Development of a selective HDAC inhibitor aimed at reactivating the HIV latent reservoir

Article abstract of DOI:10.1016/j.bmcl.2020.127367

The synthesis and SAR development of a trisubstituted imidazole HDAC inhibitor is described. The compounds were synthesized with high diastereocontrol by leveraging Ellman sulfinyl imine chemistry. Structural elucidation provided insight into binding mode and supported design rational. Pharmacokinetic properties of lead compounds were determined.

Full text of DOI:10.1016/j.bmcl.2020.127367

Bioorganic&MedicinalChemistryLetters30(2020)127367
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Development of a selective HDAC inhibitor aimed at reactivating the HIV
latent reservoir
⁎
Dane J. Clausena, , Jian Liua, Wensheng Yua, Joseph L. Duffya, Christine C. Chunga,
Robert W. Myersa, Daniel J. Kleinb, James Fellsb, Kate Hollowayb, Jin Wua, Guoxin Wub,
Bonnie J. Howellb, Richard J.O. Barnardb, Joseph Kozlowskia
a Merck & Co., Inc., 2000 Galloping Hill Road, Kenilworth, NJ 07033, USA
b Merck & Co., Inc., 770 Sumneytown Pike, West Point, PA 19486, USA
A R T I C L E I N F O
A B S T R A C T
Keywords:
The synthesis and SAR development of a trisubstituted imidazole HDAC inhibitor is described. The compounds
were synthesized with high diastereocontrol by leveraging Ellman sulfinyl imine chemistry. Structural eluci-
dation provided insight into binding mode and supported design rational. Pharmacokinetic properties of lead
compounds were determined.
Medicinal chemistry
HDAC
HIV
Latent reservoir
X-ray structure
SAR analysis
The critical role histone deacetylases (HDACs) perform within eu-
karyotic cells for gene transcription has made them important biolo-
gical targets. These enzymes play a key role in chromatin structure and
how DNA is packed around the histone to form the nucleosome.1–3 By
removing acetyl functionality from lysine residues present on the his-
tone surface, HDACs increase the overall local positive charge present.
This phenomenon results in a stronger ionic interaction with DNA’s
negatively charged backbone ultimately leading to a highly condensed
structure that is unreadable to the cell’s transcription machinery.4
Eighteen unique isoforms of HDACs have been discovered and are
separated into four distinct classes.5 Class I (HDACs 1, 2, 3, and 8), Class
competent virus that was integrated into the individual’s genome
during the initial infection.9 A therapy known as “shock and kill” has
been proposed as a possible cure.10 This treatment utilizes an epigenetic
therapeutic to awaken the infected dormant CD4+ T-cell to express the
viral information within to be targeted by a kill mechanism. Margolis
and coworkers verified the “shock” aspect of the therapy by showing an
increase in HIV RNA expression in resting CD4+ T-cells in infected
patients after administrating the pan-HDACi vorinostat (1).11 This pi-
votal observation calls for the creation of HDACi that meet stricter
safety, selectivity, and therapeutic profiles than those currently ap-
proved for the use in oncology.
II (HDACs 4, 5, 6, 7, 9, and 10), and Class IV (HDAC11) are Zn2+
-
Our internal effort is focused on developing more advanced HDACi.
This account expands the scope previously described for the dis-
ubstituted imidazole HDAC inhibitor 4 that contains a ketone zinc
binding group (bidentate coordination to the metal in its hydrated
form).12,13 Incorporation of key findings discovered through these re-
search projects will accelerate our newly proposed series.14 We look to
improve the selectivity for HDAC 1, 2, and 3 inhibition that is needed to
reactivate the infected, dormant CD4+ T-cells while achieving a clean
off-target profile with acceptable ADME properties.13
dependent and Class III are the NAD+-dependent Sirtuins (SirT1-7).
Research efforts have led to the development of multiple advance his-
tone deacetylase inhibitors (HDACi) that display either pan inhibition
or selectivity for certain classes/isoforms (Fig. 1).6,7 A vast majority of
historical HDACi clinical research has focused on combating cancer.
