Discovery of 2-Pyridinone Aminals: A Prodrug Strategy to Advance a Second Generation of HIV-1 Integrase Strand Transfer Inhibitors

Article abstract of DOI:10.1021/acs.jmedchem.5b01037

The search for new molecular constructs that resemble the critical two-metal binding pharmacophore required for HIV integrase strand transfer inhibition represents a vibrant area of research within drug discovery. Here we present the discovery of a new class of HIV integrase strand transfer inhibitors based on the 2-pyridinone core of MK-0536. These efforts led to the identification of two lead compounds with excellent antiviral activity and preclinical pharmacokinetic profiles to support a once-daily human dose prediction. Dose escalating PK studies in dog revealed significant issues with limited oral absorption and required an innovative prodrug strategy to enhance the high-dose plasma exposures of the parent molecules.

Full text of DOI:10.1021/acs.jmedchem.5b01037

Subscriber access provided by UNIV MASSACHUSETTS BOSTON
Article
Discovery of 2-Pyridinone Aminals: A Prodrug Strategy to Advance
a Second Generation of HIV-1 Integrase Strand Transfer Inhibitors
Izzat Raheem, Abbas Walji, Daniel Klein, John M. Sanders, David Powell, Pravien Abeywickrema,
Guillaume Barbe, Amrith Bennet, Sophie-Dorothee Clas, David Dubost, Mark Embrey, Jay
Grobler, Michael Hafey, Timothy John Hartingh, Daria J. Hazuda, Michael D. Miller, Keith P.
Moore, Natasa Pajkovic, Sangita Patel, Vanessa Rada, Paul Rearden, John D. Schreier, John
Sisko, Thomas G. Steele, Jean-François Truchon, John Wai, Min Xu, and Paul J Coleman
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.5b01037 • Publication Date (Web): 23 Sep 2015
Downloaded from http://pubs.acs.org on September 26, 2015
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 36
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Discovery of 2ꢀPyridinone Aminals: A Prodrug
Strategy to Advance a Second Generation of HIVꢀ1
Integrase Strand Transfer Inhibitors
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Izzat T. Raheem*‡[a], Abbas M. Walji*‡[a], Daniel Klein[b], John M. Sanders[b], David A.
Powell[a,i], Pravien Abeywickrema[c], Guillaume Barbe[i], Amrith Bennet[d], Sophie
−Dorothee
Clas[j], David Dubost[e], Mark Embrey[a], Jay Grobler[f], Michael J. Hafey[g], Timothy J.
Hartingh[a], Daria J. Hazuda[f], Michael D. Miller[f], Keith P. Moore,[a] Natasa Pajkovic[g],
Sangita Patel,[b] Vanessa Rada[a], Paul Rearden[g], John D. Schreier[a], John Sisko[a], Thomas G.
Steele[a], Jean François Truchon[i], John Wai[a], Min Xu[h], and Paul J. Coleman[a].
−
Departments of Discovery Chemistry[a], Global Structural Chemistry[b], Protein Sciences[c],
Safety Assessment[d], Discovery Pharmaceutical Sciences[e], Infectious Disease Biology[f],
Pharmacokinetics, Pharmacodynamics, and Drug Metabolism[g], and In Vitro Pharmacology[h]
Merck Research Laboratories, West Point, PA 19486.
Merck Frosst Centre for Therapeutic Research[i], Kirkland, QC, Canada.
PharmaSolv Consulting[j], Montreal, QC, Canada.
ACS Paragon Plus Environment
1

Journal of Medicinal Chemistry
Page 2 of 36
1
2
3
4
ABSTRACT
5
6
7
8
9
The search for new molecular constructs that resemble the critical 2ꢀmetal binding
pharmacophore required for HIV integrase strand transfer inhibition represents a vibrant area of
research within drug discovery. Here we present the discovery of a new class of HIV integrase
strand transfer inhibitors based on the 2ꢀpyridinone core of MKꢀ0536. These efforts led to the
identification of two lead compounds with excellent antiviral activity and preclinical
pharmacokinetic profiles to support a onceꢀdaily human dose prediction. Dose escalating PK
studies in dog revealed significant issues with limited oral absorption and required an innovative
prodrug strategy to enhance the highꢀdose plasma exposures of the parent molecules.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
INTRODUCTION
Over the past 20 years, Merck & Co. Inc. has made a strong commitment towards identifying
new antiretroviral (ARV) treatments to help patients infected with the HIVꢀ1 virus. To this end,
we validated strand transfer (ST) inhibition as a new therapy, and in 2007 successfully launched
raltegravir (1, RAL) as the first HIVꢀ1 integrase strand transfer inhibitor (InSTi).1ꢀ3 RAL has
established robust efficacy and tolerability in treatmentꢀnaïve patients and is considered a
preferred firstꢀline ARV agent as part of highly active antiretroviral therapy (HAART).4
The success of RAL represents an important addition to the armamentarium of ARV agents
and has sparked significant interest in developing the next generation of HIV InSTis. The main
challenges in enhancing efficacy and patient compliance are associated with reducing the dosing
frequency to once daily, and improving the genetic barrier to resistance.5 As a result there has
been a surge of research activity to address these limitations with significant contributions from
both academia and industry.4,6,7 The main focus has been on the identification of new chemical
ACS Paragon Plus Environment
2

