Structure-based amelioration of PXR transactivation in a novel series of macrocyclic allosteric inhibitors of HIV-1 integrase

Article abstract of DOI:10.1016/j.bmcl.2020.127531

Previous studies have identified a series of imidazo[1,2-a]pyridine (IZP) derivatives as potent allosteric inhibitors of HIV-1 integrase (ALLINIs) and virus infection in cell culture. However, IZPs were also found to be relatively potent activators of the pregnane-X receptor (PXR), raising the specter of induction of CYP-mediated drug disposition pathways. In an attempt to modify PXR activity without affecting anti-HIV-1 activity, rational structure-based design and modeling approaches were used. An X-ray cocrystal structure of (S,S)-1 in the PXR ligand binding domain (LBD) allowed an examination of the potential of rational structural modifications designed to abrogate PXR. The introduction of bulky basic amines at the C-8 position provided macrocyclic IZP derivatives that displayed potent HIV-1 inhibitory activity in cell culture with no detectable PXR transactivation at the highest concentration tested.

Full text of DOI:10.1016/j.bmcl.2020.127531

Bioorganic&MedicinalChemistryLetters30(2020)127531
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Structure-based amelioration of PXR transactivation in a novel series of
macrocyclic allosteric inhibitors of HIV-1 integrase
⁎
Prasanna Sivaprakasama, , Zhongyu Wangb,h, Nicholas A. Meanwellb,i, Javed A. Khanc,
David R. Langleya,j, Stephen R. Johnsonc, Guo Lib,i, Annapurna Pendrib,k, Timothy P. Connollyb,l,
Mian Gaod, Daniel M. Camacc, Cheryl Klakouskie,m, Tatyana Zvyagae,n, Christopher Ciancif,
Brian McAuliffef,o, Bo Dingf,o, Linda Discottof, Mark R. Krystalf,o, Susan Jenkinsg,o
,
Kevin M. Peeseb,o, B. Narasimhulu Naidub,o
a Computer-Aided Drug Design, Bristol-Myers Squibb Research and Development, 5 Research Parkway, Wallingford, CT 06492, USA
b Discovery Chemistry, Bristol-Myers Squibb Research and Development, 5 Research Parkway, Wallingford, CT 06492, USA
c Molecular Structure & Design, Bristol-Myers Squibb Research and Development, P.O. Box 4000, Princeton, NJ 08543-4000, USA
d Protein Sciences, Bristol-Myers Squibb Research and Development, P.O. Box 4000, Princeton, NJ 08543-4000, USA
e Lead Discovery and Optimization, Bristol-Myers Squibb Research and Development, 5 Research Parkway, Wallingford, CT 06492, USA
f Virology Discovery Biology, Bristol-Myers Squibb Research and Development, 5 Research Parkway, Wallingford, CT 06492, USA
g Pharmaceutical Candidate Optimization, Bristol-Myers Squibb Research and Development, 5 Research Parkway, Wallingford, CT 06492, USA
A R T I C L E I N F O
A B S T R A C T
Keywords:
PXR
Previous studies have identified a series of imidazo[1,2-a]pyridine (IZP) derivatives as potent allosteric inhibitors of
HIV-1 integrase (ALLINIs) and virus infection in cell culture. However, IZPs were also found to be relatively potent
activators of the pregnane-X receptor (PXR), raising the specter of induction of CYP-mediated drug disposition
pathways. In an attempt to modify PXR activity without affecting anti-HIV-1 activity, rational structure-based design
and modeling approaches were used. An X-ray cocrystal structure of (S,S)-1 in the PXR ligand binding domain (LBD)
allowed an examination of the potential of rational structural modifications designed to abrogate PXR. The in-
troduction of bulky basic amines at the C-8 position provided macrocyclic IZP derivatives that displayed potent HIV-1
inhibitory activity in cell culture with no detectable PXR transactivation at the highest concentration tested.
