
⁎
Manoj Patela,e, , B. Narasimhulu Naidua,e, Ira Dickerb,e, Helen Higleyb, Zeyu Linb, Brian Terryb,
Tricia Protackb, Mark Krystalb,e, Susan Jenkinsa,e, Dawn Parkera,e, Chiradeep Panjac,
Richard Rampullad, Arvind Mathurd, Nicholas A. Meanwella, Michael A. Walkera,f
a Departments of Discovery Chemistry and Molecular Technologies, 5 Research Parkway, Wallingford, CT 06492, USA
b Virology, Bristol-Myers Squibb Research and Development, 5 Research Parkway, Wallingford, CT 06492, USA
c Biocon-Bristol-Myers Squibb Research Center, Plot 2 & 3, Bommasandra Industrial Estate - Phase-IV, Bommasandra-Jigani Link Road, Bengaluru, Karnataka 560099,
India
d Department of Discovery Chemistry Synthesis, Bristol-Myers Squibb Research and Development, PO Box 4000, Princeton, NJ 08543-4000, USA
e ViiV Healthcare, 36 East Industrial Parkway, Branford, CT 06405, USA
f Assembly Biosciences, Inc. 331 Oyster Point Blvd, San Francisco, CA 94080, USA
A R T I C L E I N F O
A B S T R A C T
Keywords:
The design, synthesis and structure-activity relationships associated with a series of bridged tricyclic pyr-
imidinone carboxamides as potent inhibitors of HIV-1 integrase strand transfer are described. Structural mod-
ifications to these molecules were made in order to examine the effect on potency towards wild-type and
clinically-relevant resistant viruses. The [3.2.2]-bridged tricyclic system was identified as an advantageous
chemotype, with representatives exhibiting excellent antiviral activity against both wild-type viruses and the
G140S/Q148H resistant virus that arises in response to therapy with raltegravir and elvitegravir.
HIV
HIV Integrase
Integrase inhibitor
Strand transfer inhibitor
Bridged tricyclic pyrimidinone
1. Introduction
Q148R and, especially, G140S/Q148H (GQ).5
Herein we describe the investigation of a series of bridged tricyclic
Human immunodeficiency virus type-1 (HIV-1) integrase is one of
the three enzymes involved in the HIV-1 replication cycle. The in-
tegrase enzyme catalyzes the insertion of viral DNA into the genome of
the host, which is an essential step in the viral replication cycle and
therefore represents an important therapeutic target for HIV-1 drug
development.1,2 Integration is a complex process consisting of two
biochemical steps: 3′-processing and strand transfer. In the 3′-proces-
sing step, integrase binds to the blunt end termini of the double-
stranded viral DNA and removes two nucleotides from each of the 3′
ends, leaving a recessed CA-OH terminus. In the strand transfer step,
the integrase enzyme inserts the processed 3′-hydroxyl into the host
DNA at the site of integration. Raltegravir (Fig. 1) was the first integrase
inhibitor approved for treating HIV-1 infection, while elvitegravir, do-
lutegravir and bictegravir are three additional integrase strand transfer
inhibitors that have been approved after completing clinical trials.3
Dolutegravir and bictegravir are next-generation integrase strand
transfer inhibitor (INSTI) since they retain activity toward viruses that
have developed resistance to raltegravir and elvitegravir.4 Resistant
mutations that arise in response to first generation INSTIs are N155H,
pyrimidinone-carboxamide derivatives as inhibitors of the HIV-1 in-
tegrase strand transfer reaction that retain potent activity toward
viruses resistant to raltegravir and elvitegravir. After an extensive
survey, the [3.2.2]-template represented by 3 (Fig. 2) was selected for
further optimization and the design, synthesis and antiviral activity of
this series of compounds as potential second generation HIV-1 integrase
inhibitors will be discussed.