Page 3 of 36
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
classes able to retain the intrinsic potency and structural elements of a bidentate metalꢀbinding
pharmacophore, which is essential for strand transfer inhibition. These efforts have provided
several distinct structural classes and resulted in the notable discoveries of elvitegravir (2,
ELV)8,9 from Gilead Sciences, dolutegravir (3, DTG)10 developed by a ShionogiꢀViiV
HealthcareꢀGlaxoSmithKline joint venture, and MKꢀ2048 (4)11,12 and MKꢀ0536 (5)13,14 from
Merck & Co. Inc (Scheme 1).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 1. Current leading HIV integrase strand transfer inhibitors: twoꢀmetal binding
pharmacophore highlighted in blue.
ELV (2), approved in 2012, has a reduced dosing frequency to onceꢀdaily (150 mg) when coꢀ
dosed with a cytochrome P450 inhibitor, but confers significant cross resistance to RAL. DTG
(3), approved in 2013, has differentiated potency against viral clones resistant to RAL and ELV,
and does not require pharmacokinetic boosting to enable onceꢀdaily dosing. MKꢀ2048 and MKꢀ
0536 provide a similar superior barrier to resistance as DTG, with MKꢀ2048 exhibiting potential
ACS Paragon Plus Environment
3

Journal of Medicinal Chemistry
Page 4 of 36
1
2
as a microbicide for prevention of HIV infection.12b MKꢀ0536 represents a new structural class
based on the 2ꢀpyridinone core which is currently under investigation in our laboratories.
We have previously disclosed our lead finding and lead optimization efforts culminating in the
identification of MKꢀ0536.14 MKꢀ0536 exhibits excellent antiviral activity as measured in a viral
kinetics assay (IP 50% NHS = 53 nM) and also retains activity against RALꢀresistant mutants
with a maximum 3ꢀfold shift against N155H (Table 1). This robust activity against a wide range
of integrase variants suggests that MKꢀ0536 has the potential for a high genetic barrier for the
development of resistance. Furthermore, it displays a reasonable preclinical pharmacokinetic
(PK) profile in rat (CLp = 2.3 mL/min/kg, %F = 29) and dog (CLp = 2.2 mL/min/kg, %F = 44),
and low intrinsic clearance (CLint) in vitro in human hepatocytes (CLint = <20 mL/min/kg).
Using singleꢀspecies allometric scaling of dog plasma clearance and correcting for the difference
in plasma protein binding between the species, the human plasma clearance was predicted to be
approximately 1 mL/min/kg.15
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 1. Inhibitory potency and mutation profiles of RAL and MKꢀ0536.
Fold shift of IP in potency assay[a] for mutant
virus vs. wildꢀtype (WT) virus (50% NHS)
Compound WT IP[b] E92Q
RAL 9 / 53 nM 3
MKꢀ0536 6 / 53 nM 1
Y143R Q148R Q148K N155H
13
2
16
1
22
1
8
3
[a] kinetic HIV replication assay16 [b] IP in 0 / 50% NHS.
During our discovery efforts towards MKꢀ0536, we identified several structural modifications
within the 2ꢀpyridinone scaffold which provided new opportunities for lead optimization.
ACS Paragon Plus Environment
4

Page 5 of 36
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Herein, we describe the discovery of two lead molecules from within a novel tricyclic 2ꢀ
pyridinone aminal series. In an effort to advance these molecules into preclinical safety studies,
we implemented an innovative prodrug strategy to overcome absorptionꢀlimiting
physicochemical properties and achieve our targeted plasma exposures of the parent compounds.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
RESULTS AND DISCUSSION
A deeper understanding of HIVꢀ1 integrase inhibition and resistance as well as the use of
structureꢀguided design of small molecule HIVꢀ1 integrase inhibitors has been hampered by the
inability to crystallize a ligandꢀbound HIVꢀ1 integrase intasome. However, the structure of the
related prototype foamy virus (PFV) intasome was recently solved.17 While PFV integrase and
HIV integrase share less than 20% sequence identity, there is a high degree of sequence
conservation in their active sites. As such, it is currently believed that PFV integrase structures
serve as reliable models of the HIV integrase active site, and could facilitate structureꢀguided
small molecule inhibitor design. To date, numerous structures of HIV integrase inhibitors bound
to the PFV integrase intasome have been reported.18
Using MKꢀ0536 as a starting point, and guided in part by computational docking of MKꢀ0536
in the PFV integrase active site, we sought to explore structural diversity of the 2ꢀpyridinone
core through fusion of a third ring. Our docked structure of MKꢀ0536 (Figure 1), together with
the crystal structure of RAL, supported the hypothesis that a ring fusion in the northeastern
portion of the 2ꢀpyridinone core would be tolerated by HIV integrase. Additionally, we hoped to
maintain the beneficial PK attributes and optimal mutation profile imparted by the MKꢀ0536
bicyclic framework in these elaborated inhibitors. Over the course of our lead optimization
ACS Paragon Plus Environment
5

Journal of Medicinal Chemistry
Page 6 of 36
1
2
3
4
5
6
efforts, numerous such cyclization strategies were explored, a subset of which are shown in
Figure 2.
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 1. Overlay of modeled MKꢀ0536 (cyan) and crystallographic RAL (orange) in the
catalytic active site of PFV Integrase (PDB 3oya).
MKꢀ0536 scaffold
8ꢀmembered lactam
7ꢀmembered lactam
6ꢀmembered lactam
D.
A.
B.
C.
WT potency
Mutation profile
Dog PK
WT potency
Mutation profile
Dog PK
WT potency
Mutation profile
Dog PK
WT potency
Mutation profile
Dog PK
5ꢀmembered aminal
6ꢀmembered aminal
6ꢀmembered triazinone
Cyclic sulfonamide
E.
F.
G.
H.
WT potency
Mutation profile
Dog PK
WT potency
Mutation profile
Dog PK
WT potency
Mutation profile
Dog PK
WT potency
Mutation profile
Dog PK
Figure 2. Overview of Cꢀring cyclization strategies and general scaffold profiles. General
approximations for profile rankings of scaffolds − WT Potency (IP in 10% NHS): Green= <100
nM, Yellow= 100–250 nM, Red= >250 nM; Mutation profile (IP shift vs. least sensitive
ACS Paragon Plus Environment
6