ALLINI
HIV-1 integrase
LEDGF/p75
Macrocycle
Inhibitor
Acquired immunodeficiency syndrome (AIDS) is a major health
challenge. Human immunodeficiency virus-1 (HIV-1) is the causative
agent of AIDS, and while there is no cure for this infection, highly active
antiretroviral therapy (HAART) can effectively halt progression of the
disease by reducing viral replication to levels below the limit of de-
tection. According to the World Health Organization, by the end of
2018 there were approximately 37.9 million people worldwide infected
with HIV-1, of which 23.3 million were receiving HAART.1 HIV-1 in-
tegrase is one of the three key enzymes of HIV-1 and the one that
catalyzes the integration of viral DNA into the host cell genome. Small
molecule inhibitors that block the action of HIV-1 integrase at the ac-
tive site (termed integrase strand transfer inhibitors, INSTIs) have been
quite successful and some are currently used as backbone agents in
antiretroviral therapy. There are currently four approved drugs that act
by this mechanism: raltegravir, elvitegravir, dolutegravir and bicte-
gravir, while a fifth agent, cabotegravir, is a long-acting INSTI with a
once monthly dosing regimen currently under review at the Food and
Drug Administration (FDA).2–5 While our initial efforts in this field were
⁎ Corresponding author at: Molecular Structure & Design, Bristol Myers Squibb Research and Early Development, P.O. Box 4000, Princeton, NJ 08543-4000, USA.
E-mail address: prasanna.siva@bms.co (P. Sivaprakasam).
h Present address: Sterne Kessler Goldstein & Fox, PLLC, 1100 New York Ave NW # 600, Washington, DC 20005, USA.
i Present address: Discovery Chemistry, Bristol Myers Squibb Research and Early Development, P.O. Box 4000, Princeton, NJ 08543-4000, USA..
j Present address: Arvinas Inc, 5 Science Park, 395 Winchester Ave, New Haven, CT 06501, USA.
k Present address: Discovery Chemistry, Bristol Myers Squibb Research and Early Development, 100 Binney Street, Cambridge, MA 02142, USA.
l Present address: Global Pharmaceutical Research & Development, AbbVie Inc., 1401 N. Sheridan Road, North Chicago, IL 60064, USA.
m Present address: Clinical Pharmacology Services Covance Inc., USA.
n Present address: Lead Discovery and Optimization, Bristol Myers Squibb Research and Early Development, P.O. Box 4000, Princeton, NJ 08543-4000, USA.
o Present address: ViiV Healthcare, 36 East Industrial Road, Branford, CT 06405, USA.
https://doi.org/10.1016/j.bmcl.2020.127531
Received 10 July 2020; Received in revised form 27 August 2020; Accepted 28 August 2020
Availableonline02September2020
0960-894X/©2020ElsevierLtd.Allrightsreserved.

P. Sivaprakasam, et al.
Bioorganic&MedicinalChemistryLetters30(2020)127531
to interrogate active site inhibitors of HIV-1 integrase,6–8 data on a new
target within the integrase multimer prompted an examination of al-
losteric inhibitors of the enzyme. The first allosteric inhibitors of HIV-1
integrase were described by Christ et al. who adopted the name lens
epithelium-derived growth factor (LEDGF/p75) inhibitors (LEDGINs).9
LEDGINs bind to HIV-1 integrase in a pocket remote from the active site
and where the cellular cofactor LEDGF/p75 binds. This class of in-
hibitor is now more commonly referred to as non-catalytic site integrase
inhibitors (NCINIs)10,11 or, more recently, as allosteric inhibitors of
HIV-1 integrase (ALLINIs).11 ALLINIs have multiple mechanisms of
action during the viral life cycle. They induce a hyper multimerization
of the protein, which occurs late in the infection cycle, altering the
morphogenesis of new particles, resulting in the creation of non-in-
fectious virions. ALLINIs also inhibit early in the virus life cycle, by
inhibiting proviral DNA integration through inhibition of LEDGF/p75
association.12–16
Compounds from the IZP series 3 and 4a-d that form the basis of
this study are compiled in Table 1 along with data for (S,S)-1 and 2,
which are included for the purpose of comparison. The in vitro EC50
values for 4a-d in cell culture ranged from 4.4 to 9.5 nM, a 2- to 4-fold
improvement in potency compared to 3. The macrocyclic prototype 4a,
although demonstrating excellent antiviral activity in cell culture suf-
fered from a significant PXR liability (PXR EC50 = 1.01 µM,
Ymax = 71%, where Ymax is a calculated maximal achievable response