2. Results and discussion
Screening of a proprietary collection of HIV-1 integrase inhibitors
against wild-type and Q148R viruses resulted in the identification of the
[3.2.1] bridged tricyclic pyrimidinone 1 as a promising lead. This
compound showed excellent antiviral activity toward wild-type virus
replication in cell culture, EC50 = 2 nM, and promising potency toward
the first generation INSTI-resistant viruses Q148R (EC50 = 26 nM) and
GQ, (EC50 = 39 nM). The immediate focus of further structural mod-
ification was to improve the potency toward these resistant viruses. To
that end, the achiral [2.2.2]-bridged tricyclic analog 2 was prepared as
⁎ Corresponding author at: Departments of Discovery Chemistry and Molecular Technologies, 5 Research Parkway, Wallingford, CT 06492, USA.
E-mail address: manoj.m.patel@viivhealthcare.com (M. Patel).
https://doi.org/10.1016/j.bmc.2020.115541
Received 6 February 2020; Received in revised form 25 April 2020; Accepted 29 April 2020
0968-0896/©2020ElsevierLtd.Allrightsreserved.
Pleasecitethisarticleas:ManojPatel,etal.,Bioorganic&MedicinalChemistry,https://doi.org/10.1016/j.bmc.2020.115541

M. Patel, et al.
Bioorganic&MedicinalChemistryxxx(xxxx)xxxx
Fig. 1. The structures of raltegravir, elvitegravir, dolutegravir and bictegravir.
Fig. 2. Structures of bridged tricyclic pyrimidinone-carboxamide derivatives 1–3.
an analogue of 1 with the same molecular constitution but with an
altered bridging element that is attached to the carbon atom bound to
N1. Unfortunately, despite displaying excellent activity toward wild-
type virus, EC50 = 8 nM, 2 was inactive against the GQ mutant virus,
with an EC50 value of > 8 μM. However, by restoring the methylene
spacer between N1 and the bridgehead methine, the alternative achiral
[3,2,2]-bridged tricyclic analog 3 emerged as a compound with im-
proved potency against both wild-type (EC50 = 1 nM) and GQ
(EC50 = 11 nM) viruses, representing a significant advance over the
initial lead compound 1.
Table 1
Antiviral activity of compounds 3–9 toward WT and GQ HIV-
1.
Compd
R
WT EC50
SD (nM)
GQ EC50
SD (nM)
3
4
5
6
7
8
9
Me2NCO.CO
Me2NCO
1
0.6 (n = 7)
10
7 (n = 79)
The initial investigation of 3 focused on determining the importance
of the oxamide moiety for GQ inhibitory activity, with the results
summarized in Table 1. Although compounds 4–9 displayed good po-
tency toward wild-type virus, they differed significantly in their activity
against the GQ mutant virus. The urea 4, which is devoid of one of the
carbonyl moieties of the oxamide of 3, has substantially reduced anti-
viral activity toward the GQ mutant virus, while individual reduction of
each of the oxamide carbonyls gave 5 and 6 which were also poorly
active toward the GQ virus but retained potent inhibition of wild-type
14
0.2 (n = 2)
361 (n = 1)
Me2NCH2CO
Me2NCOCH2
EtO2CCO
20 (n = 1)
> 8000 (n = 1)
> 8000 (n = 1)
> 8000 (n = 1)
7
3 (n = 2)
21 (n = 1)
Me2CHCO.CO
Me2NCOCH2CO
2
6 (n = 2)
94
42 (n = 2)
0.7 (n = 1)
363
415 (n = 3)
EC50 values were determined in the presence of 10% FBS. SD = standard de-
viation.
2

M. Patel, et al.
Bioorganic&MedicinalChemistryxxx(xxxx)xxxx
Table 3
Antiviral activity of substituted benzyl amides 19–28.
Compd
R
WT-EC50
SD (nM)
GQ-EC50
SD (nM)
3
H
1
0.6 (n = 7)
10
17
6
7 (n = 79)
19
20
21
22
23
24
25
26
27
28
3-F
0.7
0.8
0.7
0.5 (n = 2)
14 (n = 4)
3-Cl
0.6 (n = 3)
3 (n = 3)
3-Me
3-CF3
2-F
0.5 (n = 39)
3
2 (n = 52)
1.4 (n = 1)
32 (n = 1)
Fig. 3. Compound 3 (green) docked into the prototype foamy virus (PFV) in-
tegrase (PDB: 3OYL) complex with the proposed binding mode of oligonu-
cleotides (white), H-bond between the carbonyl of compound 3 that is distal to
the core heterocycle and the 3′OH of the terminal ribose of the viral DNA is
indicated by the dashed line.