Page 7 of 36
Journal of Medicinal Chemistry
1
2
3
4
5
6
mutant): Green= <5 fold shift, Yellow= 5–15 fold shift, Red= >15 fold shift; Dog PK (CLp):
Green= <10 mL/min/kg, Yellow= 10–20 mL/min/kg, Red= >20 mL/min/kg.
7
8
9
Due in part to more facile synthetic access, we initially targeted a series of variablyꢀsized
amidoꢀcontaining ring systems. The 8ꢀ and 7ꢀmembered ring systems (Figure 2B and 2C)
provided inhibitors that generally displayed excellent WT potencies and resistance profiles, but
suffered from high unbound clearances (CLun) in dog. The corresponding 6ꢀmembered ring
system (Figure 2D) provided potent inhibitors with compromised mutation and PK profiles. We
were gratified to discover that the fused 5ꢀmembered ring system (Figure 2E) provided an
optimal balance between potency and PK. In general, inhibitors in this structural class were
extremely potent on the wildꢀtype enzyme as measured in the viral kinetics assay16 (IP 0% NHS
< 10 nM), with promising mutation profiles. Importantly, they displayed low CLp and CLun in
preclinical species, akin to MKꢀ0536. Additionally, despite the presence of an aminalꢀtype
functionality, compounds were chemically and configurationally stable upon isolation.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Two distinct homologation strategies of this promising ring system were further
investigated: (a) homologation by one carbon gave the corresponding 2ꢀpyridinone 6ꢀmembered
ring scaffolds (Figure 2F) and (b) homologation by one nitrogen (Figure 2G) afforded the
corresponding triazinone 2ꢀpyridinone. While compounds in these 6ꢀmembered ring classes
exhibited goodꢀtoꢀmoderate mutation profiles, PK properties were typically compromised.
Numerous other nonꢀamido tricyclic cores were also interrogated. One specific example is the
dihydroimidazole ring system highlighted in Figure 2H. This core, like many others
interrogated, provided inhibitors with poor mutation and PK profiles.
Based on this analysis the 5ꢀmembered aminal ring system was selected for further
exploration. Through an extensive lead optimization effort, key points of diversity around the
ACS Paragon Plus Environment
7

Journal of Medicinal Chemistry
Page 8 of 36
1
2
3
4
5
6
7
8
9
periphery of the 2ꢀpyridinone aminal framework were thoroughly interrogated; an SAR overview
is presented in Figure 3. A western aromatic group (R1) is present in nearly all published HIVꢀ1
InSTis, and serves as a crucial hydrophobic πꢀclamp in the catalytic active site. Historically, a
halogenated benzyl moiety in this position has been optimal, and the same was found to be true
for the 2ꢀpyridinone aminals. While some R1 variations were identified that allowed for
modulation of compound physical properties, little structural variation was tolerated, with all
compounds of interest bearing a 4ꢀfluoroꢀ3ꢀchlorobenzyl or 4ꢀfluorobenzyl group.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
While substituent changes at the R2/R3 position did not significantly impact WT potency, they
did provide us the opportunity to modulate other important properties of our compounds. The
SAR at R2/R3 proved to be extremely narrow, with only a select number of small cyclic aliphatic
groups providing acceptable resistance profiles against key single mutants.
Finlly, the R4 position was found to be most tolerant of structural variation, facilitating
optimization of potency, resistance profile, physical properties, and PK parameters. The position
tolerated diversity in size, polarity, functional group, and Hꢀbond donor/acceptor capacity,
allowing us to prepare and characterize a number of distinct leads.
Figure 3. Overview of 2ꢀpyridinone aminal series SAR
ACS Paragon Plus Environment
8

Page 9 of 36
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Our optimization efforts led initially to 6 (Table 2), an important benchmark compound for the
series. Its key structural feature is a novel bicyclo[3.1.0]hexane ring system at the R2/R3 aminal
position; the R1 position maintains the aforementioned 4ꢀfluoroꢀ3ꢀchlorobenzyl group and R4 is
unsubstituted. Beyond its structural novelty, 6 exhibits excellent intrinsic antiviral activity (WT
IP 0% NHS = 2.5 nM), but a large shift in serum potency attributed to high plasma protein
binding (WT IP 50% NHS = 600 nM). 6 displays a good resistance profile against RALꢀresistant
mutants with a maximum 6ꢀfold shift against Q148K. Additionally, it displays low clearance
and moderate oral bioavailability in rat (CLp = 0.12 mL/min/kg, %F = 29) and dog (CLp = 0.15
mL/min/kg, %F = 27), and low Clint in vitro in human hepatocytes (CLint = <5 mL/min/kg).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 2. Inhibitory potency, mutation profile, and pharmacokinetic profile of 6.
Fold shift of IP in potency assay[a] for mutant
virus vs. wildꢀtype (WT) virus (50% NHS)
WT IP[b]
E92Q
3
Y143R
Q148R Q148K
N155H
3
2.5 / 600 nM
1
5
6
Intravenous
Oral
Clp
T1/2
(hr)
14
AUC
(ꢀM•hr)0ꢀinf
species
(IV / PO dose)
rat
(0.2 / 2 mg/kg)
dog
(0.5 / 2 mg/kg)
Vdss
(L/kg)
0.12
F
(%)
(mL/min/kg)
0.12
189
29
0.15
0.19
16
47
27
[a] kinetic HIV replication assay16 [b] IP in 0 / 50% NHS.
ACS Paragon Plus Environment
9