reported as the % of control that can be produced by the test compound
at the highest non-cytotoxic concentration relative to the known posi-
tive control rifampicin tested at a concentration of 10 μM). This ob-
servation prompted studies designed to identify a path toward miti-
gating PXR activation in order to allow further development of the
series. PXR has both a DNA-binding and a ligand-binding domain
(LBD). The expression, purification, and crystallization of PXR has
previously been described24,25 and X-ray crystallographic studies have
helped in understanding the molecular details of the binding of xeno-
biotics to the PXR LBD.26 Although we were not able to obtain co-
crystal structures of members of the IZP series of compounds with the
PXR LBD, an available co-crystal structure of the structurally-related
(S,S)-118 in the PXR LBD provided key insights. The prototype IZP
macrocycle 4a and (S,S)-1 share significantly similar three dimensional
space, enabling the use of (S,S)-1 as a surrogate with which to study 4a
and to develop models of its binding mode in the PXR LBD. We carefully
reviewed the co-crystal structure of (S,S)-1 in the PXR LBD with the
goal of identifying structural features of 4a that could be modified to in
an effort to attenuate PXR activation without negatively impacting
antiviral activity. This endeavor was further enabled by the availability
of a co-crystal structure of (S,S)-1 bound to the HIV-1 integrase catalytic
core domain (CCD) dimer which includes the ALLINI binding site. This
facilitated the identification of suitable vectors in 4a for the introduc-
tion of substituents that were predicted to be tolerated in the integrase
allosteric binding site while being detrimental to binding to the PXR
LBD to elicit an activating response. The best computationally docked
poses of 4a in the HIV-1 integrase CCD and the PXR LBD that aided in
this analysis are shown in the Supplementary material. The coordinates
of the PXR LBD: (S,S)-1 complex have been deposited in the Protein
Data Bank under the PDB code 6XP9.
As part of our discovery efforts17–21 to identify new ALLINIs, we have
described the discovery and profiling of the pyrazolo[1,5-a]pyrimidine
derivative (S,S)-1 that had its origin in the piperidine-substituted pro-
genitor 2.18 As part of the exploration of this chemotype, alternate core
heterocycles were examined that included the imidazo[1,2-a]pyridine (IZP)
derivative 3.19 This chemotype provides an opportunity to explore the ef-
fects of substitution of the heterocyclic core at the C-8 position, a vector
unavailable to (S,S)-1 and 2. The antiviral activity of compounds in cell
culture was assessed using a fully replicating NL4-3 virus in the MT-2 cells,
as previously described.17,18 The antiviral activity of 3 is comparable to 2
(Table 1) and the biphenyl moiety in this compound provided a convenient
scaffold on which to probe the effect of installation of a macrocyclic linker
between this element and the 4-position of the piperidine ring, which is
bound to the C-5 position of the core, a chemotype represented generically
by 4. The carbinol 4a (Table 1), a prototype molecule in this structural
class, offered potent antiviral activity in the cell-based HIV-1 replication
assay. However, pregnane-X receptor (PXR) activation was identified as a
major liability with this molecule. PXR belongs to the superfamily of nu-
clear hormone receptors and can be activated by a wide range of en-
dogenous and exogenous chemical substances.22 PXR activation may lead
to the induction of CYP 450 enzymes, of particular concern for a drug
anticipated to be used as part of an anti-HIV-1 combination therapy (or
with other concomitant drugs) where the induction of metabolism may
compromise efficacy by reducing the systemic exposure of other drugs. Our
attempts to mitigate PXR activation in this macrocyclic IZP series of AL-
LINIs are described herein. PXR activity was determined using a transac-
tivation assay conducted in HepG2 cells.23
It should be noted that the binary complex of the PXR LBD in
complex with (S,S)-1 crystallizes as a dimer, with similar interactions of
2

P. Sivaprakasam, et al.
Bioorganic&MedicinalChemistryLetters30(2020)127531
Table 1
Antiviral activity in cell culture assay, PXR transactivation and calculated pKa data for the ALLINIs (S,S)-1, 2, 3 and 4a-d.
