Design, Synthesis, and Pharmacokinetic Evaluation of Phosphate and Amino Acid Ester Prodrugs for Improving the Oral Bioavailability of the HIV-1 Protease Inhibitor Atazanavir

Article abstract of DOI:10.1021/acs.jmedchem.9b00002

Phosphate and amino acid prodrugs of the HIV-1 protease inhibitor (PI) atazanavir (1) were prepared and evaluated to address solubility and absorption limitations. While the phosphate prodrug failed to release 1 in rats, the introduction of a methylene spacer facilitated prodrug activation, but parent exposure was lower than that following direct administration of 1. Val amino acid and Val-Val dipeptides imparted low plasma exposure of the parent, although the exposure of the prodrugs was high, reflecting good absorption. Screening of additional amino acids resulted in the identification of an l-Phe ester that offered an improved exposure of 1 and reduced levels of the circulating prodrug. Further molecular editing focusing on the linker design culminated in the discovery of the self-immolative l-Phe-Sar dipeptide derivative 74 that gave four-fold improved AUC and eight-fold higher C_trough values of 1 compared with oral administration of the drug itself, demonstrating a successful prodrug approach to the oral delivery of 1.

Full text of DOI:10.1021/acs.jmedchem.9b00002

Article
Cite This: J. Med. Chem. XXXX, XXX, XXX−XXX
pubs.acs.org/jmc
Design, Synthesis, and Pharmacokinetic Evaluation of Phosphate
and Amino Acid Ester Prodrugs for Improving the Oral
Bioavailability of the HIV‑1 Protease Inhibitor Atazanavir
,
†
§
‡,∇
‡
†
Murugaiah A. M. Subbaiah,* Sandhya Mandlekar, Sridhar Desikan, Thangeswaran Ramar,
†
†
§
Lakshumanan Subramani, Mathiazhagan Annadurai, Salil D. Desai, Sarmistha Sinha,
Susan M. Jenkins,# Mark R. Krystal, Murali Subramanian, Srikanth Sridhar,
,¶
⊥,¶
§
‡
§
§
§
§
§
Shweta Padmanabhan, Priyadeep Bhutani, Rambabu Arla, Shashyendra Singh, Jaydeep Sinha,
Megha Thakur, John F. Kadow, and Nicholas A. Meanwell
§
∥,¶
∥
†
‡
§
Department of Medicinal Chemistry, Department of Biopharmaceutics, and Department of Pharmaceutical Candidate
Optimization, Biocon-Bristol Myers Squibb R&D Centre, Biocon Park, Bommasandra IV Phase, Jigani Link Road, Bangalore
60099, India
5
∥
⊥
#
Department of Discovery Chemistry and Molecular Technologies, Department of Virology, and Department of Pharmaceutical
Candidate Optimization, Bristol-Myers Squibb Research and Development, P.O. Box 4000, Princeton, New Jersey 08543-4000,
United States
*
S Supporting Information
ABSTRACT: Phosphate and amino acid prodrugs of the HIV-1 protease inhibitor (PI) atazanavir (1) were prepared and
evaluated to address solubility and absorption limitations. While the phosphate prodrug failed to release 1 in rats, the
introduction of a methylene spacer facilitated prodrug activation, but parent exposure was lower than that following direct
administration of 1. Val amino acid and Val-Val dipeptides imparted low plasma exposure of the parent, although the exposure
of the prodrugs was high, reflecting good absorption. Screening of additional amino acids resulted in the identification of an L-
Phe ester that offered an improved exposure of 1 and reduced levels of the circulating prodrug. Further molecular editing
focusing on the linker design culminated in the discovery of the self-immolative L-Phe-Sar dipeptide derivative 74 that gave four-
fold improved AUC and eight-fold higher Ctrough values of 1 compared with oral administration of the drug itself, demonstrating
a successful prodrug approach to the oral delivery of 1.
INTRODUCTION
based on their efficacy and high genetic barrier toward the
■
3−5
development of resistance.
Atazanavir (ATV, 1) is an
Combination antiretroviral therapy (cART), which is com-
posed of HIV-1 inhibitors that target different aspects of the
virus life cycle, is presently the most effective treatment for
azapeptide-based PI that received marketing authorization by
the US Federal Drug Administration (FDA) in 2003 and is a
6
1
designated member of the WHO list of essential medicines. It
HIV-1 infection. These drug combinations, which include
is one of the nine PIs and one prodrug that have been granted
inhibitors of HIV-1 protease, integrase, and reverse tran-
scriptase (nucleoside reverse transcriptase inhibitors (NRTIs)
and non-nucleoside reverse transcriptase inhibitors
4
,7
marketing authorization by the FDA. Although 1 is available
as a monotherapeutic agent for once daily (quaque die, QD)
dosing, it is typically used in conjunction with a PK enhancer
as a fixed-dose combination with either ritonavir [RTV, 3] or,
more recently, cobicistat that allows for a reduced dose of the
(
NNRTIs)), have transformed HIV-1 infection from a fatal
diagnosis to a manageable chronic condition, resulting in a
substantial enhancement of both life expectancy and quality for
8
2
PI while maintaining target Ctrough levels. The profile of 1
most patients. Therapy with cART usually results in long-
offers additional advantages over other PIs, including a more
beneficial effect on lipid profiles and a low capsule burden;
lasting suppression of viremia to undetectable levels that
translates into a restoration of CD4 counts. HIV-1 PIs, which
are in most cases used in combination with a pharmacokinetic
(
PK) enhancer that acts by inhibiting cytochrome P450 (CYP)
Received: January 1, 2019
enzymes, have played a critical role in the evolution of cART
©
XXXX American Chemical Society
A
DOI: 10.1021/acs.jmedchem.9b00002
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
however, 1 brings its own side effect profile which includes
hyperbilirubenimia. As part of a class effect, 1 is known for
concentration of 1 than in the absence of the PK enhancer and
6
allows for a QD rather than a twice daily (bis in die, BID)
13
modest oral bioavailability, with ∼79% of administered
radioactivity recovered in the feces of humans, suggesting
incomplete absorption and/or biliary excretion of the drug.
This is reflected in the PK profile of 1 in preclinical species
which exhibits 15% oral bioavailability in rats when dosed as a
dosing schedule at the preferred dose. However, this also
results in additional drug−drug interactions and related
toxicities, which can also contribute to reduced patient
compliance.
The consequences of the high excipient loading and pill
burden associated with higher doses of 1 and the need for a PK
enhancer inspired a feasibility study toward the development of
prodrug approaches designed to improve the systemic delivery
6
suspension and 36% in dogs from a capsule formulation.
Similar to the observation in humans, the recovery of the
unchanged drug from the feces was significant in both the rat
16
(
39%) and dog (79%) following oral administration. This
of 1. In the context of a contemporary scenario that reflects
limited drug discovery effort by the pharmaceutical industry to
develop next-generation PIs with enhanced PK properties and
profile can be attributed to suboptimal physicochemical
properties that contribute to poor absorption and the
susceptibility of 1 to metabolic modification, particularly by
CYP3A4. Most prominently, 1 is characterized by poor
aqueous solubility at higher pH values, modest absorptive
4
,5
improved resistance and safety profiles,
a strategy of
improving the absorption, distribution, metabolism, and
excretion (ADME) properties of existing PIs via formulation,
9
18−23
permeability, high secretory efflux, and, importantly, an
prodrug, and other approaches assumes significance.
For
extensive first-pass metabolism. Being a weak base, 1 shows
high solubility at pH = 1.0 (relevant for dissolution in the
stomach) but poor solubility at pH = 6.5 (relevant for
solubility and absorption in the intestine), data captured in
Table 1. Following administration of a combination of 1 and 3
example, deuterated analogues of 1, which were prepared by
replacing hydrogen atoms with deuterium isotopes at one or
more sites, including those susceptible to metabolic
modification, have been investigated as a means of enhancing
24,25
metabolic stability.
However, this approach, which takes
advantage of the kinetic isotope effect to slow or redirect
metabolism, resulted in only a marginal improvement in oral
bioavailability, failing to provide a compound amenable to
single pill combination regimens while also not addressing the
Table 1. Physicochemical Properties and PK Profiling of
6
,16,17
1
parameter
value
25
pH-dependent absorption issue associated with 1. In spite of
a
molecular weight
clog D
704.86
4.54
5
a growing interest in the exploration of prodrugs among all
a
26−29
therapeutic classes,
a prodrug approach has only rarely
a
hydrogen-bond donor (HBD) count
hydrogen-bond acceptor (HBA) count
rotatable bond count
polar surface area (PSA)
been investigated as a means of improving the PK profile of
a
16,23
13
1
.
The discovery of a prodrug that would demonstrate
a
18
171 Å
4.7
improved oral exposure of 1 and potentially eliminate the
requirement for the PK enhancer would represent a significant
advancement for this antiviral agent.
a
2
17
pKa
b
aqueous solubility at 25 °C
pH 1.0 = 1.69 mg/mL
pH 3.0 = 0.028 mg/mL
pH 4.0 = <0.001 mg/mL
pH 5.0 = <0.001 mg/mL
pH 6.5 = <0.001 mg/mL
pH 7.4 = <0.001 mg/mL
<0.001 mg/mL
A−B = <15 nm/s
B−A = 470 nm/s
efflux ratio: >31
RESULTS AND DISCUSSION
■
We report herein the synthesis and evaluation of a series of
phosphate- and amino acid ester-based prodrugs of 1 in an
attempt to identify compounds that would confer an improved
ADME profile. To the best of our knowledge, these prodrugs
have never been investigated for 1, although phosphate- and
amino acid-based prodrugs have been explored with other
b
unbuffered aqueous solubility at 25 °C
Caco-2 permeability Papp
b
2
2
PIs. We have previously described a conjugative acyl
migration prodrug approach that improved the oral delivery
b
t1/2 in rat hepatocytes
27 min
rat: 15%
1
by relying upon the sequential enzyme-mediated release of
6
absolute bioavailability
an amine that was specifically designed to trigger an
intramolecular acyl migration to deliver the parent drug.
dog: 36%
16
a
Physical properties calculated using ChemAxon Marvin Sketch
This approach was shown to enable enhanced exposure of 1
based on improved ADME properties by relying upon a
sustained release of drug from the prodrug in vivo whilst also
mitigating the pH-dependent absorption associated with 1. As
part of that initiative to identify effective prodrugs of 1, we
were particularly interested in prodrugs that could be derived
in a more straightforward fashion from the parent drug itself by
taking advantage of the secondary alcohol moiety, a transition
b
16
software. Data generated in-house.
with a 40 mg dose of a proton pump inhibitor (PPI), the
exposure of 1 was reduced by 75%, demonstrating the
10−12
dependence of the absorption of 1 on gut pH.
Because
of this pH effect on absorption, the use of histamine H2
antagonists and PPIs, stomach acid modifiers that raise gastric
pH, in conjunction with 1 requires caution, which can lead to
reduced patient compliance. The pH-dependent absorption
complicates inclusion of 1 in fixed-dose combination regimens,
thereby reducing the clinical utility of the drug. Clinically, 1 is
co-administered with a PK enhancer which suppresses
metabolic degradation by inhibiting CYP3A enzymes. This
approach facilitates an improved PK profile with higher oral
bioavailability, a longer plasma t1/2, and a higher plasma trough
22
state-mimicking pharmacophore, as a prodrug handle. The
design consideration behind the selection of phosphate and
amino acid promoieties was based on the physiological
relevance and acceptable toxicological profile of these entities.
The desired characteristics of a lead prodrug would broadly
include (1) acceptable aqueous buffer stability and solubility at
pH = 6.5 and 7.4; (2) enhanced absorptive flux; (3) reduced
secretory efflux; (4) reduced metabolic clearance associated
B
DOI: 10.1021/acs.jmedchem.9b00002
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 1. Structures of PIs 1, 3, 5, and 7 and their direct phosphate prodrugs (2 is described in this article; 4, 6, and 8 have been reported in the
14,15
literature
).
a
Scheme 1. Synthesis of Direct (2) and Methylene-Bridged (11) Phosphate-Based Prodrugs of 1
a
Reagents and conditions: (a) dibenzyl N,N-diisopropylphosphoramidite, 1H-tetrazole (0.4 M in CH CN), CH Cl , RT, 8 h, then 30% aqueous
3
2
2
H O solution, 62%; (b) CF CO H, anisole, ClCH CH Cl, 0 °C to RT, 16 h, 72%; (c) Ac O, AcOH, dimethyl sulfoxide (DMSO), RT 48 h, 79%;
2
2
3
2
2
2
2
(
4
d) N-iodosuccinimide, H PO , 4 Å molecular sieves, tetrahydrofuran (THF), 0 °C to RT, 30 min, then CH OH, 1 M Na S O , Na CO (solid),
3
4
3
2
2
3
2
3
5%.
with unproductive metabolic pathways other than prodrug
release; (5) quantitative bioconversion to the parent to
maximize drug exposure that also embraces the potential for
a sustained release of parent drug due to a slow rate of cleavage
such that the prodrug acts as a depot; and (6) an acceptable
in vivo toxicological profile for the prodrug, the released
promoieties and the byproducts.
1, which are covalently modified by the attachment of water-
33
soluble nonpeptidic oligomers, have been reported. Phos-
phate prodrugs have been investigated for other marketed PIs
including 3, amprenavir (APV; 5), and lopinavir (LPV, 7),
which are compounds 4, 6, and 8, respectively, with
16
14,15
fosamprenavir (6) approved by the FDA (Figure 1).