Journal of Medicinal Chemistry
Page 10 of 36
1
2
3
4
5
6
7
8
9
Based on the above SAR analysis of the 2ꢀpyridinone aminal core and the promising profile of
6, further optimization efforts were undertaken employing divergent functionalization strategies
starting from compound 6 (Scheme 2). The objective was to improve the serumꢀshifted potency
by increasing the polar surface area (PSA) at the predefined R2/R3 and R4 positions. For
compounds within this series, we identified the introduction of free hydroxyl groups as optimal
polarityꢀenhancing groups, providing the necessary balance among potency, mutation profile,
and preclinical PK profile. These efforts yielded two lead molecules: compound 7 bearing a
chiral secondary alcohol on the alkyl chain of the aminal nitrogen (R4 optimization), and
compound 8 incorporating a hydroxymethyl substituent (R2/R3 optimization) at the optimal
position on the bicyclo[3.1.0]hexane ring system.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 2. Lead optimization starting from compound 6: strategic introduction of free alcohols at
R2/R3 and R4 to increase polar surface area.
As summarized in Table 3, both compounds maintain excellent potency against the wildꢀtype
virus (WT IP 50% NHS <75 nM), and extremely low foldꢀshifts in potency against RALꢀ
ACS Paragon Plus Environment
10

Page 11 of 36
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
resistant mutants. Coꢀcrystal structures for compounds 7 and 8 bound to the PFV intasome
complex revealed noteworthy interactions previously suggested as important for both potency
and the favorable mutation profile required of a secondꢀgeneration strand transfer
inhibitor.17a Both 7 and 8 adopt lowꢀenergy bioactive conformations, as judged by having
relative ꢁG values <1 kcal/mol as calculated by the FreeForm program,19 and are characterized
by their direct interactions with two Mg2+ ions and burial of the hydrophobic 4ꢀfluoroꢀ3ꢀ
chlorobenzyl groups between the terminal DNA base pair and Pro214 (Figure 4A). Although of
differing stereochemistry, the substituted bicyclo[3.1.0]hexane groups of 7 and 8 each make
increased contact with Gly187 in the β4ꢀα2 loop as compared to MKꢀ0536 and other firstꢀ
generation InSTis and therefore resemble the interactions observed with DTG and MKꢀ2048
(Figure 4B).17a,18a In the observed binding conformations, the polarityꢀenhancing hydroxyl
groups of 7 and 8 make no direct interactions with protein or DNA, consistent with the apparent
selective impact of these groups on improving potency under high serum concentration
conditions. While the protein residues are largely unchanged when compared to other PFV
intasome inhibitor complexes, the 3’ꢀterminal unpaired adenosine, A(ꢀ17), adopts the less
energetically favorable of the two predominant conformations that have been previously
observed (Figure 4A).18b Taken together, the structural data demonstrate that 2ꢀpyridinone
aminals display excellent HIV integrase inhibition even without forming extensive interactions
between the tricyclic core and A(ꢀ17).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment
11

Journal of Medicinal Chemistry
Page 12 of 36
1
2
3
4
Table 3. Inhibitory potency and mutation profiles of 7 and 8.
5
6
7
8
Fold shift of IP in potency assay[a] for mutant
virus vs. wildꢀtype (WT) virus (100% NHS)
9
Compound WT IP[b]
E92Q Y143R Q148R Q148K N155H
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2 / 67 nM 1.5
1
1
1.5
1
1
1
1.5
1
7
8
3 / 32 nM
1
[a] kinetic HIV replication assay16 [b] IP in 0 / 50% NHS.
Figure 4. A) Crystal structure overlay of 7 (magenta) and 8 (green) in the PFV catalytic active
site. B) Overlay of 7, 8, DTG (blue, from PDB 3s3m) and MKꢀ2048 (yellow, from PDB 3oyb),
illustrating inhibitor contacts with β4ꢀα2 loop.
ACS Paragon Plus Environment
12

Page 13 of 36
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Compounds 7 and 8 were tested in intravenous (iv) and oral (p.o.) PK studies in beagle dogs
(Tables 4). After iv administration both compounds exhibited low plasma clearance with
terminal halfꢀlives ranging from 5−6 h. The oral bioavailabilities for 7 and 8 were 42% and
26% respectively. Plasma protein binding for both compounds was high with measured values
of 98.8 and 97.5% in dog and human for compound 7, and 98.6 and 98% in dog and human for
compound 8. The major route of elimination of these compounds was glucuronidation, while
cytochrome P450 (CYP) mediated oxidative metabolism was a minor contributing pathway.
The intrinsic clearances in vitro in human hepatocytes ranged from CLint = 14−26 mL/min/kg.
The compounds were not potent reversible inhibitors of the major isoforms of the CYP enzymes
(3A4, 2C9, or 2D6), nor timeꢀdependent inhibitors of CYP3A4. Both lead compounds exhibited
very similar pharmacokinetic properties in preclinical species to those observed with MKꢀ0536,
and using the same methods as previously described the human plasma clearance was predicted
to be approximately 1 mL/min/kg for both compounds. The steadyꢀstate plasma AUC0ꢀ24hr for
the predicted human onceꢀdaily doses for our leads was between 30 ꢀM•hr and 50 ꢀM•hr.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 4. Pharmacokinetic profiles of 7 and 8 in beagle dog.
Intravenous
Vdss
Oral
AUC
Clp
T1/2
(hr)
5.9
F
(%)
42
Compound
iv/p.o. dose
(ꢀM•hr)0ꢀinf
(L/kg)
0.6
(mL/min/kg)
3.0
0.5 / 2 mg/kg
0.5 / 1 mg/kg
10.5
7
8
1.9
0.4
5
4.9
26
ACS Paragon Plus Environment
13