0.7
0.1 (n = 2)
37 (n = 1)
2-Cl
1.3 (n = 1)
1.3 (n = 1)
8 (n = 1)
60 (n = 1)
11 (n = 1)
1325 (n = 1)
558 (n = 1)
2-Me
2-CF3
2-CONHMe
2-SO2Me
> 8000 (n = 1)
> 8000 (n = 1)
> 8000 (n = 1)
EC50 values were determined in the presence of 10% FBS. SD = standard de-
viation.
HIV-1. Similarly, replacing the N,N-dimethylamide with an ester (7)
resulted in poor GQ inhibitory activity. Interestingly, the 3,3-di-
methylpyruvic acid derivative 8, in which the N,N-dimethylamide ni-
trogen atom has been replaced by CH, retained moderate GQ inhibitory
activity while compound 9, in which the oxamide carbonyls are sepa-
rated by a CH2 spacer, shows significantly reduced potency toward the
resistant virus. It is apparent from this SAR survey that the N,N-di-
methyloxamide moiety is of importance for inhibiting GQ virus, leading
to the hypothesis that this group is engaging in a H-bonding interaction
with the enzyme. Modelling studies using a published cocrystal struc-
ture of the prototype foamy virus (PFV) intasome with a bound in-
hibitor suggested that the oxamide carbonyl of 3 that is distal to the
heterocyclic core is positioned to form a H-bond with the 3′-hydroxyl of
the terminal adenosine nucleoside of the 3′-processed strand (Fig. 3).6
The position of the oxamide with respect to the 3′-hydroxyl is con-
trolled by the 3-dimensional structure of the template and the con-
formational preferences of the oxalamide element.
dimethyloxamide moiety (10) resulted in a 65-fold reduction in GQ
inhibitory potency. On the other hand, replacing one or both oxamide
methyl substituents with ethyl, trifluoroethyl, isopropyl, cyclopropyl or
isobutyl groups were all well tolerated. These compounds, 11–16, ei-
ther retained or exhibited enhanced activity toward the GQ virus, with
the isobutyl derivative 16 the most potent compound to emerge from
this series that offered an 8-fold improvement in potency toward the
resistant mutant. Interestingly, the 2-methoxyethyl analog 17 retained
good potency toward both viruses but the basic 2-dimethylaminoethyl
analog 18 exhibits significantly diminished activity against both the
wild-type virus and GQ variant.
The next series of studies probed the effect of substituents on the
benzyl amide moiety while maintaining the N,N-dimethyloxamide at
the bridgehead position, with the results of this survey summarized in
Table 3. Small substituents at the meta-position, such as F, Cl and Me,
compounds 19–21, respectively, were well tolerated, providing a 2- to
8-fold increase in GQ inhibitory potency compared to 3. Interestingly,
the m-CF3-substituted analogue 22 is almost 25-fold less potent toward
the GQ virus than the m-CH3 homolog 21. In contrast to the effect of
substituents at the meta-position, the ortho-position proved to be highly
sensitive to the size and identity of the substituent and modifications at
this site exerted a significant effect on GQ inhibitory activity, although
the effect on wild-type virus was more modest. The antiviral data for
compounds 3 and 23–28 show that GQ inhibitory potency decreased as
the size of the ortho-substituent increased. The enhanced GQ inhibitory
potency associated with 21 compared to 3 appears to correlate with the
dissociation half life from the enzyme with is 2.3 h for the former and
8.7 h for the latter. For compounds 26–28, which incorporate relatively
bulky CF3, CONHMe and SO2Me substituents, respectively, GQ in-
hibitory activity is substantially reduced and these compounds also
display diminished activity toward the wild-type virus (Table 3). The
reduction in antiviral activity observed with these compounds is
thought to be due, in part, to unfavorable interactions between the
oxamide moiety and the integrase protein arising from a conforma-
tional adjustment that is a consequence of steric interference between
the ortho-substituent and N,N-dimethyloxamide, which are believed to
compete for the same site of the pharmacophore, as is evident from
Fig. 3 and depicted topologically in Fig. 4.