Cmpd
R
WT HIV-1 EC50
SD (nM)a
PXR EC50 (µM)b/Ymax
Calculated strongest basic pKa with reliability range (in
log units) 27
(S,S)-1
2
–
109.3
130
20.9
5.7b
4.4
61.4
5.7/99%
–
–
–
–
–
–
60
3
–
9
–
4a
4b
4c
C(CH3)2OH
1.01/71%
> 50
> 50
> 50
CH2N(CH3)2
CH2NH.CH(CH3)2
CH2NH.CH2C(CH3)2OH
0.5
3
10.3
10.8
9.7
0.8
0.8
7.3
4d
9.5b
0.5
a
b
EC50 values were the result of multiple determinations (n ≥ 2).
EC50 values were obtained after single determination (n = 1).
(S,S)-1 in each of the binding pockets; however, for the purpose of
comparison only one chain (chain A) is discussed here.
(Fig. 1A). In contrast, the carboxylic acid group of (S,S)-1 in the PXR
LBD is submerged into a space filled with water molecules. However,
His407 lies close to the carboxylic acid moiety of (S,S)-1, allowing for
the possibility of a salt bridge interaction. The carboxylic acid group of
(S,S)-1 in HIV-1 integrase projects above the plane of the central bi-
cyclic core, whereas in the PXR LBD it points below the plane of the
(S,S)-1 occupies the LEDGF/p75 binding site of the integrase dimer.
The carboxylic acid group of (S,S)-1 makes a bidentate hydrogen-
bonding interaction with the N-Hs of the backbone amides of Glu170
and His171 and the side chain of Thr174 at the dimer interface
Fig. 1. X-Ray co-crystal structure of (S,S)-1 in: A – the HIV-1 integrase CCD (PDB code 6UM8); and B – the PXR LBD (PDB code 6XP9). The figures were prepared
using the program Pymol (The Pymol Molecular Graphics system, v.2.3.4 Schrödinger, LLC).
3

P. Sivaprakasam, et al.
Bioorganic&MedicinalChemistryLetters30(2020)127531
of the bicyclic core of (S,S)-1 loosely interacts with Gln95 through two
water molecules in the HIV-1 integrase CCD, suggesting the potential
for the introduction of a substituent at this position (i.e. at the C-8
position of the IZP macrocyclic series) that would not negatively impact
the anti-HIV-1 activity. Unlike the HIV-1 integrase CCD structure, the 2-
phenyl moiety of (S,S)-1 is deeply positioned in a hydrophobic pocket
in the PXR LBD. The 2-phenyl ring of (S,S)-1 in the PXR LBD interacts
tightly with Leu239, Pro227, Tyr225, Leu213, Tyr306, and Val211. The
bicyclic core of (S,S)-1 sits above Met243 from the α3-helix and faces
the distal end of the α2-helix in the PXR LBD where there is no ob-
servable electron density for residues Cys207 to Leu209. Grafting this 3-
residue structural element from other PXR LBD structures into the
missing region of the (S,S)-1-bound structure revealed a tight junction
around the core N-4 position in (S,S)-1. This region is fairly solvent
exposed in the HIV-1 integrase CCD co-crystal structure and enabled the
hypothesis that suitably spaced bulky substituents projecting from this
region of 4 would sterically collapse with the distal end of the α2-helix
in the PXR LBD. Upon visualizing the IZP macrocycle 4 with a relatively
small group at this region which is close to the core heterocycle, this
hypothesis was tested by introducing extended, bulky, and polar sub-
stituents at the R position, as captured in Table 1. Taking advantage of
the aldehyde intermediate S19 used in the synthesis of 4a, analogues
4b-d, which incorporate basic amines, were quickly generated (Scheme
1; see Supplementary material for syntheses details). All three of these
compounds demonstrated EC50 values of > 50 µM in the PXR transac-
tivation assay while retaining single digit nM EC50 inhibitory values