Journal of Medicinal Chemistry
Page 14 of 36
1
2
3
4
5
6
7
8
9
Chemistry. Lead compounds 7 and 8 are both prepared from trisꢀpivaloylated 2ꢀpyridinone
intermediate 14. The synthesis of 14 is shown in Scheme 3, and begins with known α,βꢀ
unsaturated sulfone 9.4 Conjugate addition of ethyl 2ꢀ((diphenylmethylene)ꢀamino)acetate
followed by imine hydrolysis provides 10, which is treated with DIPEA to effect elimination to
11. Condensation with ethyl oxalyl chloride furnishes the full bicyclic 2ꢀpyridinone core 12,
which is then hydrolyzed and tripivaloylated, to afford key intermediate 14.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 3. Synthesis of trisꢀpivaloylated intermediate 14.
The known chiral ketone building block 18 is synthesized in three steps (Scheme 4). The
reaction of (R)ꢀepichlorohydrin with allyl magnesium chloride and copper iodide furnishes 16,
which undergoes baseꢀmediated cyclization to provide the bicyclic framework. Oxidation of the
secondary alcohol with TPAP/NMO gives ketone 18.
ACS Paragon Plus Environment
14

Page 15 of 36
Journal of Medicinal Chemistry
1
2
Scheme 4. Synthesis of chiral ketone building block 18.20
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
With 14 and 18 in hand, compound 7 is prepared in two steps, as shown in Scheme 5. First, 14
is reacted with commercially available chiral amine 19. Under these conditions, one of the
pivaloyl moieties acylates the primary amine while the other two are hydrolyzed, furnishing 20
in near quantitative yield. Intermediate 20 is then condensed with chiral ketone 18 under forcing
conditions to provide the tricyclic 2ꢀ pyridinone aminal 7.
Scheme 5. Synthesis of 7.
Compound 8 is prepared from intermediate 14 in four steps (Scheme 6). Treatment of 14 with
ammonium hydroxide furnishes primary amide 21, which undergoes condensation with known
chiral ketoꢀester 2221 under conditions similar to those described above. The resulting aminal is
methylated under standard conditions, and the pendant ester is reduced with LiBH4 to furnish 8.
ACS Paragon Plus Environment
15

Journal of Medicinal Chemistry
Page 16 of 36
1
2
3
4
Scheme 6. Synthesis of 8.
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Prodrug Strategy to Address PK Limitations. Our focused lead optimization efforts from
compound 6 were successful in discovering two 2nd generation HIV InSTis, which offered
potential improvements over known InSTis. First, both compounds exhibited superior barrier to
resistance compared to RAL and ELV, and second, they displayed preclinical PK profiles
consistent with onceꢀdaily human dose predictions without the need for a PK enhancer.
However, we uncovered a significant limitation as we evaluated these compounds in nonꢀ
rodent (beagle dog) dose escalating oral PK studies. We observed an early plateau in plasma
exposure following the first 5ꢀfold increase in dose (2 mg/kg to 10 mg/kg) for both lead
compounds (Figure 5); this exposure plateau persisted further as the dose was increased to 30
mg/kg. In light of low plasma clearance and moderate bioavailabilities, we speculated that the
less than dose proportional increase in exposure was likely due to poor physicochemical
properties for both lead molecules. Indeed, this was verified by low crystalline solubility in
fasted simulated intestinal fluid (FaSSiF) and simulated gastric fluid (SGF). Further, in vitro
passive permeability in LLCꢀPK1 and MDCKꢀII cell lines was moderateꢀtoꢀhigh for compound 7
and low for compound 8 (Table 5).
ACS Paragon Plus Environment
16

Page 17 of 36
Journal of Medicinal Chemistry
1
2
3
4
5
6
Table 5. In vitro solubility and permeability of compounds 7 and 8.
7
8
9
Compound Solubility
(FaSSiF)[a]
Solubility
(SGF)[b]
Permeability
(cmꢀ6/sec)
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
LLCꢀPK1
14.8
MDCKꢀII
22.9
0.058 mg/mL 0.03 mg/mL
0.03 mg/mL
7
8
0.017 mg/mL 6.7
8.4
[a] FaSSiF at pH= 6.5, t= RT, time= 4 h. [b] SGF pH= 1.8, t= RT, time= 4 h.
16
15
14
14
12
12.5
11
11
10
8
6
Dosed 7
5
4
Dosed 8
2
0
2
10
30
Dose (mg/kg)
Figure 5. Oral dose escalation studies in dog with compound 7 and 8.
Considering the exceptional virological and PK profiles of 7 and 8, we were now challenged to
address the major constraint of insufficient oral exposure in our preclinical species. As such, our
goal turned to the identification of prodrugs of 7 and 8 that would achieve the highest possible
plasma exposures of the parent compound to enable longꢀterm oral nonꢀrodent safety studies, and
subsequently enable IND human studies. A thorough inspection of the scientific literature22,23
led us to consider several opportunities to address the identified physicochemical limitations.
ACS Paragon Plus Environment
17