Fosamprenavir (6) was designed as a solubility-enhancing
phosphate prodrug of 5 with the objective of allowing
formulation of the molecule with a reduced excipient-to-drug
ratio, thereby decreasing the high pill burden associated with 5
As a potential approach to solving the pharmaceutical issues
associated with the inherently poor aqueous solubility of 1 that
result in dissolution-limited and pH-dependent absorption, the
initial focus was directed toward the synthesis of phosphate
derivatives, with the direct phosphate prodrug 2 selected as the
14
and providing a more patient-compliant regimen. The direct
phosphate prodrug 2 was prepared from 1 in two synthetic
steps that involved derivatization of the alcohol moiety with
dibenzyl N,N-diisopropylphosphoramidite followed by oxida-
tion with H O to afford the dibenzylphosphate intermediate 9
3
0−37
first iteration.
This prodrug was designed to enhance the
solubility of the drug in the gastrointestinal tract, with the
release of the parent drug and nontoxic inorganic phosphate as
a byproduct through a presystemic activation mechanism
mediated by alkaline phosphatases (ALPs) that are abundantly
2
2
(Scheme 1). Unmasking of the phosphate moiety by O-
debenzylation using Pd/C-mediated hydrogenolysis was found
to be cumbersome because of the concurrent cleavage of the
N-4-(pyridin-2-yl)benzyl moiety of the parent. However, this
was resolved by exposing 9 to CF CO H in the presence of
30−32
expressed on the brush border membranes of enterocytes.
While direct phosphate derivatives of 1 have not been
described, phosphate monoesters of structural analogues of
3
2
C
DOI: 10.1021/acs.jmedchem.9b00002
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
a
Table 2. Aqueous Stability and Solubility Data of 1 and Its Prodrugs in the pH Range of 1−6.5
stability (t1/2, h) at 37 °C
solubility (mg/mL) at 25 °C
b
compound #
prodrug moiety
phosphate
clog D (pH 6.5)
pH 1
pH 4
pH 6.5
pH 1
pH 4
pH 6.5
1
2
4.5
1.6
1.8
4.5
4.5
3.9
3.9
3.9
3.9
3.4
4.2
4.5
3.7
3.7
4.9
4.9
4.4
4.4
5.7
5.7
5.4
5.8
5.8
6.1
3.7
3.9
3.9
3.7
5.1
ND
334.3
0.2
ND
ND
250
1.69
>2.65
ND
>1.39
>1.71
>2.15
>2.4
>2.0
>1.71
>1.2
>2.2
>1.4
<0.001
0.25
>0.93
>1.58
>1.95
>1.2
>2.2
>2.0
>1.42
>1.6
>2.4
>1.7
>1.1
>1.7
>0.67
>0.62
>1.53
>0.89
>1.1
<0.001
>2.87
>1.06
0.5
0.67
0.78
334.3
>500
376.3
>500
>500
>500
>500
>500
>500
>334.3
>500
>500
>500
>500
>500
>500
>500
>500
>500
376.3
>500
376.3
>500
>500
9.03
11
23
24
29
30
31
32
44
45
46
47
48
49
50
51
52
53
54
59
64
66
71
72
74
78
81
82
POM
L-Val
>500
334.3
>500
>500
>500
>500
>500
>500
>500
>500
>500
>500
>500
>500
>500
>500
>500
>500
>500
>500
>500
>500
>500
>500
>500
>500
376.3
>500
>500
>500
>500
376.3
376.3
D-Val
L-Val-L-Val-1
D-Val-D-Val-1
D-Val-L-Val-1
L-Val-D-Val-1
Gly
0.82
0.62
0.93
0.96
c
c
39.09
Sar
25
188
0.93
0.36
c
c
N,N-dimethyl-Gly
L-Ala
37.6
37.6
71.6
>1.3
>1.3
0.75
0.52
0.001
0.041
0.044
D-Ala
c
c
L-Leu
>0.45
>0.66
>1.63
>0.99
>1.5
c
D-Leu
ND
103.8
97.1
c
c
L-Met
c
c
D-Met
0.045
L-Phe
>500
>500
214.9
334.3
167.2
0.027
D-Phe
>1.2
>1.2
<0.001
c
c
L-Tyr
>0.98
>1.38
>1.04
>0.99
>2.9
1.17
>2.1
>2.7
>1.2
>1.42
>0.98
>0.91
>0.71
>3.5
ND
>2.3
ND
>2.2
0.009
c
c
L-F-Phe
<0.001
0.001
c
c
L-Trp
N-BnGly
L-Phe-Gly-1
L-Phe-Sar-1
D-Phe-Gly-1
N-Me-L-Phe-Sar-1
N,N-dimethyl-L-Phe Sar-1
334.3
73.4
0.08
88.5
0.08
<0.001
>0.64
ND
0.23
ND
>500
15.6
>500
>334.3
0.02
a
For stability studies, 30% CH CN in buffer was used. Buffers used were 0.1 N aqueous HCl (pH = 1.0), acetate buffer (pH = 4.0), and phosphate
3
b
buffer (pH = 6.5/7.4) at a concentration of 50 mM. ND = not determined. clog D values were calculated using ChemAxon Marvin Sketch
software. pH = 7.4.
c
anisole which facilitated selective O-debenzylation to afford 2.
Attempts to derive 2 directly from 1 in a single step using
moiety at intestinal pH. To understand this in vivo outcome in
rats, 2 was incubated with recombinant rat intestinal ALP
(rIALP) with the result that negligible dephosphorylative
turnover was observed within the incubation period of 120
min. Moreover, 2 was not susceptible to phosphate cleavage in
the presence of human rIALP. On the basis of these in vitro
results, the in vivo observation of the lack of systemic exposure
of 1 from phosphate 2 can be attributed to its inefficient
bioconversion to the parent drug in the intestine. To gain
further insights into the structural impact of 2 on activation by
ALP, rates of conversion of direct phosphate prodrugs of other
PIs including 4, 6, and 8 were considered for a comparative
review. Fosamprenavir (6) has been shown to undergo a high
turnover in the ALP assay (t1/2 = 13.2 min) and also furnishes
systemic exposure of 5 (APV) following oral administration to
rats (Table 3). In contrast, the direct phosphate derivatives 4
and 8 of the PIs 3 and 7, respectively, were not substrates for
ALP-mediated dephosphorylation in vitro and 4 failed to
furnish detectable systemic levels of 3 in rats following oral
administration. An inspection of the structures of these
prodrugs reveals a differentiated architecture associated with
6 that presumably facilitates its recognition by ALP enzymes
when compared with the phosphates of the other PIs that
include 1 (Figure 1). The peptidomimetic core of 1, which
closely resembles that of 3 and 7, represents a pseudo-C2-
symmetric carboxamide-based, dipeptide isostere with an (S)-
POCl were not successful because of the absence of an
3
isolable product.
For prodrug evaluation, we followed a simple, two-level
screening tier that is composed of the determination of
aqueous solution solubility and stability over the pH range of
1
−6.5 followed by an in vivo PK evaluation in Sprague-Dawley
(
SD) rats involving oral administration of the prodrug at a 3
mg/kg-equivalent dose of 1. Both the parent drug 1 and the
prodrugs were formulated as solutions for dosing to rats to
mitigate any potential impact of dissolution on the oral
exposure. The phosphate monoester 2 showed high chemical
stability, with a half-life of at least 250 h over the pH range of
1
−6.5 (Table 2). In contrast to the pH-dependent solubility
15
associated with 1, compound 2 showed good aqueous
solubility across the pH range of 1−6.5, with a ∼250- and
∼
2800-fold improved aqueous solubility at pH = 4 and 6.5,
respectively, compared with 1. The enhanced solubility profile
at the higher pH of 6.5 is especially suited to mitigating the
pH-dependent intestinal absorption of 1. However, 2 failed to
deliver measurable levels of 1 in the plasma of SD rats
following oral administration, in spite of the enhanced aqueous
solubility. Systemic levels of circulating prodrug were
negligible, an observation that can be attributed to the poor
passive permeability of the highly ionized phosphate pro-
D
DOI: 10.1021/acs.jmedchem.9b00002
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Table 3. Comparison of in Vitro ALP-Mediated Cleavage
Half-Life Data and Key Structural Features of Reported
respectively) and robust stability with elevated pH (t1/2 > 500
h at pH values of 4 and 6.5) (Table 2). The solution stability
profile of 11 suggests that the POM moiety may degrade to
some extent in the stomach but will be more stable in the
intestine. Prodrug 11 displayed >900-fold improved solubility
compared with 1 at pH values of 4 and 6.5. Following oral
administration to SD rats, the prodrug was found to release the
parent in vivo, providing systemic exposure of 1 that was only
slightly lower than that following direct administration of 1 in a
precipitation-resistant formulation. The relative oral bioavail-
ability of 1 after dosing of 11 was 72% (Table 4) while
systemic circulation of 11 was found to be negligible, results
that are consistent with cleavage of the POM prodrug at the
brush border membrane of the intestinal mucosa prior to
absorption. Following incubation with rat rIALP, prodrug 11
was found to undergo rapid enzyme-mediated cleavage to give
1, with the t1/2 value of less than 10 min (Table 3). In human
rIALP, the cleavage half-life was 21 min, indicative of a slower
rate of dephosphorylation compared with the rat enzyme. On
the basis of the in vitro solubility and in vivo PK profiles, the
POM prodrug in 11 appears to be a useful approach to
mitigate the solubility-related pharmaceutical issues associated
with 1. However, it may also be inferred from the PK data that
enhanced solubility alone cannot substantially increase the oral
exposure of 1 if there exists a significant role for efflux and
metabolism in influencing oral bioavailability. As a class, orally
administered phosphate prodrugs may not be able to modulate
the barriers of efflux and metabolism because of the lack of
transcellular passive absorption associated with the presence of
Phosphate Prodrugs of HIV-1 PIs with That of the Prodrugs
a
2
and 11 from the Parent Drug 1
parent
exposure
in PK
t1/2 in
rat
rIALP
t
in
1/2
human
rIALP
stereochemistry
of OH
compound #
spacer
study
6
7
NT ,
b c
S
R
none
none
no
yes
13.2
,
bc
min
b c
8
2
NT ,
S
S
none
none
no
no
>120
min
NT)
>120
min
d
d
(
(NT)
d
1
1
>10
min
21 min
S
b
CH2
yes
d
a
NT: no turnover of dephosphorylation. Reported in the literature
ref 15. Calf intestinal ALP. Data are reported in the mean of
triplicates (CV < 10%).
c
d
stereochemistry of the pharmacophoric secondary hydroxyl
1
5
moiety. In contrast, the peptidomimetic skeleton of 5
represents a sulfonamide-based dipeptide isostere with an
(
R)-configuration of the secondary alcohol. Thus, the
accessibility of the promoiety groups of 2, 4, and 8 by ALP
compared with 6 may differ significantly because of the
differences in both the stereochemistry of the secondary
alcohol and the P2′ amide structure in the vicinity of the
phosphate ester moiety. On the basis of the experimental data
obtained with 2 and structural comparisons with 4 and 8, it can
be concluded that unfavorable steric features around the
phosphate of 2 prevented it from effectively engaging with the
active site of ALP in a fashion that allows the enzymatic
dephosphorylation to proceed with efficiency.
The failure of phosphate 2 to undergo cleavage by ALP and
the insights gleaned into the potential steric issues underlying
these observations inspired the design of the homologue 11,
which involved the installation of a methylene spacer between
36,37
the polar ionized promoiety.
Considering the limitations associated with the phosphate
prodrugs 2 and 11, the next phase of prodrug synthesis was
directed toward the design of amino acid-based prodrugs. In
light of the ALP-mediated activation challenges associated with
the direct phosphate prodrug 2, a critical element was to
develop an understanding of whether direct amino acid
prodrugs could undergo enzyme-mediated activation in vivo
and further enhance the oral exposure of 1. Amino acid ester
prodrugs, which can undergo proteolytic cleavage by a number
of specific and nonspecific esterases and peptidases found in
the stomach, intestine, liver, and plasma, have found wide
1
and the phosphate in an effort to project the site of cleavage
further away from the peptidomimetic core and reduce the
steric encumbrance that presumably compromises recognition
by ALP. The phosphonooxymethyl (POM) prodrug 11 was
envisaged to undergo presystemic dephosphorylation to release
an unstable O-hydroxymethyl intermediate that would degrade
to 1 with the extrusion of formaldehyde. This acetal-linked
prodrug strategy was given consideration, in part, based on the
observation that POM prodrugs of 3 and 7 have been reported
27,28,38
application across drug classes and therapeutic areas.
Natural amino acids and their metabolites are useful
promoieties because they are ubiquitous in the human body
and their release in vivo may not pose a significant safety
hazard. This kind of prodrug derivatization has been shown to
improve aqueous solubility, enhance membrane permeability,
modulate metabolic disposition and, in some cases, demon-
34
15,31
to release their parent drugs in vivo in the rat and dog.
The
38,39
POM prodrug 11 was synthesized in two steps from 1, as
depicted in Scheme 1. The first step involved the conversion of
strate prolonged release of the parent drug in vivo.