However, recent reports have shown new utility for these inhibitors as
agents to reactivate the latent HIV reservoir.8
HIV has become a controllable chronic disease from the power of
combination antiretroviral therapy (cART) to suppress viral levels to
undetectable amounts. Unfortunately, the patient is not cured through
this treatment alone and viral infection rapidly returns when cART is
stopped. This stems from latent CD4+ T-cells containing replication
The design for this new series was aided from x-ray structure elu-
cidation of inhibitor 4 in the active site of HDAC2 (Fig. 2). This study
showed that the imidazole core interacts with the water pocket present
on the surface of the enzyme through hydrogen bonding. We
⁎ Corresponding author.
E-mail address: dane.clausen@merck.com (D.J. Clausen).
https://doi.org/10.1016/j.bmcl.2020.127367
Received 1 April 2020; Received in revised form 18 June 2020; Accepted 20 June 2020
Availableonline26June2020
0960-894X/©2020ElsevierLtd.Allrightsreserved.

D.J. Clausen, et al.
Bioorganic&MedicinalChemistryLetters30(2020)127367
Fig. 1. Representative HDAC inhibitors 1–4.
Scheme 1. (a) NaH, DMF, 0 °C, then SEMCl; (b) NBS, AIBN, CHCl3, 60 °C; (c)
PhB(OH)2, K3PO4, PPh3 Pd G2, Dioxane/H2O, 100 °C; (d) 3.0 M LiOH, EtOH; (e)
DMP, DCM; (f) 2-methylpropane-2-sulfinamide, PPTS, MgSO4, DCM; (g) LDA,
THF, −78 °C; (h) TFA, DCM, then 4.0 M HCl in Dioxanes; (i) Boc2O, DIPEA,
DCM; (j) TMS-Diazomethane, MeOH/Toluene; (k) Pent-1-en-3-one, Zhan 1B,
DCM/Toluene, 50 °C; (l) Pd/C, H2, MeOH; (m) TFA, DCM; (n) thiazole-5-car-
boxylic acid, PyBOP, DIPEA, DMF; (o) 3.0 M LiOH, MeOH, 60 °C; (p) R2NH,
HATU, DIPEA, DMF.
Fig. 2. X-Ray structure of HDACi 4 in HDAC2. Access codes for X-ray co-
ordinates in RCSB Protein Data Bank (PDB) database are 6WBW.
through a metathesis/reduction protocol.18 TFA revealed the amine
present in compound 11 for a PyBOP mediated coupling. Another
round of deprotection and amide couplings provided imidazole amide
compounds A-D.
Chiral primary amide E was easily constructed for a close compar-
ison against known HDACi 4 (Scheme 2). Amine 11 was subjected to
the same sequence previously mentioned using the correct corre-
sponding reactants. The desired stereochemistry was isolated through
SFC purification.
The number of primary amide substituted imidazole variants was
expanded to include the optimal amide linkages found within closely
related series (Scheme 3). Acid 14 was obtained in a similar manner as
above, however a more electron poor ring system was installed instead
to improve the metabolism. A more divergent synthesis was adopted by
incorporating the zinc binging group prior to joining the two halves.
Therefore, the electrophile for the Ellman reaction was altered to sul-
finyl imine 15 previously described by Schultz-Fademrecht and
Fig. 3. Proposed trisubstituted heterocycle variant.
hypothesized we could extend further into this pocket and form
stronger interactions by substituting the heterocycle with polar func-
tionality. Fig. 3 outlines our proposed trisubstituted imidazole variants.
The regiochemstiry of the imidazole was reversed to obtain a carbon
atom orientated towards the open pocket that was functionalized with
either an amide or cyano group. We tested our hypothesis by synthe-
sizing a diverse set of compounds with varying imidazole substitution,
aryl groups, and amide linkages.
The synthesis for a select variety of amide substituted imidazole
compounds is outlined in Scheme 1. Commercially available imidazole
5 was protected and exposed to radical bromination conditions to fur-
nish bromide 6. Phenyl boronic acid readily reacted with this inter-
mediate through a Suzuki cross-coupling and after saponification acid 7
was obtained.15 Alcohol 8 was converted to imine 9 through DMP-
mediated oxidation and condensation with tert-butylsulfinamide.