Journal of Medicinal Chemistry
Page 18 of 36
1
2
3
4
5
6
7
8
9
The molecular structure of compounds 7 and 8 provided two distinct hydroxyl functional groups
for the introduction of permeability and/or solubility enhancing proꢀmoieties (a and b, Figure 6).
By employing in silico tools,24 we were able to systematically design prodrugs to independently
evaluate the impact of an increase in solubility and then permeability. As part of the discovery
efforts, we evaluated established promoeities known to enhance plasma exposure via
mechanisms involving active transport, colonic absorption, circulating prodrugs, and lipidic
prodrug uptake.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 6. Hydroxyl groups (a and b) for introduction of proꢀmoieties.
Our initial prodrug strategy sought to address the limited aqueous (FaSSiF/SGF) solubilities of
our lead molecules. To this end, we selected parent compound 7 for further interrogation. We
focused our efforts on the installation of solubilityꢀenhancing phosphate and aminoester
promoieties at both positions, but found challenges with chemical stability and synthesis with
such groups installed at the aromatic hydroxyl (b) of the core. However, numerous prodrugs at
the aliphatic hydroxyl position (a) were successfully prepared and characterized, and two distinct
leads emerged, P1ꢀ7 and P2-7, which are highlighted in Table 6. Importantly, both prodrugs
ACS Paragon Plus Environment
18

Page 19 of 36
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
provided the intended marked improvements in FaSSiF solubility to >2 mg/mL. P1-7 exhibited
the anticipated loss in cellꢀbased potency due to the inefficient permeability imparted by the
ionizable phosphate group, while P2-7 maintained its potency due to either efficient uptake of
the prodrug or extracellular esteraseꢀmediated bioconversion to parent.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 6. Profiles of solubilityꢀenhancing prodrugs P1-7 and P2-7.
Compound WT IP[a]
Solubility
Stability at 37 °C
(FaSSiF, SGF)
(FaSSiF)[b]
1580 nM
40 nM
2.3 mg/mL
>2 mg/mL
>98% for 4 h
90% for 4 h
P1-7
P2-7
[a] kinetic HIV replication assay IP in 0% NHS.16 [b] FaSSiF at
pH= 6.5, t= RT, time= 4 h.
We identified efficient means by which to prepare these classes of prodrugs directly and
selectively from their parent compound 7 (Scheme 7). P1-7 was synthesized in high yield by
treatment of 7 with diphosphoryl chloride at low temperature followed by an in situ basic
hydrolysis to the desired phosphate. Correspondingly, P2-7 was prepared by selective
esterification of the secondary hydroxyl group followed by Bocꢀdeprotection to afford the
aminoester prodrug in high yield.
ACS Paragon Plus Environment
19

Journal of Medicinal Chemistry
Page 20 of 36
1
2
3
4
Scheme 7. Synthesis of P1-7 and P2-7.
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The PK profiles of P1-7 and P2-7 in dog were evaluated and are summarized in Figure 7.
Over the timeꢀcourse of these studies, circulating levels of both prodrug and parent were
measured, and due to the known plasma instability of both compounds, collected blood samples
were immediately stabilized with dichlorovos.25 When dosed iv, P1-7 and P2-7 displayed
expected high clearance with no significant detectable levels of circulating prodrug. At a low
oral dose of 2 mg/kg, both compounds provided good exposures of parent (AUC= 6.8−9.6
ꢁM•hr), consistent with those observed from a direct oral dose of 7. Unfortunately, dose
escalation revealed that the previously observed exposure plateau was not sufficiently remedied
through this prodrug strategy. Oral dose ranging studies up to 300 mg/kg provided relatively
little improvement in AUC, with this highest dose providing a maximum 6ꢀfold improvement in
exposure over the 2 mg/kg dose. As a result, efforts toward solubilityꢀenhancing prodrugs were
suspended. These exposure data suggested that parent solubility was not the only factor limiting
ACS Paragon Plus Environment
20

Page 21 of 36
Journal of Medicinal Chemistry
1
2
3
4
5
6
compound absorption, and rather indicated that limited parent permeability may be an additional
culprit.
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
50
43
45
Dosed P1-7;
measured 7
Dosed P2-7;
measured 7
40
35
30
25
20
20
15
10
13
10
5
7
0
2
10
300
Dose (mg/kg)
Figure 7. Oral dose escalation studies in dog with compound P1-7 and P2-7 measuring
exposures of 7. The indicated doses have been normalized to account for the promiety of each
prodrug, and represent the actual dose of parent administered.
We next turned our focus towards the phenolic hydroxyl group (b) of the requisite bidentate
metal binding pharmacophore inherent to all HIV InSTis, including compounds 7 and 8. Based
on our previous experience with prodrugs of RAL,26,27 we suspected that the acidity and the
ionizability of this hydroxyl group was likely impacting the permeability of the parent
compounds. Therefore we devised a similar strategy utilizing the acetal carbonate functionality
as a permeabilityꢀenhancing proꢀmoiety. The acetal carbonate group provided two distinct
points of chemical diversity that could be used to modulate physicochemical properties. Indeed,
using in silico methods, the isopropyl acetal carbonate was identified to provide the greatest
predicted increase in permeability with minimal negative impact on solubility, and was therefore
initially prioritized for synthesis on both parent compounds.
ACS Paragon Plus Environment
21

Journal of Medicinal Chemistry
Page 22 of 36
1
2
3
4
5
6
7
8
9
Chemical synthesis of the acetal carbonate prodrugs was initiated from parent molecules
7 and 8 using commercially available haloalkyl carbonates or haloalkyl carbonates synthesized
using established procedures.28 From parent 7, alkylation by the in situꢀgenerated corresponding
alkyl iodide of chloromethyl isopropyl carbonate provided prodrug P3-7 in good yield (Scheme
8). Prodrug P3-8 was synthesized using a similar procedure starting with parent 8 and
iodomethyl isopropyl carbonate which provided improved chemical yields (Scheme 9).
Synthesis of both prodrugs was robust and provided multiꢀgram quantities for highꢀdose PK
studies.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 8. Synthesis of P3-7.
Scheme 9. Synthesis of P3-8.
ACS Paragon Plus Environment
22