Some
amino acid-based prodrugs have shown to improve intestinal
absorption by taking advantage of active transport processes
that can enhance the oral exposure of poorly permeable drugs,
exemplified most effectively by valacyclovir (13) and
valganciclovir (15), clinically successful prodrugs of acyclovir
(12) and ganciclovir (14), respectively. In the case of 13 and
15, the enhanced exposure can potentially be attributed to
active transport mediated by oligopeptide transporters that
includes the proton-coupled peptide transporter 1 (PepT1)
expressed in the brush-border membrane of intestinal mucosa
and/or reduced efflux associated with a lower affinity of the
1
into the corresponding O-methylthiomethyl ether 10
through a Pummerer-type rearrangement accomplished by
treatment with DMSO and Ac O. The second step involved
35
2
the conversion of 10 to 11 through an N-iodosuccinimide-
mediated nucleophilic displacement with orthophosphoric acid
that was conducted in the presence of 4 Å molecular sieves
1
5
(
Scheme 1). Because phosphate 11 was found to be
somewhat unstable in the free acid form, it was isolated as
its disodium salt by adding Na CO at the completion of the
2
3
reaction and purifying the crude product using buffer-free,
reversed-phase high-performance liquid chromatography
40
prodrugs to efflux transporters like P-glycoprotein (P-gp).
(
HPLC). Phosphate 11 exhibited poor chemical stability at
Design considerations involving the precedent of Val-based
prodrugs 13 and 15 and previous exploration of the Val
low pH (t1/2 = 12 min and 1.95 h at pH values of 1 and 2,
E
DOI: 10.1021/acs.jmedchem.9b00002
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Table 4. PK Data in SD Rats Obtained after Oral Administration of 1 at 3 mg/kg and Selected Prodrugs at an Equivalent Dose
a
of 3 mg/kg of the Parent Drug
compound
dosed
C
t
AUC
last
relative ATV AUC from the
prodrug (%)
relative total AUC (prodrug +
parent) (fold change)
prodrug/
max
max
(h)
promoiety
POM
analyte (nM)
(nM h)
parent ratio
b
1
1
1
1
1
2
1
2
1
2
2
1
2
3
1
4
1
4
1
4
1
4
1
4
1
4
1
5
1
5
1
5
1
5
1
6
1
6
1
7
1
7
1
7
1
7
1
8
1
266
257
NA
10
1529
6
990
4
30
0.5
0.25
NA
1.7
0.5
1.5
0.6
5.0
3.0
3.0
1.0
0.5
0.3
9.0
1.3
0.5
0.8
0.3
0.3
3
384
NA
0.72
11
23
24
29
277
BLQ
42
4731
23
3845
20
85
72
11
6
NA
113
167
1
3
4
L-Val
12.4
10.1
D-Val
c
c
L-Val-L-Val-1
5
9.6
184
3
9
1179
12
3598
53
c
c
3
0
D-Val-D-Val-1
14
13.9
99
4
0
1430
384
162
617
9.0
8.9
66.8
1407
39
235
3
52
72
241
37
4894
374
224
484
29.6
23.7
45.7
1519
94
423
8
82
130
406
53
44
45
46
47
48
49
51
53
54
59
64
66
71
72
74
78
81
82
Gly
58
8
1.84
0.14
4.10
1.35
0.23
1.40
0.37
3.28
4.50
0.46
2.3
2.2
4
5
6
7
8
9
1
3
4
9
4
6
1
2
4
8
1
Sar
0.8
N,N-dimethyl-Gly
L-Ala
12
24
2
33.2
4.5
2
D-Ala
4.3
1.8
0.5
0.5
0.83
0.42
0.5
0.5
0.5
0.33
0.4
0.3
0.5
0.3
0.8
0.7
0.6
0.3
0.4
NA
1
10.3
3.1
L-Leu
34
14
149
40
27
121
49
26
4
L-Met
1.7
73
90
L-Phe
456
474
64
649
99
573
686
154
1575
103
75
463
430
190
215
100
833
16
BLQ
1533
1
307
1124
92
1.2
D-Phe
10.2
0.73
0.93
1.13
8.33
NA
0.001
3.66
NA
2.80
L-Tyr
63
L-F-Phe
L-Trp
305
288
90
134
55
701
9
NA
794
3
1.1
N-BnGly
L-Phe-Gly-1
L-Phe-Sar-1
D-Phe-Gly-1
N-Me-L-Phe-Sar-1
2.4
NA
4.0
399
80
24
93
0.3
0.5
0.4
0.3
NA
0.5
51
3.7
300
160
NA
285
NA
3.6
BLQ
359
N,N-dimethyl-L-
Phe-Sar-1
8
2
432
0.4
1033
a
Values shown are mean ± SD (n = 3). Parent drug 1 was dosed at 3 mg/kg, while the prodrugs were dosed at an equivalent dose of 3 mg/kg of 1.
NA = not available, BLQ = below limit of quantification. The formulation used for these experiments was 10% v/v DMAc, 40% v/v PEG-400, 50%
b
c
v/v of 30% w/v HPßCD in pH 4.0 citrate buffer (50 mM). The oral bioavailability of 1 = 20%. AUC of the intermediate is also included in the
AUC of the prodrug.
prodrugs 17, 18, and 20 of PIs like 7, saquinavir (16), and
nelfinavir (19), guided the initial investigation of prodrugs of 1
the L-Val ester prodrug 13 which is recognized as a substrate
by PepT1 and unmasked in vivo by valacyclovirase, a serine
hydrolase enzyme that cleaves an α-amino ester bond with
41−44
which focused on valine derivatives.
The oral bioavail-
45
ability of 12 was significantly improved by taking advantage of
high specificity. Stereoisomeric Val amino acid and Val-Val
F
DOI: 10.1021/acs.jmedchem.9b00002
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
demonstrated at least 500-fold higher solubility at pH 6.5 than
1
, suggesting the potential to mitigate not only dissolution-
limited absorption but also pH-dependent absorption.
Although the isomeric prodrugs might be expected to show
different physical and chemical properties because of their
diastereomeric nature, no significant differences in solution
stability or solubility parameters were observed.
dipeptide prodrugs of 16 have been reported to reduce efflux
With these data in hand, select Val-based compounds were
progressed to oral PK studies in SD rats. Although the direct L-
Val and other amino acid prodrugs of PIs like 7, 16, and 19
have been reported, these studies have mainly focused on in
vitro characterization for determining chemical stability,
aqueous solubility, permeability, active transport, and/or
ratio in Caco-2 cells and to improve metabolic stability in rat
42
liver microsomes (RLM) relative to 16. We have briefly
described the L-Val amino acid prodrug of 1 in a previous
publication in the context of designing a traceless acyl
1
6
migration prodrug.
18−20,23,41,44,46−48
metabolic stability.
Because PK data for
most of these prodrugs are not readily available, it is difficult
to assess their susceptibility to bioconversion in vivo and their
capacity to improve the oral bioavailability of their respective
parent PIs. In addition, there is an absence of a description of
the effects of amino acid composition on the PK outcome that
might have provided insights for the present study to guide the
selection of amino acids. As is evident from the data compiled
in Table 4, the amino acid-based prodrugs 23, 24, 29, and 30
delivered lower plasma exposure of 1 than that observed
following the direct administration of the parent drug.
However, a significant systemic exposure of circulating
prodrugs was observed and, in the case of the dipeptide-
based prodrugs, varying levels of mono-amino acid ester
intermediates were also noted. These in vivo PK studies
facilitated three key conclusions: First, the direct Val amino
acid and Val-Val dipeptide prodrugs were not susceptible to a
substantial level of bioactivation, indicating that the crucial
ester bond at the pharmacophoric 2° alcohol of 1 is recognized
poorly by esterases and/or peptidases in the rat. Second, the
observation of the formation of intermediate mono-amino acid
esters in the case of dipeptide derivatives suggests a process of
stepwise degradation of the dipeptide moieties to release
amino acids, with subsequent degradation of these inter-
mediates to release 1. Third, as a result of the observation of a
10- to 14-fold higher relative total AUC (total AUC =
AUCprodrug + AUCintermediate + AUCATV) of amino acid- and
dipeptide-based prodrugs compared with that of 1, it appears
that these prodrugs significantly mitigated the barriers to oral
bioavailability associated with 1, including solubility-limited
absorption, efflux, and, possibly, CYP-mediated metabolism. In
terms of the influence of stereochemistry of the amino acid
moieties, the L-Val prodrug 23 showed slightly better exposure
The diastereomeric Val amino acid prodrugs 23 and 24
derived from 1 were synthesized in two steps involving
esterification of 1 with N-Boc-L-Val or N-Boc-D-Val using
N,N′-dicyclohexylcarbodiimide (DCC) to afford 21 and 22,
respectively, followed by deprotection of the amine using 4 N
HCl in dioxane to give 23 and 24, which were isolated as their
HCl salts (Scheme 2). The four stereoisomeric Val-Val
dipeptide prodrugs 29−32 were synthetically derived from
the corresponding amino acid prodrugs 23 and 24 by coupling
with N-Boc-L-Val or N-Boc-D-Val using the HATU reagent
with subsequent acid-mediated unmasking of the amino group
(
Scheme 2).
The Val-based amino acid and dipeptide prodrugs of 1
showed excellent chemical stability across the pH range of 1−
6
.5, with t values of greater than 330 h, data that are
1
/2
summarized in Table 2. These compounds exhibited high
aqueous solubility in the pH range of 1−6.5, with solubility
decreasing with increasing pH. Importantly, these prodrugs
a
Scheme 2. Synthesis of Stereoisomeric Val-Based Ester 23 and 24 and Val−Val-Based Dipeptide Ester Prodrugs 29−32 of 1
a
Reagents and conditions: (a) N-Boc-L-Val for 21 or N-Boc-D-Val, DCC, 4-dimethylaminopyridine (DMAP), CH Cl , RT, 14 h; yields: 21 = 41%;
2
2
2
2 = 23%; (b) 4 N HCl in dioxane, CH Cl , 0 °C to RT, 1 h; yields: 23 = 85%; 24 = 88%. (c) N-Boc-L-Val for 25 and 27 or N-Boc-D-Val for 26
2 2
and 28, HATU, DIPEA, CH Cl , RT, 14 h; yields: 25 = 45%, 26 = 42%, 27 = 58%, 28 = 64%; (d) 4 N HCl in dioxane, 0 °C to RT, 1 h; yields: 29
=
2
2
76%, 30 = 80%, 31 = 85%, 32 = 81%.
G
DOI: 10.1021/acs.jmedchem.9b00002
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
than the D-Val prodrug 24. In the case of the dipeptide-based
prodrugs, the D-Val-D-Val prodrug 30 showed slightly higher
exposure than the corresponding L-Val-L-Val prodrug 29. In
addition, the D-Val amino acid ester intermediate 24 from the
D-Val-D-Val prodrug 30 was observed in circulation at
significantly higher levels than the L-Val amino acid ester 23
derived from the L-Val-L-Val dipeptide 29, indicating an effect
of stereochemical recognition on the rate of enzymatic
activation of the peptidic bond.
With the limitations posed by the L-Val-based prodrugs, an
expansion of the scope of the amino acid promoiety was
explored in an attempt to further develop SARs associated with
the susceptibility to bioactivation. A small set of 5 different
amino acid ester prodrugs (Gly and its N-methyl analogues,
Ala, Leu, Met, and Phe) was considered. These can be divided
into three categories: amino acids that lack an α-carbon
substituent (Gly and its mono- and di-N-methyl analogues);
amino acids that possess a smaller Cα-substituent than that of
L-Val (Ala); and amino acids that have a larger Cα-substituent
than L-Val (Leu, Met, and Phe). Of particular interest were
amino acid promoieties like Gly and Ala with smaller Cα-
substituents than that of L-Val, based on the hypothesis that
the corresponding prodrugs may be susceptible to a higher
level of bioactivation compared with L-Val because of the
reduced steric hindrance adjacent to the ester moiety. In
support of this hypothesis, brivanib alaninate (BMS-582664,
solubility of greater than 360 μg/mL at pH = 6.5, comparable
to that of 23 (500 μg/mL). In contrast, the Leu, Met, and Phe
esters 50−54 displayed significantly reduced buffer solubility,
in the range of 1−45 μg/mL at pH = 6.5, which can be
attributed to the increased lipophilicity associated with the Cα-
substituents which are bulkier than that in 23.
Select prodrugs were progressed into rat PK studies to
compare the AUC profiles and prodrug-to-parent ratios with
that of the L-Val prodrug 23. The L-Phe derivative 53 showed
significantly improved oral exposure of 1, with a 50% higher
plasma exposure (AUC ) observed than that following direct
last
dosing of 1 and a 14-fold higher plasma AUC of 1 than that
following dosing of 23 (Table 4). The oral exposure of 53 was
also two-fold higher than that of the solubility-enhancing POM
prodrug 11, suggesting potential advantages of amino acid-
based prodrugs to overcome additional barriers to oral
bioavailability. Prodrug 53 also exhibited a significantly
reduced prodrug/parent ratio of 1.2 compared with the ratio
of 113 observed with the L-Val prodrug 23, suggesting that the
L-Phe ester 53 was more readily recognized by proteolytic
enzymes than 23. In addition, 53 showed a nine-fold lower
prodrug/parent ratio when compared with the D-Phe-based
prodrug 54, indicative of a more efficient proteolytic
bioactivation of the natural L-Phe prodrug 53 than the
unnatural D-Phe congener 54 (Table 4). All of the other
prodrugs in this set produced a lower exposure of 1 than that
obtained following the direct administration of parent drug,
although prodrugs 44 (Gly), 47 (L-Ala), and 49 (L-Leu)
showed at least a two-fold better exposure of 1 when compared
with 23. Of the three Gly-based prodrugs examined, 44
exhibited the highest exposure of 1, with the relative AUC
values increasing in the order of Sar (45, 12%) < N,N-
dimethylGly (46, 22%) < Gly (44, 58%). Of the two pairs of
stereoisomeric prodrugs that were evaluated, 47 and 53
showed at least three-fold higher exposure than their
corresponding D-amino acid homologues 48 and 54,
respectively. In terms of relative total AUC (combined
exposures of 1 + prodrug), the D-Phe derivative 54 exhibited
the highest exposure (4.5-fold) followed by the N,N-dimethyl
glycine ester 46 (4.1-fold) and the L-Phe derivative 53 (3.3-
fold). However, the total AUC values associated with these
prodrugs were still lower than that of the Val-based prodrugs
(10−12-fold higher AUC compared with the AUC obtained
following direct dosing of 1).