The key transformation relied on using Ellman sulfinyl imine
chemistry to unite imidazole 7 with imine 9.16,17 Acidic conditions
removed the liable functionality and the resulting intermediate was
reprotected to grant terminal olefin 10. This provided a synthetic
handle to introduce different zinc binding groups into the inhibitor
Scheme 2. (a) 1-methylazetidine-3-carboxylic acid, PyBOP, DIPEA, DMF; (b)
3.0 M LiOH, MeOH, 60 °C; (c) NH3, HATU, DMF; (d) SFC Purification.
2

D.J. Clausen, et al.
Bioorganic&MedicinalChemistryLetters30(2020)127367
Scheme 3. (a) 4F-PhB(OH)2, K3PO4, PPh3 Pd G2, Dioxane/H2O, 100 °C; (b)
3.0 M LiOH, EtOH; (c) LDA, THF, −78 °C; (d) HCl, Dioxane/H2O; (e) Boc2O,
DIPEA, DMF; (f) NH3, HATU, DMF; (g) TFA, DCM; (h) RCOOH, HATU, DIPEA,
DMF.
Scheme 5. (a) 3.0 M LiOH, EtOH; (b) NH3, HATU, DMF; (c) Tf2O, Pyridine,
DCM, 0 °C; (d) LDA, THF, −78 °C; (e) HCl, MeOH/Dioxanes, then TFA, MeOH;
(f) (S)-6-methyl-6-azaspiro[2.5]octane-1-carboxylic acid, HATU, DIPEA, DMF;
(g) ArBR2, K3PO4, XPhos Pd G3, Dioxane/H2O, 100 °C; (h) TFA, DMC, 50 °C.
coworkers.19 The union of the two fragments proceeded with great
diastereocontrol to give compound 16.20 Global deprotection occurred
with HCl and the nascent amine was reprotected to avoid in-
tramolecular cyclization during the primary amide installation. In-
hibitors F-I were formed using HATU and the required acid once in-
termediate 17 had its amine functionality denuded.
cells) at two different serum levels to gauge potency shift caused by
plasma protein binding.22–24 Clear SAR trends were exposed and the
results of these studies are presented in Table 1. Overall these trisub-
stituted imidazole HDACi showed excellent selectivity for HDAC1, 2,
and 3. The selectivity over HDAC6 was also significantly improved
compared to its disubstituted imidazole parent (trisubstituted HDACi B
selectivity for HDAC1/HDAC6 = ~28000x versus disubstituted HDACi
4 selectivity for HDAC1/HDAC6 = ~52x).12
The nitrile imidazole core was accessed by taking advantage of the
primary amide present in compound 17 (Scheme 4). The dehydration
occurred smoothly using triflic anhydride and pyridine. Imidazole 18
was rapidly diversified using modern parallel synthesis practices for the
generation of cyano substituted variants J-P.
The type of amide present on the central imidazole core had a direct
effect on the potency of the inhibitor. Primary amide B was more active
(cell assay (0.1%, 5% NHS) = 154, 537 nM) compared to secondary
amide A (cell assay (0.1%, 5% NHS) = 570, 1804 nM) or tertiary amide
D (cell assay (0.1%, 5% NHS) = 2729, 3771 nM). This observation can
be explained by the primary amide having more H-bond donors avail-
able to form a greater number of favorable interactions in the water
pocket. At this point, the primary amide functionality was locked onto
the heterocycle and explored in more detail.
The synthetic strategy presented in Scheme 5 showcases the route
used to vary the aryl substituent present on the nitrile imidazole scaf-
fold. Bromide 6 was treated with the reproducible process to convert
the ester to the desired cyano group. Sulfinyl imine 15 underwent nu-
cleophilic attack once heterocycle 13 was deprotonated by LDA. Se-
lective deprotection of compound 19 was achievable and general amide
forming conditions granted completion of Suzuki coupling partner
21.14 Multiple aryl groups were introduced using a highly active pal-
ladium source and simple SEM cleavage afforded HDACi Q-X.