Page 23 of 36
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
P3-7 and P3-8 displayed excellent SGF stability up to 4 hours and provided an increase in
FaSSiF and SGF solubility over their parent compounds (Table 7). As expected based on
literature precedent, the acetal carbonate promoieties underwent rapid esteraseꢀmediated
bioconversion in plasma, and exhibited overall improved in vitro permeability compared to the
parent compounds. Both prodrugs P3-7 and P3-8 met the in vitro selection criteria and were
evaluated in rising oral dose PK studies in dog.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 7. In vitro profiles of permeabilityꢀenhancing prodrugs P3-7 and P3-8.
Compound WT IP[a]
Solubility
Stability at 37 °C
(FaSSiF, SGF)
Permeability
(cmꢀ6/sec)
(FaSSiF)[b]
LLCꢀPK1 MDCKꢀII
178 nM
16 nM
0.25 mg/mL
0.50 mg/mL
>98% for 4 h
>98% for 4 h
26.7
26.5
23.5
19.7
P3-7
P3-8
[a] kinetic HIV replication assay IP in 0% NHS.16 [b] FaSSiF at pH= 6.5, t= RT, time= 4 h.
At an initial 2 mg/kg dose of P3-7, a 2ꢀfold increase in exposure of 7 was observed compared
to exposure from an equivalent 2 mg/kg dose of 7 (23 ꢁM•hr vs. 11 ꢁM•hr) (Figure 8). We were
gratified to observe a doseꢀproportional increase in exposure of 7 to 128 ꢁM•hr after a 5ꢀfold
increase in the P3-7 dose (2 mg/kg to 10 mg/kg). However, a plateau in exposure of 7
manifested after a further 3ꢀfold dose increase of P3-7 (10 mg/kg to 30 mg/kg). We suspected
that the insufficient solubility of P3-7 likely contributed to the early exposure plateau and
therefore initiated efforts towards addressing this limitation (vide infra).
ACS Paragon Plus Environment
23

Journal of Medicinal Chemistry
Page 24 of 36
1
2
3
4
5
6
7
8
9
300
250
200
150
100
50
284
Dosed P3-7;
measured 7
Dosed P3-8;
measured 8
128.0
117
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
23
31
8
0
2
10
30
Dose (mg/kg)
Figure 8. Oral dose escalation studies in dog with prodrugs P3-7 and P3-8 measuring exposures
of 7 and 8. The indicated doses have been normalized to account for the promoiety of each
prodrug, and represent the actual dose of parent administered.
In contrast to P3-7, an initial 2 mg/kg dose with P3-8 did not provide a similar increase in
exposure of 8 compared to exposure from an equivalent 2 mg/kg dose of 8 (5 ꢁM•hr vs. 8
ꢁM•hr). After the standard 5ꢀfold increase in dose of P3-8 (2 mg/kg to 10 mg/kg), a near 4ꢀfold
increase in exposure of 8 was observed. Interestingly, at the next 3ꢀfold increase in dose of P3-8
(10 mg/kg to 30 mg/kg), exposures of 8 reached 284 ꢁM•hr, which was considered to provide
sufficient in vivo exposure of 8 to enable toxicology studies.
To study the relative contribution of hepatic versus extrahepatic conversion of P3-7 and P3-8
to their respective parent molecules, P3-7 and P3-8 were dosed to portal veinꢀcannulated dogs
and the concentrations of prodrugs and parent molecules were determined in plasma collected
from the portal and jugular veins (Tables 8 and 9). Following oral administration of both
prodrugs at doses corresponding to 10 mg/kg of the respective parent molecules, low levels of
P3-7 and P3-8 were measured in the portal vein along with high plasma levels of the respective
ACS Paragon Plus Environment
24

Page 25 of 36
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
parents, suggesting most of the prodrug converts rapidly to parent before reaching the liver. The
prodrug levels diminished quickly in systemic circulation, highlighting efficient in vivo
conversion to both parent molecules.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 8. Profiles of prodrug P3-7 in portal veinꢀcannulated dogs.
10 mg/kg p.o.
Percent (%) P3-7
measured
AUC (ꢀM·hr)
dose P3-7[a]
P3-7
1.12
7
Portal vein
23.9
33.9
5%
Jugular vein
0.034
0.1%
[a] The indicated dose has been normalized to account for
the promoiety, and represents the actual dose of 7
administered.
Table 9. Profiles of prodrug P3-8 in portal veinꢀcannulated dogs.
10 mg/kg p.o.
Percent (%) P3-8
measured
AUC (ꢀM·hr)
dose P3-8[a]
P3-8
2.55
8
Portal vein
27.3
30.4
9%
Jugular vein
0.037
0.1%
[a] The indicated dose has been normalized to account for
the promoiety, and represents the actual dose of 8
administered.
Oral administration of the carbonate acetal prodrug P3-7 provided significant improvements in
highꢀdose parent exposure over direct administration of 7. However, we remained interested in
further exploring the limitations of our prodrug strategy. While we identified and addressed
parent compound permeability as a primary contributing factor to the poor highꢀdose exposures
of 7, the observed exposure plateau of P3-7 (10 to 30 mg/kg doses) suggested to us that
ACS Paragon Plus Environment
25