3
3) has been shown to undergo rapid and efficient enzyme-
mediated hydrolysis, both in vitro and in vivo, to release the
active drug.
4
9,50
Prodrugs 44−54 were synthesized following the two-step
procedure depicted in Scheme 3, and the data for these
compounds are compiled in Table 2 (solubility and stability)
and 4 (PK profiles). This set of prodrugs showed excellent
stability (t1/2 greater than 330 h) at pH = 1 and 4 but stabilities
varied at pH = 6.5, with t1/2 values ranging from 25 to 500 h
(Table 2). With the exception of the Phe derivatives, these
prodrugs displayed lower stability at pH 6.5 (t1/2 = 25−188 h),
when compared with the Val-based prodrugs. The solution
stability at pH = 6.5 increased in the order of Gly ≈ Ala < Leu
Encouraged by the improved plasma exposure of 1 following
dosing of the L-Phe derivative 53 at a dose of 3 mg/kg, the PK
performance of the prodrug in a dose-escalation study was
investigated (Table 5). At doses equivalent to 6 mg/kg and 30
mg/kg of 1, prodrug 53 exhibited approximately 2-fold and 34-
fold higher exposure of the parent drug compared with the
exposure obtained following administration of 53 at a dose
equivalent to 3 mg/kg of 1. Dosing of 53 afforded 4-fold and
65-fold higher total exposure of 1 + prodrug at 6 and 30 mg/kg
equivalent doses, respectively, when compared with the
exposure obtained following the direct dosing of 1 at a dose
of 3 mg/kg. While the systemic exposure of 1 after dosing at
the equivalent of 6 mg/kg of 53 was dose-proportional, the
oral exposure at 30 mg/kg dose was more than dose-
proportional. Prodrug 53 exhibited a decreasing prodrug-to-
parent ratio with increasing dose, indicating a higher level of
bioactivation as the dose was escalated. At 3- and 6-mg/kg-
equivalent doses, the trough concentration of 1 from dosing
with 53 measured at 24 h post-dose was below the limit of
<
Met < Phe ≈ Val, suggesting a trend of decreasing hydrolytic
susceptibility with increasing steric crowding around the amino
acid ester bond of 1. Among the Gly-based prodrugs, the
chemical stability increased in the order of Sar (45) < Gly (44)
<
N,N-dimethyl-Gly (46) at pH = 6.5, indicating that the
introduction of one methyl group resulted in a slightly lower
stability, but the introduction of two methyl groups resulted in
∼4-fold improved stability when compared with the Gly-based
prototype 44. As was observed in the case of Val prodrugs,
there were no significant differences in buffer stability and
solubility between the diastereomeric prodrugs. All of these
prodrugs showed very high solubility at pH = 1 and 4 (≥450
μg/mL) but more variable solubility at pH = 6.5 (<1 to 930
μg/mL). Amino acid prodrugs like 44−46 (Gly and its N-Me
derivatives) and 47 and 48 (Ala), which had smaller Cα-
substituents than the L-Val derivative 23, showed aqueous
H
DOI: 10.1021/acs.jmedchem.9b00002
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Scheme 3. Synthesis of Amino Acid Prodrugs of 1 in Which Smaller or Larger Amino Acid Residues than That of L-Val Are
a
Explored
a
Reagents and conditions. (a) N-Boc-amino acid except in the case of 46, for which dimethylgycine was used, DCC, DMAP, CH Cl , RT, 14−16
2
2
h; yields: 34 = 76%, 35 = 64%, 46 = 57%, 36 = 64%, 37 = 60%, 38 = 37%, 39 = 42%, 40 = 49%, 41 = 45%, 42 = 74%, 43 = 71%; (b) for all N-Boc
derivatives: 4 N HCl in dioxane, 0 °C to RT, 1−2 h.; yield: 44 = 89%, 45 = 19%, 47 = 98%, 48 = 99%, 49 = 76%, 50 = 80%, 51 = 94%, 52 = 94%,
5
3 = 85%, 54 = 95%.
a
Table 5. Dose-Escalation PK Data for the L-Phe Prodrug 53 at Doses Equivalent to 6 and 30 mg/kg of 1
C
t
AUC
last
(nM h)
Ctrough at 24 h relative ATV exposure from
relative total AUC (prodrug +
parent) (fold change)
prodrug/
parent ratio
max
max
(h)
b
compound dose analyte (nM)
(nM)
the prodrug (%)
1
5
3
3
1
1
5
1
5
1
5
266
456
474
550
874
4720
2314
0.5
0.5
0.5
0.5
0.2
0.8
0.2
384
573
686
913
680
19670
5409
0.25
BLQ
3
149
3.3
4.1
1.2
0.7
0.3
3
3
3
6
3
BLQ
1.0
238
0
5122
65.3
a
Values shown are mean ± SD (n = 3). NA = not available, BLQ = below limit of quantification. The formulation used for these experiments was
b
1
0% v/v DMAc, 40% v/v PEG-400, 50% v/v of 30% w/v HPßCD in pH 4.0 citrate buffer (50 mM). Dose for 53 = equivalent dose of the parent.
quantification (BLQ), although 1 was detectable at a
concentration of 1 nM following a 30 mg/kg-equivalent
dose. While the parent drug 1 furnished a Cmin of 0.25 nM at 3
mg/kg dose, the absence of measurable trough levels of 1 from
Table 6. Extended aqueous Solubility Profiling of the L-Phe
Prodrug 53
a
aqueous solubility
(mg/mL)
5
3 indicates the inferiority of this prodrug compared with the
fold improved solubility of
parent on this important PK parameter. One of the important
criteria for the selection of a prodrug candidate is that it should
exhibit a Ctrough of 1 that is comparable to or improved over the
Ctrough obtained following direct administration of the parent.
Having identified 53 as the lead prodrug from this cohort, a
set of profiling studies was undertaken that were designed to
understand pharmaceutical aspects, mechanism of bioactiva-
tion, and key safety aspects associated with hepatic uridine
diphosphate glucuronosyltransferase 1A1 (UGT1A1) inhib-
ition. The aqueous solubility of 53 was evaluated in additional
buffers where it differentiated from 1 in the pH range of 3−6.5,
with 27- to 4100-fold improved solubility (Table 6). The
medium
53
1
53 vs 1
pH 1 buffer
pH 3 buffer
pH 4 citrate buffer
pH 4 acetate buffer >4.1
pH 5 buffer
pH 6.5 buffer
pH 7.4 buffer
FaSSIF
>1.5
>3.2
>1.1
1.69
0.028
<0.001
<0.001
<0.001
<0.001
<0.001
ND
114
1100
4100
1300
27
1.3
0.027
0.001
0.13
FeSSIF
>0.5
ND
a
ND = not determined.
I
DOI: 10.1021/acs.jmedchem.9b00002
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
dissolution and solubility profile of 53 in biorelevant simulated
small intestinal media which more effectively mimic the
physiological conditions of the small intestine than simple
buffers was also evaluated because this can be a useful
approach to predicting the performance of a formulation in
vivo while assessing the potential for a food effect as part of the
modeling of the drug absorption process (Table 6). In both
fasted-state simulated intestinal fluids (FaSSIF) and fed-state
simulated intestinal fluids (FeSSIF), which contain natural
solubilizers like bile salts and lecithin in amounts similar to
natural intestinal fluids, 53 displayed an acceptable solubility of
parent drug (Table 7). In the presence of PMSF, the rate of
metabolic depletion of the prodrug in rat hepatocytes was
reduced further, with the t1/2 value extending to 77 min.
Following co-incubation of 53 in rat hepatocytes in the
presence of ABT, the half-life exceeded the assay time of 120
min, indicating that the metabolism of 53 was almost
completely halted under these assay conditions. The combined
presence of both PMSF and ABT in rat hepatocytes further
improved the metabolic stability of 53 (Table 7 and Figure 4a,
the Supporting Information). The significant enhancement in
the elimination half-life of 53 in the presence of ABT
compared with that in rat hepatocytes alone suggests that
the prodrug is subject to direct metabolic clearance mediated
by CYP 450 enzymes. In addition, the impact of esterase
inhibition on the elimination rate of 53 was significantly less
than the impact of CYP inhibition, indicating that the CYP-
mediated metabolism was substantially higher than esterase-
mediated degradation.
51
0
.13−0.5 g/mL.
Prodrug 53 was found to be stable following incubation with
simulated gastric fluid (SGF) containing pepsin, an endopep-
tidase that is one of the main digestive enzymes produced in
the stomach (see Figure 3, the Supporting Information).
Pepsin is known to specifically cleave peptides at amino acid
residues L-Phe, L-Met, L-Leu, and L-Trp when adjacent to a
hydrophobic amino acid and the high gastric stability of 53
may be linked to the lack of susceptibility of the prodrug ester
The formation of 1 from 53 was also monitored when the
prodrug was incubated in rat hepatocytes alone or in the
presence of PMSF and/or ABT (Table 7 and Figure 4b, the
Supporting Information). The generation of 1 from 53 was
minimal in the presence of PMSF alone or with a mixture of
PMSF and ABT in rat hepatocytes, results that suggest that
PMSF inhibits the esterase-mediated cleavage of 53 to deliver
5
2
bond to pepsin. However, 53 was found to be unstable in
simulated intestinal fluid (SIF) containing pancreatin, with the
formation of 1 as the only product observed following
incubation. This suggests that pancreatin, which is a mixture
of several digestive enzymes (including lipase, amylase, and
protease) produced in the pancreas by exocrine cells and is
mainly responsible for the intestinal breakdown of peptides,
might have contributed to the cleavage of the ester prodrug
1
. In contrast, the abundance of 1 was maximal in the presence
of ABT alone, with the amount of 1 observed in the presence
of ABT reduced by more than 90% by the addition of PMSF.
This observation indicates that suppression of unproductive
metabolism of 53 and/or 1 by ABT in the absence of
inhibition of esterase-mediated prodrug conversion led to the
highest levels of 1 in rat hepatocytes. Combining these results,
it can be concluded that the formation of 1 from 53 is
predominantly mediated by esterase-dependent prodrug
activation. Although the in vitro system demonstrates efficient
cleavage of 53 in hepatocytes, conversion of 53 into 1
following IV dosing is lower when compared with PO dosing,
suggesting the potential for a contribution from intestinal
metabolism in prodrug activation. Assuming proteolytic ester
cleavage of 53, the prodrug is expected to release L-Phe as the
byproduct in vivo, a ubiquitous natural amino acid that should
not present a safety concern. However, regulation of the intake
of L-Phe may be required in those individuals who express a
genetic polymorphism in phenylalanine hydroxylase, an
52
moiety of 53 in SIF. These data also suggest the potential for
significant metabolic activation in the intestine and are
consistent with the high intestinal turnover observed in rat
closed loop studies (vide infra).
To further understand the mechanistic aspects of bio-
conversion, the metabolic susceptibility of 53 was evaluated in
four different experiments as follows: (a) incubation in rat
hepatocytes in the absence of additives; (b) incubation in rat
hepatocytes in the presence of the esterase inhibitor phenyl-
methylsulfonyl fluoride (PMSF); (c) incubation in rat
hepatocytes in the presence of the CYP 450 inhibitor 1-
aminobenzotriazole (ABT); and (d) incubation in rat
hepatocytes in the presence of both PMSF and ABT. PMSF
is known to inhibit proteases (e.g., trypsin, chymotrypsin,
thrombin, and papain) by reacting with the active site serine
residue. As a non-isoform-specific and mechanism-based
inhibitor of CYP enzymes, ABT is widely used to determine
the relative contribution of oxidative metabolism in preclinical
54
enzyme that metabolizes L-Phe to L-Tyr in the human body.
An assessment of the activation of 53 in the intestine using a
closed loop rat intestinal model was conducted to understand
the bioconversion of the prodrug prior to or during absorption.
The in situ intestinal perfusion technique in rodents has been
useful as a model to understand the mechanisms underlying
5
3
studies. Significant turnover of 53 was observed in control rat
hepatocytes, with a half-life of 42 min (Table 7 and Figure 4,
the Supporting Information). However, this half-life was longer
than that of the parent (27 min), indicating a modest reduction
in the metabolic clearance of the prodrug compared with the
5
5
intestinal drug absorption.