HDACi A-X were evaluated for inhibitory activity and selectivity
against HDAC 1, 2, 3, 6, and 8.22 In addition, the compounds ability to
reactivate dormant CD4+ T-cells into producing silenced genetic in-
formation was measured using a Jurkat model of HIV latency (2C4
The lessons learned during the development of the disubstituted
heterocyclic core provided direction. Incorporation of the known aze-
tidine fragment removed the potency shift present in the cell assay and
improved the overall off target profile at no cost to potency. The loss in
activity caused by installing a more metabolically stable aryl group was
gained back by using the optimal 6-azaspiro[2.5]octane (phenyl azeti-
dine HDACi E (cell assay (0.1%, 5% NHS) = 294, 269 nM) or 4F-phenyl
spiro amine HDACi I (cell assay (0.1%, 5% NHS) = 193, 196 nM).
To help mitigate PK concerns, the primary amide was converted to
the nitrile to remove two H-bond donors. Regrettably, this modification
resulted in decreasing enzymatic inhibition. Comparing the azetidine
containing primary amide F versus cyano J shows a ~7.5-fold loss in
HDAC1 potency with a ~3-fold drop in efficacy in the cell assay. The
more advanced spiro amine containing nitrile HDACi O and P showed
the same degree of enzyme potency degradation, however only dis-
played minimal loss of activity on the dormant CD4+ T-cells compared
to their primary amide versions. The trends for the amide functionality
within the nitrile variant mirrored those previously discussed.
Aryl group manipulation was explored further within the cyano
imidazole series and the SAR did not tract with its disubstituted
Scheme 4. (a) Tf2O, Pyridine, DCM, 0 °C; (b) TFA, DCM; (c) RCOOH, HATU,
DIPEA, DMF.
3

D.J. Clausen, et al.
Bioorganic&MedicinalChemistryLetters30(2020)127367
Table 1
The HDAC and cell assay potency data of compounds A-X.21
Entry Compound
HDAC1 IC50
HDAC2 IC50
(nM)
HDAC3 IC50
(nM)
HDAC6 IC50
(nM)
HDAC8 IC50
(nM)
Cell Assay 0.1%
NHS (nM)
Cell Assay 5%
NHS (nM)
(nM)
A
27.6
91.4
26.7
53.1
4.7
29,180
16,400
> 45000
7471
4701
5335
570
1804
537
B
C
7.2
< 1.5
4.5
154
13.5
1142
853
D
E
43.6
1.9
159
19.1
2.5
18,960
9829
16,780
5422
2729
294
3771
269
18.1
F
3.5
34.4
73.8
3.6
3.4
15,940
27,850
5383
7566
775
643
G
12.3
1834
1504
H
I
0.6
0.4
6.3
5.0
0.9
0.7
17,650
17,460
6802
6364
166
193
276
165
(continued on next page)
4

D.J. Clausen, et al.
Bioorganic&MedicinalChemistryLetters30(2020)127367
Table 1 (continued)
Entry Compound
HDAC1 IC50
(nM)
HDAC2 IC50
(nM)
HDAC3 IC50
(nM)
HDAC6 IC50
(nM)
HDAC8 IC50
(nM)
Cell Assay 0.1%
NHS (nM)
Cell Assay 5%
NHS (nM)
J
25.9
117
13.8
36,930
19,650
1564
2071
K
6.2
26.9
2.4
> 45000
6929
1731
2207
L
47.9
10.6
219
26.8
3.1
43,320
29,490
12,260
1605
254
3532
1111
M
51.3
> 45000
N
35.6
164
12.1
> 45000
16,190
558
2726
O
P
Q
R
S
3.8
19.7
11.7
177
96
2.9
26,610
23,610
8972
16,750
9370
233
161
548
566
107
409
2.8
1.8
259
97.9
17.0
3.7
101
10.7
1.9
> 45000
39,610
19,190
1351
1162
175
22,540
8249
15.6
(continued on next page)
5

D.J. Clausen, et al.