Journal of Medicinal Chemistry
Page 26 of 36
1
2
3
4
5
6
7
8
9
compromised compound solubility remained a contributing liability (vide supra). Therefore, we
sought to combine two promoieties targeting both solubility and permeability into a single
compound through a “double” prodrug approach to provide additional exposure enhancements.
To this end, we synthesized prodrug P4-7 bearing a phosphate group on the aliphatic hydroxyl
(a) and a carbonate acetal on the aromatic hydroxyl (b).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
As expected, P4-7 maintained a profile similar to the previously described prodrug P1-7, with
poor cellꢀbased potency, excellent solubility, good FaSSiF and SGF stability, and poor plasma
stability (Table 10). Upon oral dose escalation of P4-7 in the dog, we were gratified to achieve
high and dose proportional exposures of parent 7 up to 100 mg/kg (Figure 9). At this highest
dose tested, we were able to attain a parent plasma exposure of 570 ꢁM•hr, the highest oral
exposure achieved to date from this compound class. This significantly exceeded our initial
exposure target, and provided a path by which to also move 7 into toxicology studies and support
compound advancement.
Table 10. Profile of “double” prodrug P4-7.
Compound
WT IP[a]
Solubility
Stability at 37 °C
(FaSSiF, SGF)
(FaSSiF)[b][
>4200 nM >2 mg/mL
>98% for 4 h
P4-7
[a] kinetic HIV replication assay IP in 0% NHS.16 [b] FaSSiF at
pH= 6.5, t= RT, time= 4 h.
ACS Paragon Plus Environment
26

Page 27 of 36
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
600
500
400
300
200
100
0
570
9
Dosed P3-7;
measured 7
Dosed P4-7;
measured 7
Dosed P1-7;
measured 7
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
168
128
117
23
20
2
10
30
100
Dose (mg/kg)
Figure 9. Comparison of oral dose escalation studies in dog with P1-7, P3-7 and P4-7
measuring exposures of parent 7. The indicated doses have been normalized to account for the
promoiety of each prodrug, and represent the actual dose of parent administered.
CONCLUSION
We have described the discovery of two nextꢀgeneration HIV integrase strand transfer
inhibitors, 7 and 8, based on the 2ꢀpyridinone core of MKꢀ0536. Both compounds exhibit
excellent WT potency, a flat mutation profile, and preclinical PK projecting a onceꢀdaily human
dose. In advancing both compounds toward nonꢀrodent safety and tolerabilty studies, we learned
that 7 and 8 suffered from less than dose proportional oral exposure, with plasma concentrations
plateauing below a 10 mg/kg dose. These exposures were attributed to compromised absorption
due to poor physicochemical properties, and a prodrug strategy was undertaken to address these
issues by targeting compounds to independently enhance both parent solubility and parent
permeability. While solubilityꢀenhancing prodrugs on their own did not sufficiently address
ACS Paragon Plus Environment
27

Journal of Medicinal Chemistry
Page 28 of 36
1
2
3
4
5
6
7
8
9
limitations of either parent compound, we were able to achieve high parent plasma exposures by
masking the acidic phenol of the InSTi pharmacophore with a carbonate acetal promoiety.
Carbonate acetal prodrug P3-8 provided high oral exposures of parent 8, while “double“ prodrug
P4-7 (carbonate acetal, phosphate), was required to achieve similar high exposures of 7. Our
innovative prodrug strategy successfully addressed key limitations of 7 and 8, and provided a
path by which to further advance both compounds.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
EXPERIMENTAL SECTION
All experimental procedures, characterization data for new compounds, in vitro and in vivo
experimental protocols, and details on computational modeling and Xꢀray crystallographic data
are available in the Supporting Information section free of charge via the Internet at
http://pubs.acs.org.
1
Compound purity and identity for intermediates were determined by LCꢀMS and/or HNMR.
Compound purity and identity for final compounds were determined by LCꢀMS and 1HNMR. In
all cases, final compounds possess purity ≥95%.
ASSOCIATED CONTENT
Supporting Information. All Supporting Information is available free of charge on the ACS
Publications website at http://pubs.acs.org.
PDB ID Codes. Coordinates and structure factors for PFV Integrase with compounds 7 and 8
have been deposited in the Protein Data Bank29 with accession numbers 4ZTF and 4ZTJ,
respectively.
ACS Paragon Plus Environment
28

Page 29 of 36
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
AUTHOR INFORMATION
Corresponding Author
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
*Eꢀmail: izzat_raheem@merck.com. Phone: (215) 652ꢀ2734. Fax: (215) 652ꢀ7310.
*Eꢀmail: abbas_walji@merck.com. Phone: (215) 652ꢀ3379. Fax: (215) 652ꢀ7310.
Author Contributions
‡ I.T.R and A.W. contributed equally.
ACKNOWLEDGMENTS
The authors thank the Merck West Point NMR and Mass Spectrometry facilities for assistance
in characterizing the compounds presented in this manuscript. The authors also gratefully
acknowledge Dr. Deanne Rudd and Kenneth Anderson for identifying ex vivo prodrug
stabilization conditions to support prodrug bioanalysis. We are grateful to the scientists at
WuXiꢀAppTec for their assistance in synthesizing valuable intermediates and final compounds.
ABBREVIATIONS USED
AUC, area under the plasma concentrationꢀtime curve; CLint, intrinsic clearance; CLp, plasma
clearance; CLu, unbound plasma clearance;
CYP, cytochrome P450; DABCO, 1,4ꢀ
diazabicyclo[2.2.2]octane; DIPEA, diisopropyl ethylamine; DMA, dimethylacetamide; DMF,
ACS Paragon Plus Environment
29