The introduction of the
mesenteric vein cannulation, which allows for the determi-
nation of the appearance of compounds in the blood, is
particularly useful for evaluating mechanistic aspects of the
absorption of compounds and prodrugs undergoing intestinal
metabolism. Following cassette dosing of 53 and the low
permeability marker nadolol at a dose of 1 mg/kg to rats, the
observed levels of 1 in the mesenteric vein plasma were about
Table 7. In Vitro Metabolic Stability Profile of L-Phe
Prodrug 53 in Rat Hepatocytes in the Absence or Presence
of Esterase and/or CYP Inhibitors
additive
degradation data
PMSF
ABT
t1/2 (min)
kobs (min−1)
5
0% of prodrug 53, suggesting the occurrence of significant
no
yes
no
no
no
yes
yes
42
77
>120
>120
0.0164
0.0009
0.0042
0.0014
prodrug activation in rat intestine (Table 8 and Figure 5, the
Supporting Information). The parent drug 1 was also observed
to be produced in intestinal tissue (29% of the prodrug 53)
and intestinal contents (50% of 53). In the case of nadolol, the
yes
J
DOI: 10.1021/acs.jmedchem.9b00002
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Table 8. Assessment of the Activation of 53 in the Intestine
Using a Closed Loop Intestinal Model Conducted in the
Rat
concentrations of both the released parent 1 and 53 at the
efficacious dose.
a
With these data in hand, the design considerations for
further optimization of the profile of 53 focused on analogues
that encompassed both relatively simple structural variations
that included substitution on the phenyl ring of the promoiety
and more complex variations in which the phenyl ring was
replaced with a heteroaryl variant or the benzyl group was
translocated from the α-carbon to the amino group. This phase
of the study explored four different amino acid ester prodrugs,
all bearing an aryl group, with the objective of further
optimizing the prodrug performance as a means of improving
exposure and, particularly, the trough concentration of 1
measured at 24 h postdose. The L-Tyr prodrug 59 was
explored in an effort to understand whether a potential
metabolite of 53 would offer differentiation in terms of rate of
in vivo conversion to the parent. This is in line with the results
of biotransformation studies conducted with 53 which
suggested that hydroxylation of L-Phe promoiety was one of
the metabolic pathways. Compound 59 was synthesized in 4
steps as depicted in Scheme 4. The phenolic hydroxyl of N-
Boc-L-Tyr 55 was protected as its silyl ether by treatment with
TBDMS-Cl to afford 56 which was converted to the ester 57
by DCC-mediated coupling with 1. Treatment of 57 with
TBAF removed the silicon protecting group and afforded
intermediate 58 which was subjected to amine deprotection on
treatment with 4 N HCl in dioxane to afford the L-Tyr
derivative 59.
recovered percentage of dose of prodrug administered
intestinal
mesenteric vein
plasma (%)
homogenate
(%)
intestinal
recovery
(%)
compound
content (%)
5
3
2.8 ± 0.02
1.4 ± 0.2
4.2
5.6 ± 1.0
1.6 ± 0.2
7.2
32 ± 3.8
16.1 ± 1.9
48
41 ± 2.8
19 ± 1.9
60
1
total
nadolol
1.3 ± 0.1
14 ± 1.0
45 ± 11
61 ± 10
a
Cassette dosing of the prodrug 53 and nadolol at 1 mg/kg each;
formulation vehicle = PEG 400/TPGS/pH 6.5 buffer (25:5:70);
values shown are mean ± SD (n = 3).
amount absorbed into plasma was 1.3% and recovery was 61%.
Overall, these data imply that in the rat, the intestine plays a
significant role in the conversion of 53 to 1. While the ratio of
prodrug 53 to parent 1 in plasma in the closed loop study was
∼
2:1 (32% of parent 1; 68% of 53) following a dose of 1 mg/
kg, the relative plasma concentration of parent and prodrug in
the conventional oral PK study was 1:1. In addition, the
prodrug was found to be highly stable in rat blood in vitro,
with a t1/2 value of greater than 120 min. Taking all of these
factors into account, it can be suggested that the initial
formation of 1 from 53 occurs in the intestine with further
conversion occurring in the liver.
Patients who are treated with 1 can experience asympto-
matic hyperbilirubinemia because 1 is known to inhibit
UGT1A1, thereby interfering with the glucuronidation and
The effect introducing a fluorine substituent to the phenyl
ring of the L-Phe element of 53 was examined because the
strategic incorporation of this element can productively
modulate several aspects of molecular properties that
56,57
elimination of bilirubin.
The potential for inhibition of
bilirubin glucuronidation by the L-Phe prodrug 53 was assessed
in vitro following separate incubations with recombinant
human and cynomolgus monkey UGT1A1 protein as well as
human and RLMs. For the purpose of comparison under the
same experimental conditions, UGT1A1 IC50 data for 1 were
also generated. As is evident from the data compiled in Table
encompass molecular conformation, pK , lipophilicity, intrinsic
a
potency, permeability, metabolic pathways, and PK proper-
58−61
ties.
Thus, the 4-fluoro-Phe derivative 64 was prepared
because this substitution may influence oxidative metabolism
of 53 should that occur in vivo. Prodrug 64 was synthesized in
four steps, as delineated in Scheme 5. The readily available
amino ester intermediate 60 was protected as its Boc derivative
9
, prodrug 53 showed four- to nine-fold lower inhibition of
6
1, which in turn was subjected to hydrolysis using LiOH to
afford the N-Boc amino acid 62. DCC-mediated coupling of
2 with 1 afforded 63 which was converted to the final
Table 9. Cross-Species UGT1A1 Inhibition Profile of 1 and
the L-Phe-Based Prodrug 53
6
enzyme
prodrug 53
IC (μM)
parent 1
prodrug
compound 64 following acid-aided removal of the Boc group.
The L-Trp prodrug 66 was prepared from 1 following the
two-step procedure captured in Scheme 6 while the N-
benzylglycine (N-BnGly) prodrug 71 was given consideration
as part of the effort to understand the potential impact of
migrating the Cα-benzyl group from carbon to nitrogen on PK
properties. This prodrug was synthesized in four steps from
commercially available N-benzylglycine ethyl ester, as depicted
in Scheme 6.
substrate
bilirubin
preparation
IC50 (μM) selectivity
5
0
human rUGT1A1
2.8
11.9
0.3
0.7
9.3
17.0
cynomolgus
monkey
rUGT1A1
β-estradiol rat rUGT1A1
12.6
2.5
19.5
4.2
0.6
3.4
3.0
4.2
5.7
HLM
RLM
Compounds 59, 64, 66, and 71 exhibited acceptable stability
across the pH range, although these prodrugs showed lower
stability than 53 at pH = 6.5 (Table 2). However, these
prodrugs showed improved stability compared with the Gly,
Ala, Leu, and Met prodrugs 44, and 47−52, suggesting that the
presence of an aromatic group in the amino acid residue
contributes to reduced hydrolytic susceptibility at pH = 6.5.
While 59, 64, 66, and 71 showed excellent solubility of at least
700 μg/mL at pH = 1 and 4, these prodrugs showed
significantly reduced solubility of less than 10 μg/mL at pH =
6.5, attributable to the negative impact of the presence of a
UGT1A1 when compared with 1. In a human rUGT1A1 study,
3 exhibited ∼9-fold lower inhibitory potency than 1 while in
human liver microsomes (HLM), 53 demonstrated ∼4-fold
reduced potential for inhibition of glucuronidation than 1. The
lower activity of prodrug 53 toward UGT1A1 inhibition
compared with the parent was consistent in other species
including the cynomolgus monkey (rUGT1A1) and rat
5
(
rUGT1A1 and RLM). However, the potential advantage of
the reduced inhibitory effect on UGT1A1 needs to be carefully
considered in the context of the Cmax and free drug
62
hydrophobic aryl group. This series of prodrugs exhibited
K
DOI: 10.1021/acs.jmedchem.9b00002
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
a
Scheme 4. Synthesis of the L-Tyrosine-Based Amino Acid Ester Prodrug 59
a
Reagents and conditions. (a) TBDMS-Cl, imidazole, DMF, RT, 16 h, 62%; (b) 1, DCC, DMAP, CH Cl , RT, 16 h, 68%; (c) TBAF, THF, 0 °C, 1
2
2
h, 72%; (d) 4 N HCl in dioxane, 0 °C to RT, 1 h, 84%.
a
Scheme 5. Synthesis of 4-Fluorophenylalanine Amino Acid Ester Prodrug 64
a
Reagents and conditions. (a) (Boc) O, Et N, CH Cl , RT, 2 h, 95%; (b) LiOH, THF, H O, 0 °C, 2 h, 93%; (c) 1, 62, DCC, DMAP, CH Cl , RT,
2
3
2
2
2
2
2
1
6 h, 58%; (d) 4 N HCl in dioxane, 0 °C to RT, 1 h, 74%.
Scheme 6. Synthesis of L-Trp Amino Acid Ester Prodrug 66
and N-Benzylglycine Amino Acid Ester Prodrug 71
conversion in vivo resulting in an improved AUC profile of
1 from 64. Unfortunately, none of these prodrugs offered an
advantage in terms of higher parent exposure over the lead
prodrug 53.
a
While 53 and 64 offered the benefit of improving the oral
exposure of the parent, scope existed to further improve the
oral bioavailability of 1 from these prodrugs and importantly
demonstrate measurable levels of 1 at 24 h postdose that is
critical for the systemic suppression of viral load. In this
direction, the installation of a linker between 1 and the L-Phe
promoiety that was simpler than the L-Val elements studied
earlier was explored in an effort to improve the efficiency of
release of parent from the prodrug (Scheme 7). This design
concept was inspired by published observations that dipeptide
ester prodrugs can undergo intramolecular cyclization at
physiological pH in which the terminal amine the dipeptide
moiety attacks the ester carbonyl group, leading to the release
of the parent drug along with the formation of a
63−65
diketopiperazine byproduct.
This chemo-activatable pro-
drug strategy was considered to offer the potential to address
the bioactivation challenges that were encountered with the
direct amino acid ester prodrugs discussed earlier. As a
precedent, neladenoson bis-alanate, a clinical prodrug
candidate of the partial adenosine A1 receptor agonist
neladenoson, has recently been shown to significantly improve
the oral bioavailability of the parent compound without the
presence of circulating prodrug. The stability pattern of the
Val-Val dipeptides gave some insight into the design of the
requisite linker. For example, the Val-Val dipeptide 29
displayed very high aqueous stability of t1/2 of more than
500 h at pH = 6.5, indicating that this prodrug is not able to
efficiently engage in an intramolecular cyclization to generate
the parent 1 under the conditions of elevated pH. This can be
attributed to the high steric crowding around the ester bond of
29 because of the presence of the isopropyl element adjacent
to the Val ester. In the case of dipeptide prodrugs of 12, it has
been shown that the rate of decomposition is significantly
a
Reagents and conditions. (a) N-Boc-Trp, DCC, DMAP, CH Cl ,
2
2
RT, 14 h, 52%; (b) 4 N HCl in dioxane, 0 °C to RT, 1 h, 68%. (c)
(
Boc) O, Et N, CH Cl , 0 °C to RT, 12 h, 92%; (d) LiOH, THF,
2
3
2
2
H O, 0 °C, 2 h, 95%; (e) 69, DCC, DMAP, CH Cl , RT, 16 h, 89%;
2
2
2
(
f) 4 N HCl in dioxane, 0 °C, 1 h, 95%.
66
plasma exposure of the parent that increased in the order 71 <
9 < 66 < 64 following oral administration to rats (Table 4).
5
Prodrug 64 demonstrated a higher plasma AUC of 1 than that
of the direct administration of the parent drug but the trough
concentration of 1 from 64 was BLQ. The introduction of a 4-
fluoro or 4-hydroxy substituent on the phenyl ring of 53 did
not result in an improved exposure of 1. Considering the total
AUC of parent and prodrug, the L-F-Phe (64) and N-Bn-Gly
(
71) promoieties were found to be better than the promoieties
L-Tyr (59) and L-Trp (66). While 64 and 71 provided a similar
total AUC, 64 was found to undergo ∼8-fold higher
L
DOI: 10.1021/acs.jmedchem.9b00002
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Scheme 7. Linker-Enabled Chemo-Activatable Prodrug Design To Improve the Bioactivation of Prodrug 53 with the Prodrug
Moieties Depicted in Blue
Scheme 8. Synthesis of L-Phe-Gly-1 (72), D-Phe-Gly-1 (78), L-Phe-Sar-1 (74), and the N-Me and N,N-Dimethyl Homologues
a
8
1 and 82 as Potential Dipeptide Ester-Based Prodrugs of 1
a
Reagents and conditions. (a) 44, N-Boc-L-Phe for 76 or N-Boc-D-Phe for 77, HATU, DIPEA, CH Cl , RT, 8 h; yields: 76 = 64%; 77 = 65%; (b)
2
2
4
=
5, N-Boc-L-Phe for 79, N-Boc-L-N-MePhe for 80, or L-N,N-dimethylPhe for 82, HATU, DIPEA, CH Cl , RT, 8 h; yields: 79 = 81%; 80 = 32%; 82
35%; (c) 4 N HCl in dioxane, 0 °C to RT, 1 h; yields: 72 = 72%; 78 = 93%; 74 = 98%; 81 = 87%.
2 2
decreased following the incorporation of sterically hindered
amino acids into the dipeptide backbone and that the effect is
more pronounced when a bulky β-branched amino acid like L-
Val is present as the C-terminal residue that is directly attached
rate of ring closure, allowing a measure of control over drug
release. Thus, the amide conformer topology would be
influenced by N-methylation which would assist in bringing
the terminal amine and ester carbonyl group in close
6
7
to the drug. It was postulated that Gly, which does not
possess a Cα-substituent, might be a good choice as a linker to
promote a facile, self-immolative internal cyclization of 72 to
generate the diketopiperazine 73 and parent 1, as depicted in
Scheme 7. The Gly−Phe dipeptide prodrugs of various drugs
that include acyclovir have previously been investigated as
carrier prodrugs for delivering the active parents via a
65
proximity, thereby enhancing the propensity for cyclization.