Bioorganic&MedicinalChemistryLetters30(2020)127367
Table 1 (continued)
Entry Compound
HDAC1 IC50
(nM)
HDAC2 IC50
(nM)
HDAC3 IC50
(nM)
HDAC6 IC50
(nM)
HDAC8 IC50
(nM)
Cell Assay 0.1%
NHS (nM)
Cell Assay 5%
NHS (nM)
T
35.2
58.9
0.3
155
229
1.5
18.9
34.8
0.2
14,820
17,060
6301
20,730
38,080
9835
1233
2112
105
77
1260
1669
97
U
V
W
0.2
1.7
0.3
9466
12,620
55
X
0.4
2.1
0.3
9044
12,420
247
322
imidazole counterpart. The methoxyquinoline found within compound
4 was less tolerated (HDACi Q HDAC 1 = 97.9 nM, cell assay (0.1%, 5%
NHS) = 548, 1351 nM). In addition, there was a preference for para
versus ortho substitution that was not present prior to revering the
imidazole’s orientation. This different regiochemistry resulted in un-
favorable electronic interactions between the heterocyclic core and the
ortho substitution present on the aryl fragment ultimately leading to a
potency loss. Large ring systems with heteroatoms further out resulted
in activity enhancement (HDACi W HDAC 1 = 0.2 nM, cell assay
(0.1%, 5% NHS) = 77, 55 nM).
HDACi E is oriented exactly as its parent 4 and makes the same key
interactions between Asp104 with both the N–H present within the
imidazole and the secondary amide bond (Fig. 4). The ketone zinc
binding group was revealed to be in its hydrate state, forming a bidentate
interaction with the Zn2+ present in the active site. The primary amide
adhered to the heterocyclic core participated extensively within the
water pocket present on the enzyme’s surface. The carbonyl is bridged
through water to His183, one of the zinc chelating residues. In addition,
the amine is coordinated to two separate water molecules that are tied
back to the enzyme’s peptide backbone. This observation is consistent
with the potency increase seen across the amides tested (primary >
secondary > tertiary). The greater number of H-bond donors available
relates to the more interactions the inhibitor can make within the pocket.
The x-ray structure of HDACi O in HDAC2 is shown in Fig. 5. Once
again, the key interactions of the imidazole, secondary amide, and zinc
binding group are similar to disubstituent variant 4. The water molecules
and the cavity itself are in the same general space with no changes
compared to the enzyme pocket when primary amide E is bound. The
nitrile present also interacts with His183 through a water bridge. There is
a Van der Waals interaction between the cyano functionality and Leu276.
Fig. 4. X-Ray structure of HDACi E in HDAC2. Access codes for X-ray co-
ordinates in RCSB Protein Data Bank (PDB) database are 6XEB.
The pharmacokinetic properties of primary amide substituted imi-
dazole HDACi E and nitrile substituted imidazole HDACi O were pro-
filed in rat (Table 2). These compounds displayed low clearance and
good half-life values. Unsurprisingly, both the polar amide and nitrile
variant were not bioavailable orally. This information can be con-
tributed to the poor permeability these molecules possess. The nitrile
variant was synthesized in order to remove two H-bond donors and
improve the overall PK profile, but only a trace enhancement was seen.
This outcome revealed that these potent and safer inhibitors are most
6

D.J. Clausen, et al.
Bioorganic&MedicinalChemistryLetters30(2020)127367
References
1. Grunstein M. Histone acetylation in chromatin structure and transcription. Nature.
1997;389:349.
2. Pazin MJ, Kadonaga JT. What’s up and down with histone deacetylation and tran-
scription? Cell. 1997;89:325.
3. Wolffe AP. Histone deacetylase: a regulator of transcription. Science. 1996;272:371.
4. Yoo CB, Jones PA. Epigenetic therapy of cancer: past, present, and future. Nat Rev
Drug Discov. 2006;5:37.
5. Gregoretti L, Lee Y, Goodson HV. Molecular evolution of the histone deacetylase
family: functional implications of phylogenetic analysis. J Mol Biol. 2004;17.
6. Dokmanovic M, Clarke C, Marks PA. Histone deacetylase inhibitors: overview and
perspectives. Mol Cancer Res. 2007;5:981.
7. Paris M, Porcelloni M, Binaschi M, Fattori D. Histone deacetylase inhibitors: from
bench to clinic. J Med Chem. 2008;51:1505.