With a preference for an oxamide at the bridgehead position es-
tablished, the effect of substitution on the oxamide terminal nitrogen
atom was examined next. An extensive SAR study was performed, with
representative examples compiled in Table 2. In general, modifications
to the terminal nitrogen atom did not significantly affect activity
against wild-type virus but several of these molecular edits did exert a
significant impact on inhibitory activity toward the GQ mutant virus.
Removal of one of the methyl substituents from the N,N-
Table 2
Antiviral activity of oxamides 10–18.
Compd
R1, R2
WT EC50
SD (nM)
GQ EC50 (nM)
3
Me, Me
1
0.6 (n = 7)
10
7 (n = 79)
10
11
12
13
14
15
16
17
18
Me, H
1
0.5 (n = 3)
0.2 (n = 2)
714 (n = 1)
Me, Et
0.6
12
13 (n = 3)
Me, CF3CH2
Et, Et
0.3 (n = 1)
5 (n = 1)
0.9
1.4
1.5
1.2
1.2
0.6 (n = 2)
8
3 (n = 2)
Me, i-Pr
0.5 (n = 3)
0.3 (n = 2)
0.1 (n = 2)
0.1 (n = 2)
14
14 (n = 3)
Me, c-Pr
8 (n = 1)
Based on the in vitro antiviral data, compounds 3 and 21 were se-
lected for evaluation in a pharmacokinetic (PK) screen in rats following
IV administration, with the results summarized in Table 4. In the in vivo
screen in which plasma exposure was monitored for 4 h, 21 demon-
strated a lower clearance (CL) and a longer half life when compared to
Me, i-Bu
1.4 (n = 1)
Me, MeOCH2CH2
Me, Me2NCH2CH2
8
5 (n = 2)
7 (n = 1)
129 (n = 1)
EC50 values were determined in the presence of 10% FBS. SD = standard de-
viation.
3

M. Patel, et al.
Bioorganic&MedicinalChemistryxxx(xxxx)xxxx
resulted in low confidence in predicting the human PK profile of 21 and
further development of the compound was halted.
3. Synthesis
The synthesis of the [3.2.2] bridged tricyclic pyrimidinone template
was performed as shown in Scheme 1, starting from alcohol 29, which
was prepared according to a previously-described procedure.9 Alcohol
29 was converted into chloride 30 by treatment with SOCl2 in pyridine.
Deprotection of the ketal was followed by a Strecker reaction, with the
resulting amine protected with a CBz moiety to give the aminonitrile
derivative 31. Conversion of the nitrile of 31 to the corresponding
amidoxime was accomplished by treatment with 50% aqueous hydro-
xylamine and the product subjected to a Michael addition reaction with
diethyl acetylenedicarboxylate, providing a mixture of the cis and trans
adducts 32. Pyrolysis of 32 in refluxing o-xylene afforded the desired
pyrimidinone ester 3310,11 which was cyclized with K2CO3 in DMF to
complete construction of the bridged [3.2.2] scaffold, affording 34 in
25% yield over the four steps. The ester 34 was transformed into the
benzyl amide 35 and then to the amine 36 in a two step process that
involved reaction of 34 with benzyl amine in refluxing EtOH followed
by the removal of the CBz group by hydrogenolysis. Amine 36 was
coupled with a carboxylic acid derivative to furnish 3, 8, 9 and 10–18
or acylated to give 4, 5 and 7. Heating amine 36 with 2-chloro-N,N-
dimethylacetamide at 60 °C in THF for 2 h provided 6.
Fig. 4. Topological representation of the potential steric clash between the aryl
ortho- substituent and the oxamide moiety.