Taking all of these factors into consideration, Gly and Sar were
chosen as the amino acid linkers, leading to the design of L-
Phe-Gly-1 (72) and L-Phe-Sar-1 (74) as potential prodrugs.
On the basis of the design principle, these prodrugs would be
expected to be stable in the stomach as the hydrochloride salt
but hypothesized to convert to a free base in the intestine or
during or after absorption, triggering a facile intramolecular
cyclization to liberate the parent 1 and the diketopiperazine
byproduct 73 or 75.
67
nonenzymatic process. While the kinetics associated with
self-immolation of prodrug 72 are inherent to the structure, it
was hypothesized that the introduction of a substituent on the
Phe−Gly amide NH would provide a means of modulating the
M
DOI: 10.1021/acs.jmedchem.9b00002
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
A set of 5 prodrugs was prepared by the routes depicted in
Scheme 8 to develop SARs around solution stability, aqueous
solubility, the rate of cyclization, and the potential for in vivo
delivery, as measured by the exposure of 1 and prodrugs in
plasma following oral administration. While prodrugs 72 and
prodrug to undergo facile intramolecular chemo-activated
cyclization. The facility with which the L-Phe-Gly prodrug 72
cyclizes to release 1 contrasts with the relative stability of the L-
Val-L-Val derivative 29, indicating that the absence of the Cα-
alkyl moiety enables intramolecular cyclization through
population of a cis-amide conformation favorable for
cyclization. While prodrugs 72 and 78 showed excellent
solubility of at least 2000 μg/mL at pH values of 1 and 4, these
prodrugs demonstrated acceptable solubility of more than 200
μg/mL at pH = 6.5, representing a more than 200-fold
improvement in solubility compared with 1. The installation of
the Gly linker in prodrugs 72 and 78 was found to offer more
than 20-fold and 200-fold increased solubility, respectively,
compared with the monoester counterparts 53 and 54 that lack
the linker. Prodrugs 74 and 81 demonstrated excellent aqueous
solubility of more than 1000 μg/mL at pH = 1, but the
solubility could not be determined at pH = 4 and 6.5 because
of the limited chemical stability under these conditions.
However, the enhanced stability of 82 allowed determination
of the solubility across the pH range with this prodrug,
demonstrating solubility of greater than 1200 μg/mL at pH 1−
7
8 were prepared from 44 in two steps, the remaining
prodrugs 74, 81, and 82 were prepared from the readily
available Sar prodrug 45 in two steps. Along with 72 and 74,
the synthesis of the three additional prodrugs 78, 81, and 82
was undertaken with the primary objective of broadening the
SARs associated with these cyclization-activated prodrugs. The
D-Phe-Sar-1 prodrug 78 was explored to understand whether
the stereochemical switch in the promoiety would differentiate
this diastereomer from 72 in terms of cyclization kinetics and
oral exposure profile. Prodrug 81 was designed to understand
the impact of N-methylation of the nucleophilic amine on
modulating conversion kinetics of 74 and on plasma exposure
of 1, whereas the N,N-dimethyl-L-Phe-Sar-1 prodrug 82 was
evaluated specifically as a means of gleaning insight into the
behavior of a prodrug incapable of undergoing cyclization-
based self-immolation at physiological pH due to blockade of
the amino terminus. Contrary to the possibility of chemical-
mediated activation mechanism of 72, 74, 78, and 81,
conversion of prodrug 82 would be expected to proceed by
either an enzyme-mediated cleavage or by a general base-
catalyzed process to deliver the parent in vivo.
4
although the solubility was a more modest 20 μg/mL at pH
6.5.
=
Following oral administration of 72, 74, 78, 81, and 82 at a 3
mg/kg-equivalent dose of 1, these prodrugs were found to
deliver the parent in vivo, with the measured AUC values of 1
increasing in the order of 72 < 81 < 78 < 82 < 74, results that
are compiled in Table 4. Prodrug 74 demonstrated the highest
exposure of 1, with a relative oral AUC of ∼400% compared
with that of 1, indicating a superior performance compared
with all of the other prodrugs that have been investigated as
part of this study. Importantly, 74 delivered an eight-fold
higher Ctrough concentration of 1 in the rat measured at 24 h
postdose compared with administration of 1 in contrast with
the earlier prodrugs, none of which furnished a measurable
Ctrough level of 1. The relative oral bioavailability of 1 from 74
was four-fold higher (80%) compared with that of 1 following
direct administration (20%). The prodrug/parent ratio
measured for 82 was 2800-fold higher than that of 74. This
indicated that the bioconversion of 82, which was mainly
designed to undergo enzymatic hydrolysis, was a slow process,
reflecting the challenges associated with prodrug activation
mediated by proteases. However, the PK profile of 82 suggests
that this prodrug may be suitable for a sustained release of the
parent. The D-Phe-Gly-1 prodrug 78 afforded 20% higher oral
exposure than the stereoisomeric L-Phe-Gly-1 prodrug 72
while the introduction of methyl group on the N-terminal
amino group of 74 was found to be less favorable, with the
observation of 17-fold reduced systemic exposure of parent
from 81.
Encouraged by the superior plasma exposure and Ctrough
value of 1 released from the L-Phe-Sar prodrug 74 at an
equivalent dose of 3 mg/kg, the PK performance of the
prodrug in a dose-escalation study was investigated (Table 11
and Figure 6, the Supporting Information). At a dose
equivalent to 30 mg/kg of 1, prodrug 74 exhibited 11-fold
higher exposure of 1 compared with the exposure obtained
following administration of 74 at a dose equivalent to 3 mg/kg
of 1. This indicates that the systemic exposure of 1 at a 30 mg/
kg-equivalent dose was dose-proportional. Prodrug 74 also
delivered a 16-fold higher Ctrough level of 1 at the 30 mg/kg-
equivalent dose of 1 than that of the 3 mg/kg-equivalent dose.
Systemic circulation of 74 was not observed, indicating a
While the L-Phe (53) and D-Phe (54) derivatives exhibited
high aqueous stability across the pH range of 1−6.5, the
introduction of a Gly linker, as in the case of 72 (L-Phe-Sar-1)
and 78 (D-Phe-Sar-1), resulted in more than five-fold reduced
stability at pH 6.5, with no impact on the high chemical
stability observed at pH values of ≤4 (Table 2). Prodrug 78,
which incorporates an unnatural D-Phe in the promoiety (t
=
1
/2
8
9 h), revealed slightly higher stability than the diasteromeric
prodrug 72 with a natural L-Phe in the promoiety (t1/2 = 74 h).
The reduced stability of the dipeptide prodrugs at pH 6.5 can
be attributed to the tendency to undergo deprotonation at
higher pH values which will facilitate intramolecular cyclization
to release 1. Self-immolative activation was further accelerated
in the L-Phe-Sar prodrug 74 where the N-methyl group would
lead to significant population of the cis-amide topology which
is primed for cyclization to the diketopiperazine to release 1.
At a pH value of 6.5, L-Phe-Sar-1 (74) cyclized completely
within 60 min, with a t1/2 estimated to be on the order of 5
min. The intramolecular cyclization of 74 in pH 6.5 buffer was
confirmed with the observation of formation of the
corresponding diketopiperazine 75 by LCMS as the only
byproduct (Scheme 7 and Table 10). Monomethylation of the
nucleophilic terminal amine of 74 gave 81 which offered no
differentiation in solution stability at pH = 6.5; however, the
dimethyl homologue 82 was considerably more stable at pH
values of ≥4 than 74, consistent with the inability of this
Table 10. Rate of Formation of the Cyclic Byproduct 75
along with 1 Associated with the Degradation of Prodrug 74
a
at pH = 6.5 in Buffer
time (min)
2
74
75 (%)
1 (%)
86.1%
2.0%
3.8
29.7
31.0
10.0
68.4
69.0
3
6
0
0
a
NA
a
NA = data not available.
N
DOI: 10.1021/acs.jmedchem.9b00002
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Table 11. PK Data for the Prodrug 74 Following Dose Escalation
a
dose (mg/kg) analyte Cmax (nM) tmax (h) AUClast (nM h) Ctrough (nM) at 24 h relative AUC of 1 from the prodrug (%) prodrug/parent ratio
3
1
7
1
7
794
3
8756
BLQ
1
0.3
1
1533
1
16580
BLQ
1.9
399
0.001
4
BLQ
30.4
BLQ
3
0
4318
NA
4
BLQ
a
The prodrug 74 was dosed at an equivalent dose of 3 mg/kg and 30 mg/kg of the parent. Values shown are mean ± SD (n = 3). Abbreviations
used: NA = not applicable or not available, BLQ = below limit of quantification.
Table 12. Evaluation of the Permeability of Select Amino acid Prodrugs across a Caco-2 Monolayer
Caco-2 P
app
b
(
nm/s)
AUC (nM h)
HBD/HBA
efflux
total AUC(prodrug+ATV)/
AUCATV
a
a
a
compound
promoiety
L-Val
clog D (pH 6.5)
count
PSA
A−B B−A
ratio
prodrug
1
1
4.5
4.5
4.5
5.7
5.4
5.8
6.1
5/7
5/7
4/8
5/8
6/9
5/8
5/8
171
205
181
203
223
203
189
<15
<15
<15
<15
<15
<15
<15
203
122
132
131
71
>14
384
42
23
46
53
59
64
71
>8.1
>8.8
>8.7
>4.7
>5.5
>6.4
4731
1519
686
75
430
833
12.4
4.0
3.3
0.46
2.3
2.4
N,N-dimethyl-Gly
L-Phe
45.7
573
103
463
100
L-Tyr
4-F-L-Phe
BnGly
82
96
a
b
Physical properties (clog D, HBD count, HBA count, and PSA) were calculated using ChemAxon Marvin Sketch software. Targeted
concentration is 2 μM. Transport medium: 4% HSA in HBSS; permeability values have been calculated based on prodrug appearance only.
Negligible levels of 1 were observed at the end of the incubation.
robust reversal of the prodrug to 1 following dose escalation.
The potential in vivo byproduct 75 was not monitored in these
studies and although this compound has been reported in the
were similar to or higher than that of 1 (Table 12). However,
the prodrugs differentiated from 1 in that they showed reduced
potential for transepithelial efflux, suggestive of lower affinity of
the prodrugs for the efflux transporters when compared with 1
(Figure 7, the Supporting Information). Amongst the
prodrugs, 46 showed the highest efflux potential despite
having the lowest HBD count (4), whereas 59, which has the
highest HBD count (6), exhibited the lowest efflux ratio,
suggesting an absence of a correlation between efflux and HBD
count. With the exception of the L-Tyr prodrug 59, all of the
prodrugs evaluated showed a 3- to 12-fold improvement in
total oral systemic exposure of prodrug and 1 in rats, when
compared with the direct administration of 1. On the basis of
the in vitro Caco-2 permeability data, it can be hypothesized
that reduced efflux partly contributes to the in vivo observation
of improved total AUC of prodrug + 1. This is supported by
the observation of a 12-fold increase in the brain-to-plasma
ratio (Cbrain/Cplasma) following IV administration of 1 to mice
treated with the P-gp inhibitor elacridor compared with
vehicle-treated mice. Surprisingly, the Tyr prodrug 59 did
not significantly improve the total AUC of prodrug + 1,
although it showed the lowest efflux potential out of the
prodrugs tested. Unlike other prodrugs, the Tyr prodrug 59
possesses a phenolic functionality that has calculated a pKa
value of 9.5, indicating that ionization in the intestine as a
limiter of oral absorption can be ruled out.
An analysis of the complete set of the solubility data over the
pH range of 6.5−7.4, which is considered to be the most
relevant for intestinal absorption, indicates that the phosphate
prodrugs 2 and 11 exhibited superior solubility, with the direct
phosphate ester 2 showing the highest solubility (Figure 8, the
Supporting Information). The direct phosphate prodrug 2
exhibited two-fold higher solubility than the POM prodrug 11,
indicating that the introduction of a methylene spacer
contributes to slightly reduced aqueous solubility. Among the
amino acid prodrugs, those compounds with no amino acid
side chain elements 44 (Gly), 45 (Sar), and 46 (N,N-dimethyl-
68
literature, its toxicological profile remains to be understood.
The potential safety concerns associated with the prodrug 74
itself appear to be limited because of the absence of systemic
circulation of the prodrug.
The parent drug 1 appears to be the substrate for several
transporters which, in turn, contributes to reduced oral
systemic bioavailability, intracellular concentration, and brain
bioavailability. As a class, PIs are the substrates for the ATP-
binding cassette family of drug efflux transporters including P-
gp, multidrug resistant proteins (MRPs), and breast cancer
resistance protein (BCRP) and also anionic and cationic
transporters like organic anion transporting polypeptide
9
,69,70
(
OATP) and organic cation transporter (OCT).
These
transporters are differentially expressed at the important sites
of drug disposition, including the intestine, lymphocytes, liver,
brain, and kidney and, hence, play a significant role in the
disposition of PIs, influencing oral absorption, tissue
distribution, and cellular accumulation. Amino acid prodrugs
of PIs have previously been shown to enhance absorptive
permeability by engaging an active transport system and
69
2
3
reducing efflux. The apparent bidirectional Caco-2 perme-
ability coefficients for 1 and select prodrugs was determined in
an effort to understand whether the prodrugs could evade
efflux transporters. The Caco-2 permeability assay represents
an important in vitro system with which to predict human
intestinal permeability and the absorption of oral drugs. In the
co-solvent Caco-2 assay, 1 showed low apical-to-basolateral
and high basolateral-to-apical permeability (Table 12) with the
secretory flux of 1 >14-fold higher than its absorptive flux.