toward HIV-1 replication in cell culture. The calculated pKa values
support the possibility that 4b-d exist in the protonated form under the
assay conditions.27 The basic amine moieties installed in 4b-d would
incur destabilizing interactions in the hydrophobic region of the PXR
LBD binding pocket. While binding and desolvation of a positively
charged side chain projecting from the 8-position is expected to be
unfavorable in the interior of the PXR LBD, it is predicted to bind just
fine at the solvent front in the HIV-1 integrase CCD. In fact the C-8
position in the IZP macrocycle chemotype offers an ideal site to install
such solubilizing groups without compromising potency, a vector not
available to the ALLINIs (S,S)-1 and 2. It is believed that the combi-
nation of high polarity and steric bulk at the C-8 position that sterically
destabilizes the α2-helix of the PXR LBD of the IZP macrocyclic core
contributes to attenuating the PXR transactivation liability.
Fig. 2. Overlay of X-ray co-crystal binding modes of (S,S)-1 in the HIV-1 in-
tegrase CCD dimer (orange) and the PXR LBD (green). The figures were pre-
pared using the program Pymol (The Pymol Molecular Graphics system, v.2.3.4
Schrödinger, LLC).
bicyclic core (Fig. 2). In the HIV-1 CCD co-crystal structure, the tert-
butoxy group of (S,S)-1 sits in a well-defined hydrophobic pocket lined
with Thr125, Gln95, Tyr99 and Thr174. Similarly, the tert-butoxy group
of (S,S)-1 occupies a hydrophobic sub-pocket created by Trp299,
Met323, His327, and Gln285 in the PXR LBD structure. These ob-
servations highlight a notable difference between the binding mode of
the carboxylic acid element of (S,S)-1 in the HIV-1 integrase and the
PXR LBD (Fig. 2). The PXR LBD takes advantage of the conformational
flexibility of the carboxylic acid moiety in (S,S)-1 and in the bound
form, this element is flipped upside down whereby the carboxylic acid
and the tert-butoxy groups are favorably positioned in the hydrophilic
and hydrophobic sub-pockets, respectively, relative to its conformation
in the HIV-1 integrase CCD. One could conceive of conformationally
locking the carboxylic acid and the tert-butoxy groups in their observed
states from the HIV-1 integrase CCD structure which would place the
two groups in the opposing sub-pockets in the PXR LBD. However,
synthetically arriving at such conformationally locked versions in this
system would be quite challenging.
In summary, a novel series of macrocyclic IZP-based ALLINIs with
potent HIV-1 inhibitory activity in cell culture is described. Relative to
the previously disclosed ALLINI 2, this new series demonstrates 14 to
23-fold better antiviral potency. However, 4a suffered from a sig-
nificant PXR transactivation liability that would compromise its use as a
therapeutic agent. Using the co-crystal structure complex of a structu-
rally similar ALLINI bound to the PXR LBD, it was hypothesized that the
PXR liability in this series could be abrogated by the judicious in-
troduction of substituents at C-8 of the core heterocycle. The validity of
the design hypothesis was confirmed with the synthesis of several de-
rivatives that presented greatly improved PXR activation profiles in a
The C7-chromane ring of (S,S)-1 occupies a deep hydrophobic
pocket formed by Leu102, Ala129, Thr174, Met178, Trp132 and
Ala169 in the HIV-1 integrase CCD. Similarly, the C7-chromane ring is
caged in a hydrophobic pocket created by Tyr306, Met246, Cys301,
Phe288, and Trp299 in the PXR LBD.