Table 4
Pharmacokinetic profiles of 3 and 21 in male Sprague-Dawley rats.
Compound
3
21
IV dose (mg/kg)a
0.5
1
Cmax (nM)
t1/2 (h)
SD
1,925 (2,390, 1,460)
0.45 (0.62, 0.28)
30.5 (23.7, 37.3)b
0.63 (0.57, 0.70)
5,303
12.3
12.6
1.2
1,119
5.5
4.1
0.6
SD
CL (mL/min/kg)
Vss (L/kg)
SD
a
b
Vehicle: PEG-400/ethanol (90:10) solution. n = 2. Data in parentheses
Reagents and conditions: (a) SOCl2, pyridine, 85%; (b) (i) 1 N HCl,
THF, 24 h; (ii) NH4OH, NaCN, THF, 72 h; (iii) CBz-Cl, Na2CO3, DCM, 0
°C to rt, 16 h, 61%; (c) (i) 50% aq. NH2OH, THF reflux, 5 h; (ii)
EtOOCC^CCOOEt, THF, rt, 16 h; (d) o-Xylene, reflux 16 h; (e) K2CO3,
DMF, 90 °C, 3 h, 25%; (f) BnNH2, TEA, EtOH, 90 °C, 16 h; (g) Pd/C, 1 N
HCl, EtOH, 16 h; (h) R1R2NCOCO2H, HATU, DMAP, DMF, 16 h; (i)
RCOCl, TEA, DCM, rt, 16 h; (j) Me2NCOCH2Cl/K2CO3/THF/60 °C/2h.
The synthesis of 19–28 is outlined in Scheme 2. The carbamate 34
was converted to the oxamide 37 which was isolated as a mixture of
esters by first removing the CBz protecting group (H2/MeOH) followed
by coupling with N,N-dimethyloxamic acid using HATU in the presence
of DMAP. The esters of 37 were reacted with a series of benzyl amines
in refluxing EtOH in the presence of Et3N to afford the desired amides
19–28 in good yields.
are values from the individual rats.
Table 5
Pharmacokinetic profiling data for 21 in the rat, monkey and dog.
Species
Rata
Dogb
Monkeyc
IV dose (mg/kg)d
CL (mL/min/kg)
Vss (L/kg)
1
1
1
12.6
1.2
4.1
0.6
5.5
7.0
0.3
1.6
5
2.2
13.5
0.26
5.5
0.01
0.9
0.11
t1/2 (h)
12.3
5
1.6
5
0.75
PO dose (mg/kg)d
AUC (nM*h)
% F
14,657
103
4,939
20,780
83
11,749
2,886
23
912
a
c
d
Sprague-Dawley rats.b beagle dogs. cynomolgus monkey. Vehicle: PEG-
400/ethanol (90:10) solution.
Reagents and conditions: (a) Pd/C, MeOH, 16 h; (b) Me2NCOCO2H,
HATU, DMAP, DMF, 16 h; (c) ArCH2NH2, TEA, EtOH, 90 °C, 16 h
3.
The superior IV PK of 21 in the rat prompted more detailed PK
profiling studies in the rat, dog and monkey, with the results sum-
marized in Table 5. Following IV administration, 21 demonstrated low
to moderate clearance in all three species. After PO dosing, high oral
bioavailability was observed in rat and dogs; however, the oral bioa-
vailability of 21 in non-human primates was significantly lower, which
may be due to glucuronidation of the heterocyclic hydroxyl substituent
in primates. This glucuronidation pathway was identified as the major
metabolic pathway for raltegravir in humans.7,8 In the presence of
45 mg/mL of human serum albumin (HSA), the EC50 value for 21 to-
ward the GQ virus was 21 nM, a 7 fold shift. With a Ctrough target of 3-
fold the EC90 value in the presence of 45 mg/mL HSA, human dosing
predictions based on allometric scaling of the preclinical PK data pro-
jected a high dose in humans of more than a gram of 21 per day. The
combination of the high dose projection superimposed upon the chal-
lenges associated with predicting the disposition of a compound that
may be subject to glucuronidation as the major metabolic pathway
4. Conclusion
In conclusion, we have identified a series of bridged-tricyclic pyr-
imidinone carboxamides as inhibitors of HIV-1 integrase strand transfer
that exhibited potent antiviral activity toward wild-type viruses.