These results are consistent with the literature characterization,
9
indicating that 1 is a substrate for efflux transporters.
Surprisingly, none of the tested prodrugs showed improved
apical-to-basolateral permeability compared with the parent,
irrespective of whether the cLog D values of these prodrugs
O
DOI: 10.1021/acs.jmedchem.9b00002
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Gly) or with smaller residues like Me (Ala derivatives 47 and
stereoisomeric pairs studied, there was no consistent trend
with respect to the particular configuration of the amino acid
moiety.
4
8) and iPr (Val derivatives 23 and 24) exhibited high
solubility, with a more than >200-fold advantage compared
with 1. Those prodrugs with either a more lipophilic side chain
In the early stages of the program, the prodrugs of 1 were
tested for HIV-1 protease inhibition and antiviral activity in a
HIV-1 replicating virus assay hosted by MT2 cells (see Table
13 and Figure 11, the Supporting Information). Although
compounds with prodrug elements attached to the pharmaco-
phoric hydroxyl moiety would not be expected to inhibit HIV-
1 protease in a biochemical assay, the prodrug may be subject
to cleavage in cell-based antiviral assays. Although the initial
prodrugs prepared were tested in both assays, there was not a
significant correlation between the activity observed in the
biochemical and cell-based assays because factors including the
permeability of the compound and activation of the prodrug by
cellular enzymes are expected to play a role in the expression of
cellular antiviral activity. As a consequence, later prodrugs were
profiled only in the cell-based assay. Control compound 1
exhibited potent inhibition in the enzymatic assay (IC50 = 0.3
nM) as well as in the cellular assay (EC50 = 0.7 nM). Those
prodrugs tested in the HIV-1 protease assay expressed
inhibition with IC50 values in the range of 4−1145 nM, with
cellular EC50 values observed across the range of 0.9−1250
nM. Prodrugs 23, 24, 29, 30, 44, 46, and 48 displayed
enzymatic and cellular inhibitory activities that represented a
less than three-fold difference between IC50 and EC50 values.
The other prodrugs 31, 32, 47, 53, and 54 showed a significant
19- to 42-fold difference between the IC50 and EC50 values.
Prodrugs 47, 53, and 54 exhibited markedly higher HIV-1
inhibitory activity in the cell-based assay compared with the
IC50 values. In general, L-amino acid-based prodrugs exhibited
better cellular anti-HIV-1 activity than D-amino acid-based
prodrugs (compare 23 and 24, 47 and 48, 49 and 50, 51 and
(
Leu and Met derivatives 49−52) or an aryl group (Phe
derivatives 53, 54, and 64, Trp derivative 66, Tyr 59 and the
Bn-Gly derivative 71) showed significantly decreased solubility
that was in the range of 1−45 μg/mL. The dipeptide ester
prodrugs 29−32, 72, and 78 demonstrated consistent increases
in solubility of at least 230 μg/mL in spite of the presence of
two isopropyl moieties or an aromatic group. The only
exception was the dipeptide prodrug 82 bearing a 3° amine,
which showed at least 10-fold reduced solubility that can be
attributed to the lack of 1° or 2° amine group. Consistent with
this, the tertiary amine-based prodrugs 46 and 82 exhibited
∼
2.5-fold and 32-fold lower solubility than their primary amine
counterparts 44 and 72, respectively. Expansion of the mono-
Val amino acid esters 23 and 24 to the Val-Val dipeptide esters
2
9−32 did not exert a significant impact on the aqueous
solubility.
An analysis of the AUC of 1 released from the prodrugs
indicated that Phe-based, direct amino acid prodrugs with (S)-
stereochemistry or Phe-based, dipeptide promoieties with the
exception of 72 and 81 showed AUC values of 1 in the range
of 307−1533 nM h, a profile that is similar to or higher than
the AUC obtained after direct administration of 1 (Figure 9,
the Supporting Information). These results suggest that the L-
Phe derivatives 53 and 64 are more readily recognized by
hydrolytic enzymes than the other amino acid-based prodrugs,
contributing to the release of a higher amount of 1 from these
prodrug moieties. The remainder of the amino acid prodrugs
without a Cα substituent (Gly and its derivatives 44-46) or
with a Cα substituent (47 and 48 (Ala), 23 and 24 (Val), 51
and 52 (Met), 66 (Trp), and 49 and 50 (Leu)) showed a
lower AUC exposure of 1 when compared with dosing of the
parent drug. Of the three L- and D-stereisomeric pairs of amino
acid prodrugs 23/24, 47/48, and 53/54 prepared, the L-amino
acid prodrugs released a 2- to 12-fold higher amount of 1 in
vivo. However, this trend was reversed in the case of the L- and
D-stereoisomeric pairs of dipeptide prodrugs 29/30 (Val-Val-
5
2 and 53 and 54 in Table 13, the Supporting Information),
consistent with the hypothesis that the natural amino acid
esters can undergo cellular activation more readily to release
the parent than the unnatural counterparts. Another SAR
pattern in HIV-1 inhibitory activity that was apparent was the
markedly lower cellular activity of Val mono-amino acid esters
and Val-Val dipeptide esters (EC50 values range from 84 to
1
250 nM) than the other amino acid esters (EC values range
1) and 72 and 78 (Phe-Sar-1)
50
from 0.7 to 17 nM), as evident from the data compiled
graphically in Figure 10. Thus, those prodrugs which have no
branching on the β-carbon of the amino acid showed higher
antiviral activity than the Val prodrugs which possess β-
branching. One of the possible reasons behind this observation
is that the prodrugs that lack branching on the β-carbon of the
amino acid moiety may be subject to a higher rate of enzymatic
cleavage to release 1 which mediates the antiviral activity in cell
culture. Additional studies are required to confirm if amino
acid or dipeptide prodrugs are intrinsically active or if the
activity is derived from the parent after hydrolytic conversion
of the prodrugs to 1 in HIV infected cells.
An analysis of the total AUC comprising the sum of the
AUC of 1 arising from the prodrug and AUC of the circulating
prodrug was undertaken to better understand the overall
performance of these prodrugs in mitigating the physiological
barriers for oral drug delivery. The prodrugs examined
furnished a 1.1- to 14-fold higher relative total AUC of 1
compared with the AUC that was obtained after oral dosing of
1
1
, with the exception of 11, 45, 48, 51, 59, 72, and 81 (Figure
0, the Supporting Information). These data further revealed
the superior performance of the Val-based amino acid
derivatives 23 and 24 and the dipeptide prodrugs 29 and 30,
which offer 10- to 14-fold higher total oral AUCs than when
dosing 1 directly. This indicates that prodrugs with Val or Val-
Val residues are the most effective at mitigating the barriers to
oral bioavailability through a combination of enhanced
solubility, reduced efflux and reduced metabolism. However,
these prodrugs do not appear to offer advantages in improving
the absorptive permeability, another barrier for bioavailability.
The Phe-derived direct prodrugs 53, 54 and 64 and dipeptide
prodrugs 74, 78, and 82 also displayed higher relative total
AUC values of 1 with a 2- to 4-fold enhancement compared
with administration of a parent drug. With respect to the
CONCLUSION
■
A survey of the in vitro and in vivo performance of a series of
phosphate and amino acid-based prodrugs of the HIV-1 PI 1
identified the L-Phe-Sar ester 74 as a prodrug that overcomes
some of the pharmaceutics and absorption-related limitations
associated with 1. The design of 74 takes advantage of a self-
immolative process of drug release in which cyclization of the
promoiety to a diketopiperazine occurs at neutral pH and is
facilitated by the presence of a sarcosine linker to exert
P
DOI: 10.1021/acs.jmedchem.9b00002
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
conformational bias toward the presence of the cis-amide
conformation. Oral dosing of 74 to rats provided a four-fold
improved exposure and an eight-fold improved C24h value of 1
compared with an equivalent dose of the parent drug with the
observation of only very low levels of circulating prodrug. In a
dose escalation study, a dose-proportional increase in both
plasma concentration and trough concentration of 1 was
observed following oral administration of 74 at doses of up to
method C = 6.40 min and 99.60% method D = 5.16 min and 99.80%;
+
HRMS (ESI/Orbitrap) m/z: [M + H] calcd for C H N O P,
38 54
6
10
7
85.3634; found, 785.3622.
Methyl ((5S,8S,9S,14S)-8-Benzyl-9-((bis(benzyloxy)phosphoryl)-
oxy)-5-(tert-butyl)-15,15-dimethyl-3,6,13-trioxo-11-(4-(pyridin-2-
yl)benzyl)-2-oxa-4,7,11,12-tetraazahexadecan-14-yl)carbamate
(9). To a stirred solution of 1 (1 g, 1.42 mmol) in CH CN (15 mL)
3
were added tetrazole (0.099 g, 1.42 mmol) and dibenzyl
diisopropylphosphoramidite (0.72 mL, 2.13 mmol). The reaction
mixture was stirred at room temperature for 8 h and cooled to 0 °C,
and an aqueous solution of H O (0.222 mL, 2.84 mmol) was added.
3
0 mg/kg. While the performance and profile of 74 in rats is
2
2
encouraging, realization of its full potential as a delivery vehicle
for 1 will require additional PK studies in higher species to
more fully understand the relevance to human exposure
predictions. The pH-activatable immolative prodrug design
described herein appears to offer a viable strategy for
facilitating the activation of prodrugs to release the parent
drugs in those circumstances where hydrolytic enzymes may
have difficulty in accessing the promoieties because of the
steric hindrance presented by structural elements embedded in
the parent molecule and may have broader applications.
The reaction mixture was stirred at 0 °C for 10 min and then
partitioned between EtOAc and H O. The organic layer was washed
2
with 10% aqueous Na CO solution and brine, dried over anhydrous
2
3
Na SO , filtered, and concentrated under vacuum to afford a colorless
2
4
oil. The crude product was purified by preparative HPLC (column:
YMC TRIAT (20 × 250 mm) 5 μm; mobile phase A: 10 mM
NH OAc in H O; mobile phase B: CH CN; flow rate: 20 mL/min;
4
2
3
gradient (T/%B): 0/60, 2/60, 15/80). The pooled HPLC fractions
were concentrated under reduced pressure below 30 °C. The residue
was dissolved in a mixture of CH CN and H O, frozen, and
3
2
lyophilized for 12 h to obtain the title product as an off-white solid
1
(
900 mg, 62.4%). H NMR (400 MHz, CH OH-d ): δ 8.63−8.56 (m,
3
3
EXPERIMENTAL SECTION
■
1H), 7.90−7.85 (m, 1H), 7.84−7.77 (m, 3H), 7.84−7.76 (m, 3H),
Details of materials and general methods are provided in the
Supporting Information. The analytical purity of the compounds was
assessed by HPLC analysis and determined to be ≥95%, unless
otherwise noted. The following HPLC methods were used to
determine the analytical purity of the compounds.
7
7
(
4
1
.76−7.70 (m, 1H), 7.48−7.42 (m, 2H), 7.40−7.34 (m, 11H), 7.24−
.14 (m, 4H), 6.57−6.46 (m, 1H), 5.17−5.07 (m, 4H), 4.64−4.53
m, 1H), 4.19−4.08 (m, 1H), 4.00−3.89 (m, 2H), 3.72−3.60 (m,
H), 3.55−3.45 (m, 4H), 3.24−3.13 (m, 2H), 2.95−2.75 (m, 2H),
+
.02−0.56 (m, 18H); LCMS (ES) m/z: 965.5 [M + H] .
Methyl ((5S,8S,9S,14S)-8-Benzyl-5-(tert-butyl)-15,15-dimethyl-9-
(
4
i) Analytical HPLC method A: column = Sunfire C18 [150 ×
(
(methylthio)methoxy)-3,6,13-trioxo-11-(4-(pyridin-2-yl)benzyl)-2-
.6 mm] 3.5 μm; buffer = 0.05% CF CO H in H O; mobile
3
2
2
oxa-4,7,11,12-tetraazahexadecan-14-yl)carbamate (10). Acetic
phase A = buffer/MeCN [95:5]; mobile phase B = MeCN/
buffer [95:5]; 10% B to 100% B; run time = 23 min; flow rate
anhydride (1.34 mL, 14.2 mmol) and CH CO H (2.44 mL, 42.6
3
2
mmol) were added to a stirred solution of 1 (1.0 g, 1.42 mmol) in
DMSO (2.6 mL, 36.9 mmol). The homogeneous reaction mixture
was stirred at room temperature for 2 days before being partitioned
=
(
1.0 mL/min).
ii) Analytical HPLC method B: column = XBridge phenyl
C18 [150 × 4.6 mm] 3.5 m; buffer = 0.05% CF CO H in
3
2
between EtOAc and H O. The organic layer was washed with 10%
2
H O; mobile phase A = buffer: MeCN [95:5]; mobile phase B
2
aqueous Na CO solution followed by brine solution and dried over
2
3
=
MeCN: buffer [95:5]; 10% B to 100% B; run time = 23 min;
Na SO , filtered, and concentrated under reduced pressure to afford
2
4
flow rate = 1.0 mL/min).
2
.1 g of crude compound, which was divided into two equal portions.
(
[
iii) Analytical HPLC method C: column = Kinetex Evo C18
The first half was triturated with Et O, the solid filtered, washed with
2
100 × 4.6 mm] 2.6 μm; buffer = 0.05% CF CO H in H O;
3
2
2
Et O, and dried under vacuum to afford 600 mg as an off white solid.