8. Shirakawa K, Chavez L, Hakre S, Calvanese V, Verdin E. Reactivation of latent HIV by
histone deacetylase inhibitors. Trends Microbiol. 2013;21:277.
9. Choudhary SK, Margolis DM. Curing HIV: pharmacologic approaches to target HIV-1
latency. Annu Rev Pharmacol Toxicol. 2011;51:397.
10. Deeks SG. Shock and kill. Nature. 2012;487:439.
Fig. 5. X-Ray structure of HDACi O in HDAC2. Access codes for X-ray co-
11. Archin NM, Liberty AL, Kashuba AD, et al. Administration of vorinostat disrupts HIV-
1 latency in patients on antiretroviral therapy. Nature. 2012;487:482.
12. Kinzel O, Llauger-Bufi L, Pescatore G, et al. Discovery of a potent class 1 selective
ketone histone deacetylase inhibitor with antitumor activity in vivo and optimized
pharmacokinetic properties. J Med Chem. 2009;52:3453.
ordinates in RCSB Protein Data Bank (PDB) database are 6XEC.
Table 2
13. Bresciani A, Ontoria JM, Biancofiore I, et al. Improved selective class I HDAC and
novel selective HDAC3 inhibitors: beyond hydroxamic acids and benzamides. ACS
Med Chem Lett. 2019;10:481–486.
Pharmacokinetic Properties of HDACi E and O in rat25.
Entry
Clu (mL/min)/kg
Vdu (L/kg)
t1/2 (h)
F (%)
Papp (10e−6 cm/s)
14. Yu W, Liu J, Yu Y, et al. Discovery of ethyl ketone-based HDACs 1,2, and 3 selective
inhibitors for HIV latency reactivation. Bioorg Med Chem Lett. 2020,:127197.
15. Miyaura N, Suzuki A. Palladium-catalyzed cross-coupling reactions of organoboron
compounds. Chem Rev. 1995;95:2457.
E
166
244
45
40
6.4
4.0
0%
1%
1.1
1.7
O
16. Robak MT, Herbage MA, Ellman JA. Synthesis and applications of tert-butanesulfi-
namide. Chem Rev. 2010;110:3660.
17. Ellman JA, Owens TD, Tang TP. N-tert-butanesulfinyl imines: versatile intermediates
for the asymmetric synthesis of amines. Acc Chem Res. 2002;35:984.
18. Furstner A. Olefin metathesis and beyond. Angew Chem, Int Ed. 2000;39:3012.
19. Schultz-Fademrecht C, Kinzel O, Marko IE, et al. A general approach to homochiral
likely suited for iv dosing.
In summery, potent trisubstituted imidazole HDACi have been de-
veloped that show excellent selectivity for HDAC1, 2, and 3. Their
potency and SAR trends were rationalized through x-ray structure
analysis. These compounds have improved off target profiles compared
to their disubstituted heterocycle parents and acceptable PK parameters
for non-oral administration.
Î
-amino substituted bromo-heteroaromatics suitable for two-dimensional rapid
analogue synthesis. Tetrahedron. 2009;65:9487.
20. Further exploration of the scope of the chemistry underway and interrogation of the
transformation will be published forthcoming.
21. All compounds tested in this publication have purity > 95% determined by LCMS or
HPLC, and their structures are fully characterized by LCMS and 1H-NMR spectra. All
reported values are the mean of two assay measurements. Standard deviation was
within three-fold of the reported value for all compounds.
Declaration of Competing Interest
22. Supporting information contains detailed description of methods used to perform the
bilogical assays.
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.
23. Tyagi M, Pearson RJ, Karn J. Establishment of HIV latency in primary CD4+ cells is
due to epigenetic transcriptional silencing and P-TEFb restriction. J Virol. 2010:6425.
24. Friedman J, Cho W, Chu CK, et al. Epigenetic silencing of HIV-1 by the histone H3
lysine 27 methyltrasferase enhancer of zeste 2. J Virol. 2011;85:9078.
25. PK was in Winstar Han rats at IV 2.0 mpk and PO 10.0 mpk dosed in DMSO/PEG400/
H2O (20/60/20). Compounds in DMSO solution (10 mM) was used.
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://
doi.org/10.1016/j.bmcl.2020.127367.
7