Furthermore, several of these compounds also displayed excellent po-
tency toward the clinically-relevant double mutant G140S/Q148H (GQ)
resistant virus with 3 and 21 demonstrating potent association with
both wild-type and GQ-mutant HIV-1 integrase enzymes in a binding
assay.12 While 21 presented the targeted antiviral properties, evalua-
tion of the PK profile of this compound in preclinical species resulted in
unfavorable human dosing projections that, when superimposed with
the uncertainties associated with predicting the PK properties of a
molecule subject to glucuronidation as the major metabolic pathway,
resulted in low confidence in predicting a successful clinical outcome
and further development of this compound was not pursued.
4

M. Patel, et al.
Bioorganic&MedicinalChemistryxxx(xxxx)xxxx
Scheme 1. Synthesis of compounds 3–13.
Scheme 2. Synthesis of compounds 14–21.
Declaration of Competing Interest
Acknowledgement
We thank Kathy Mosure for assistance with pharmacokinetic data
gathering and analysis.
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influ-
ence the work reported in this paper.
5

M. Patel, et al.
Bioorganic&MedicinalChemistryxxx(xxxx)xxxx
Appendix A. Supplementary material
Fatkenheuer G, Dittmer U, Pfister H, Esser S. J Antimicrob Chemother. 2009;64:25–32;
(f) Marinello J, Marchand C, Mott BT, Bain A, Thomas CJ, Pommier Y. Biochemistry.
2008;47:9345–9354.
Supplementary data to this article can be found online at https://
doi.org/10.1016/j.bmc.2020.115541.
5. (a) Seki T, Suyama-Kagitani A, Kawauchi-Miki S, et al. Antimicrob Agents Chemother.
2015;59:2596–2606;
(b) Pollicita M, Surdo M, Di Santo F, et al. J Antimicrob Chemother.
2014;69:2412–2419.
References
6. Hare S, Vos AM, Clayton RF, Thuring JW, Cummings MD, Cherepanov P. PNAS.
2010;107:20057–20062.
7. Kassahun K, McIntosh I, Cui D, et al. Drug Metab Disp. 2007;35:1657–1663.
8. Berry LM, Liu J, Colletti A, Krolikowski P, Zhao Z, Trffera Y. Drug Metab Disp.
2014;42:707–717.
1. Pace P, Rowley M. Curr Opin Drug Discov Develop. 2008;11:471–479.
2. Xuei G, Aida T. Curr Chem Biol. 2010;4:188–199.
3. Prada N, Markowitz M. Expert Opin Invest Drugs. 2010;19:1087–1098.
4. (a) Tsiang M, Jones G, Goldsmith J, Mulato A, Hansen D, Kan E. Antimicrob Agents
Chemother. 2016;60:7086–7097;
9. Pearson AJ, Fang X. J Org Chem. 1997;62:5284–5292.
10. (a) Naidu BN, Sorenson ME. Org Lett. 2005;7:1391–1393;
(b) Naidu BN. Synlett. 2008;4:547.
(b) Wu G, Abraham T, Saad N. Expert Rev Anti-infective Ther. 2014;12:535–544;
(c) Blanco JL, Whitlock GG. Expert Opin Pharmacother. 2014;15:573–582;
(d) Baldanti F, Paolucci S, Gulminetti R, Barbarini G, Maserati R. J Med Virol.
2010;82:116–122;
11. Pye PJ, Zhong YL, Jones GO, Reamer RA, Houk KN, Askin D. Angew Chem Int Ed.
2008;47:4134–4136.
12. Compound 3 (WT Ki = 2nM, GQ Ki = 17 nM) and Compound 21 (WT Ki = 2 nM,
GQ Ki = 8 nM).
(e) Sichtig N, Sierra S, Kaiser R, Daumer M, Reuter S, Schulter E, Altmann A,
6