The bicyclic core of (S,S)-1 and the 2-phenyl substituent are well
positioned onto the hydrophobic surface of one of the two subunits of
the dimer in the HIV-1 integrase CCD. The 2-phenyl ring of (S,S)-1
makes van der Waals contact with Ala128 in the α3-helix(124–133) of
one of the subunits in the HIV-1 integrase CCD. The N-4 nitrogen atom
Scheme 1. Synthetic protocol for the synthesis of 4b-d from S19.
4

P. Sivaprakasam, et al.
Bioorganic&MedicinalChemistryLetters30(2020)127531
cell-based PXR transactivation assay, without affecting HIV-1 antiviral
activity in cell culture. This affords this inhibitor series a pathway to
developing safer drug candidates. The deposited ALLINI:PXR LBD co-
crystal structural complex in the PDB adds value to the existing che-
mical space available for the study of PXR as well as the ALLINI re-
search area.
5. Hassounah SA, Alikhani A, Oliveira M, et al. Antimicrob Agents Chemother.
2017;61:e01695.
6. Naidu BN, Walker MA, Sorenson ME, et al. Bioorg Med Chem Lett. 2018;28:2124.
7. Naidu BN, Sorenson ME, Patel M, et al. Med Chem Lett. 2015;25:717.
8. Peese KM, Narasimhulu NB, Patel M, et al. Bioorg Med Chem Lett. 2019:30.
9. Christ F, Voet A, Marchand A, et al. Nat Chem Biol. 2010;6:442.
10. Balakrishnan M, Yant SR, Tsai L, et al. PLoS ONE. 2013;8:e74163.
11. Jurado KA, Engelman A. Expert Rev Mol Med. 2013;15:e14.
12. Koneru PC, Francis AC, Deng N, et al. eLife. 2019;8:e46344.
13. Engelman AN. J Biol Chem. 2019;294:15137.
Declaration of Competing Interest
14. Jurado KA, Wang H, Slaughter A, et al. Proc Natl Acad Sci. 2013;110:8690.
15. Feng L, Sharma A, Slaughter A, et al. J Biol Chem. 2013;288:15813.
16. Desimmie BA, Schrijvers R, Demeulemeester J, et al. Retrovirology. 2013;10:57.
17. Peese KM, Allard CW, Connolly T, et al. J Med Chem. 2019;62:1348.
18. Li G, Meanwell NA, Krystal MR, et al. J Med Chem. 2020;63:2620.
19. Connolly TP, Eastman KJ, Naidu BN, Peese K. US2018/0222901A1. Example 4 is
compound 3.
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.
20. Peese KM, Wang Z, Kadow JF, Sivaprakasam P, Naidu BN. US 20150232480 A1.
August 20, 2015.
Appendix A. Supplementary data
21. A detailed full article on the discovery of the IZP macrocyclic series will be reported
elsewhere in due course.
Supplementary data to this article can be found online at https://
doi.org/10.1016/j.bmcl.2020.127531.
22. Ihunnah CA, Jiang M, Xie W. Biochim Biophys Acta. 2011;812:956.
23. Zhu Z, Puglisi J, Connors D, et al. J Biomol Screen. 2007;12:248.
24. Wang W, Prosise WW, Chen J, et al. Protein Eng Des. 2008;21:425.
25. Khan JA, Camac DM, Low S, et al. J Mol Biol. 2015;27:924.
26. Buchanan CD, Chai SC, Chen T. Exp Opin Drug Metab Toxicol. 2018;14:635.
27. pKa values were calculated using the GALAS algorithm in ACD/Percepta, version
2019.1.3, Advanced Chemistry Development, Inc., Toronto, ON, Canada, www.ac-
dlabs.com, 2019.
References
1. http://www.who.int/.
2. Summa V, Petrocchi A, Bonelli F, et al. J Med Chem. 2008;51:5843.
3. Motohide S, Hiroshi K, Takahisa M, et al. J Med Chem. 2009;52:4869.
4. Johns BA, Kawasuji T, Weatherhead JG, et al. J Med Chem. 2013;56:5901.
5