2
mobile phase A = buffer/MeCN [95:5]; mobile phase B =
MeCN/buffer [95:5]; 10% B to 100% B; run time = 23 min;
flow rate = 1.0 mL/min).
The second half was purified by reversed phase HPLC (column:
Sunfire C18 (150 × 19 mm) 5 μm; mobile phase A: 10 mM NH OAc
4
in H O; mobile phase B: CH CN/iPrOH (70:30); flow rate: 20 mL/
2
3
(
iv) Analytical HPLC method D: column = Kinetex Biphenyl
min; gradient: 0/40, 2/40, 17/70). The pooled HPLC fractions were
concentrated under high vacuum below 30 °C. The residue was
dissolved in a mixture of CH CN and H O; the solution was frozen
C18 [100 × 4.6 mm] 2.6 μm; buffer = 0.05% CF CO H in
H O; mobile phase A = buffer: MeCN [95:5]; mobile phase B
3
2
2
3
2
=
MeCN: buffer [95:5]; 10% B to 100% B; run time = 23 min;
and lyophilized for 16 h to afford 260 mg of an off white solid. The
flow rate = 1.0 mL/min).
combined title product was obtained as an off white solid (860 mg,
1
Experimental Procedures. Methyl ((5S,8S,9S,14S)-8-Benzyl-5-
tert-butyl)-15,15-dimethyl-3,6,13-trioxo-9-(phosphonooxy)-11-(4-
pyridin-2-yl)benzyl)-2-oxa-4,7,11,12-tetraazahexadecan-14-yl)-
1.13 mmol, 79% yield). H NMR (400 MHz, DMSO-d ): δ 8.98 (s,
6
(
(
1H), 8.71−8.59 (m, 1H), 7.97 (d, J = 8.5 Hz, 2H), 7.94−7.90 (m,
1H), 7.86 (dt, J = 2.0, 7.8 Hz, 1H), 7.60 (br d, J = 9.0 Hz, 1H), 7.41
carbamate (2). To a stirred solution of 9 (200 mg, 0.21 mmol) in
(
d, J = 8.5 Hz, 2H), 7.33 (ddd, J = 1.5, 4.8, 7.3 Hz, 1H), 7.25−7.08
(m, 5H), 6.81 (br d, J = 9.0 Hz, 1H), 6.63 (br d, J = 9.5 Hz, 1H), 5.11
br d, J = 11.5 Hz, 1H), 4.90 (d, J = 11.5 Hz, 1H), 4.31 (br s, 1H),
.01−3.83 (m, 3H), 3.78 (br s, 1H), 3.69 (br d, J = 9.5 Hz, 1H), 3.52
d, J = 6.0 Hz, 6H), 2.95−2.79 (m, 3H), 2.78−2.71 (m, 1H), 2.17 (s,
ClCH CH Cl (5 mL) were added CF CO H (2 mL, 26.0 mmol) and
2
2
3
2
anisole (0.2 mL, 1.83 mmol). The reaction mixture was stirred at
room temperature for 16 h, concentrated under high vacuum at 30 °C
to leave a colorless gum. The crude product was purified using RP-
HPLC (column: Sunfire C18 (19 × 150 mm) 5 μm; mobile phase A:
(
4
(
3
+
H), 0.76 (m, 18H); LCMS (ES) m/z: 765.4 [M + H] . Analytical
0
.1% CF CO H in H O; mobile phase B: CH CN; flow rate: 20 mL/
3 2 2 3
HPLC-RT and purity: method C = 4.64 min and 99.60% method D =
min; gradient (T/%B): 0/10, 10/50). The pooled HPLC fractions
were concentrated using high vacuum at 30 °C. The residue was
dissolved in a mixture of CH CN and H O, frozen, and lyophilized for
+
6
.13 min and 99.5%; HRMS (ESI/Orbitrap) m/z: [M + H] calcd for
C H N O S, 765.9905; found, 765.3977.
4
0
57
6
7
3
2
Disodium Methyl ((5S,8S,9S,14S)-8-Benzyl-5-(tert-butyl)-15,15-
1
2 h to afford the title compound as a white solid (120 mg, 0.150
dimethyl-3,6,13-trioxo-9-((phosphonooxy)methoxy)-11-(4-(pyri-
din-2-yl)benzyl)-2-oxa-4,7,11,12-tetraazahexadecan-14-yl)-
carbamate (11). Powdered 4 Å molecular sieves (200.0 mg) and
phosphoric acid (0.027 mL, 0.52 mmol) were added to a stirred
solution of 10 (0.08 g, 0.105 mmol) in THF (3.0 mL). The mixture
was cooled to 0 °C, NIS (0.035 g, 0.157 mmol) was added, and the
1
mmol, 72% yield). H NMR (400 MHz, DMSO-d ): δ 9.77−9.63 (m,
6
1
7
7
H), 8.81−8.64 (m, 1H), 8.11−7.99 (m, 2H), 7.95−7.86 (m, 2H),
.85−7.77 (m, 1H), 7.64−7.54 (m, 2H), 7.52−7.43 (m, 1H), 7.30−
.10 (m, 5H), 6.85−6.63 (m, 2H), 4.75−4.57 (m, 1H), 4.34−4.10
(
m, 2H), 4.03−3.89 (m, 1H), 3.73−3.59 (m, 2H), 3.57−3.53 (m,
H), 3.43−3.26 (m, 3H), 3.02−2.71 (m, 4H), 0.93−0.57 (m, 18H);
3
mixture was stirred at room temperature for 30 min. CH OH was
3
+
LCMS (ES) m/z: 785.2 [M + H] . Analytical HPLC-RT and purity:
added followed by 1.0 M Na S O solution until the reaction mixture
2
2
3
Q
DOI: 10.1021/acs.jmedchem.9b00002
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
became colorless. The pH of the mixture was adjusted to 10 with solid
mmol), DIPEA (0.109 mL, 0.622 mmol) and HATU (0.047 g,
0.124 mmol) was added to a solution of 23 (0.100 g, 0.124 mmol) in
Na CO and filtered. The filtrate was concentrated under reduced
2
3
pressure to afford a residue which was dissolved in a mixture of
DMF (10 mL), and the mixture was stirred overnight at RT. H O (1
2
CH CN and H O, frozen, and lyophilized for 16 h. The crude
× 50 mL) was added, and the mixture was extracted with EtOAc (3 ×
50 mL). The combined organic layer was washed with brine (1 × 50
mL), dried over anhydrous Na SO , filtered, and concentrated to
3
2
material was purified by using reversed phase preparative HPLC
column: X-Bridgephenyl (250 × 19 mm) 5 μm; mobile phase A:
Milli-Q H O; mobile phase B: CH CN; flow rate: 20 mL/min;
(
2
4
2
3
leave a gum (150 mg). The crude compound was purified by
preparative HPLC (column: XBridge phenyl (20 × 250 mm) 5 μm;
mobile phase A: 10 mM NH OAc in H O; mobile phase B: CH CN;
gradient: 0/10, 10/30). The pooled HPLC fractions were
concentrated under high vacuum below 30 °C. The residue was
dissolved in a mixture of CH CN and H O, frozen, and lyophilized for
4
2
3
3
2
flow rate: 17 mL/min; gradient: 30−70% mobile phase B in mobile
phase A in 10 min). The pooled HPLC fractions were concentrated
under reduced pressure below 30 °C. The solid residue was dissolved
in a mixture of CH CN and H O; the solution was frozen and
1
6 h to afford the title compound as an off white solid (42 mg, 0.048
1
mmol, 45% yield). H NMR (400 MHz, D O): δ 8.66−8.55 (m, 1H),
8
2
.23−8.16 (m, 1H), 8.06−7.97 (m, 1H), 7.83−7.74 (m, 2H), 7.70−
.51 (m, 3H), 7.31−7.14 (m, 5H), 5.26−5.03 (m, 2H), 4.61−4.43
3
2
7
lyophilized for 12 h to obtain the title product as an off-white solid
1
(
m, 1H), 4.08−3.96 (m, 1H), 3.91−3.72 (m, 2H), 3.68 (s, 3H),
.58−3.46 (m, 2H), 3.43 (s, 3H), 3.02−2.75 (m, 4H), 0.82−0.80 (m,
(60 mg, 45%). H NMR (300 MHz, DMSO-d ): δ 8.88−8.78 (m,
6
3
1
1H), 8.68−8.60 (m, 1H), 8.00−7.79 (m, 6H), 7.38−7.26 (m, 3H),
7.17 (s, 5H), 6.88−6.77 (m, 1H), 6.74−6.62 (m, 2H), 5.08−4.97 (m,
+
8H). LCMS (ES) m/z: 815.2 [M + H] . Analytical HPLC-RT and
purity: method C = 4.10 min and 99.30% method D = 5.22 min and
1
3
H), 4.72−4.61 (m, 1H), 4.55−4.45 (m, 1H), 4.05−3.79 (m, 4H),
.69−3.61 (m, 1H), 3.56 (s, 3H), 3.42 (s, 3H), 3.04−2.85 (m, 2H),
+
9
8.8%; HRMS (ESI/Orbitrap) m/z: [M + H] calcd for
C H N O P, 815.3739; found, 815.3726.
39
56
6
11
2.75−2.55 (m, 2H), 2.24−2.08 (m, 1H), 2.02−1.91 (m, 1H), 1.38 (s,
(
R)-(5S,10S,11S,14S)-11-Benzyl-5,14-di-tert-butyl-3,6,13,16-tet-
9
+
H), 0.95 (s, 6H), 0.90−0.70 (m, 24H); LCMS (ES) m/z: 1003.4 [M
raoxo-8-(4-(pyridin-2-yl)benzyl)-2,17-dioxa-4,7,8,12,15-pentaa-
zaoctadecan-10-yl 2-((tert-Butoxycarbonyl)amino)-3-methylbuta-
noate (22). (R)-2-((tert-Butoxycarbonyl)amino)-3-methylbutanoic
acid (0.055 g, 0.255 mmol), DCC (0.088 g, 0.426 mmol), and
DMAP (5.2 mg, 0.043 mmol) were added to a solution of 1 (0.150 g,
+
H] ; analytical HPLC-RT and purity: method B = 10.542 min and
9
4.69%.
(
S)-(5S,10S,11S,14S)-11-Benzyl-5,14-di-tert-butyl-3,6,13,16-tet-
raoxo-8-(4-(pyridin-2-yl)benzyl)-2,17-dioxa-4,7,8,12,15-pentaa-
zaoctadecan-10-yl 2-((S)-2-Amino-3-methylbutanamido)-3-meth-
ylbutanoate Dihydrochloride (29). A solution of 4 M HCl in Et O
0
.213 mmol) in CH Cl (10 mL), and the reaction mixture was
2 2
2
stirred at RT overnight before adding H O (1 × 50 mL). The mixture
2
(0.15 mL, 4.98 mmol) was added to 25 (0.050 g, 0.050 mmol) in a 5
mL round-bottom flask cooled to 0 °C. The mixture was allowed to
warm to RT and stirred for 1 h before being concentrated under
vacuum. The residue was dissolved in a mixture of CH CN and H O;
was extracted with CH Cl (3 × 50 mL), the combined organic layer
2
2
was dried over anhydrous Na SO , filtered, and concentrated to afford
2
4
an off-white solid (250 mg). The crude material was purified by
preparative HPLC (column: XBridge phenyl C18 (19 × 150 mm) 5
μm; mobile phase A: 10 mM NH OAc in H O; mobile phase B:
3
2
the solution was frozen and lyophilized for 12 h to obtain the title
1
4
2
product as light yellow solid (38 mg; 76%). H NMR (400 MHz,
CH CN; flow rate: 16 mL/min; gradient: 80% isocratic mobile phase
3
D O): δ 8.71 (d, J = 5.5 Hz, 1H), 8.60 (t, J = 7.8 Hz, 1H), 8.27 (d, J =
2
B in mobile phase A in 20 min). The pooled HPLC fractions were
concentrated under reduced pressure below 30 °C. The residue was
dissolved in a mixture of CH CN and H O; the solution was frozen
8
.3 Hz, 1H), 7.97 (t, J = 6.8 Hz, 1H), 7.76 (d, J = 8.3 Hz, 2H), 7.62
d, J = 8.0 Hz, 2H), 7.29−7.16 (m, 5H), 4.98 (br s, 1H), 4.35 (d, J =
6.8 Hz, 1H), 4.06 (d, J = 12.8 Hz, 1H), 3.90 (d, J = 5.8 Hz, 1H),
.87−3.77 (m, 2H), 3.64 (s, 3H), 3.52−3.43 (m, 2H), 3.33 (br s,
3H), 3.03−2.93 (m, 1H), 2.91−2.80 (m, 2H), 2.75−2.73 (m, 1H),
(
3
2
and lyophilized for 12 h to obtain the title product as an off-white
3
1
solid (45 mg, 23%). H NMR (400 MHz, DMSO-d ): δ 8.81 (br s,
6
1
H), 8.65 (td, J = 2.5, 1.0 Hz, 1H), 7.95−7.81 (m, 5H), 7.36−7.30
2
.28−2.12 (m, 2H), 1.06−0.96 (m, 12H), 0.83 (s, 9H), 0.66 (s, 9H);
(
=
m, 3H), 7.23−7.12 (m, 6H), 6.70 (dd, J = 9.0, 5.0 Hz, 2H), 4.99 (t, J
+
LCMS (ES) m/z: 903.9 [M + H] ; analytical HPLC-RT and purity:
6.5 Hz, 1H), 4.69 (br s, 1H), 4.16 (dd, J = 8.8, 5.8 Hz, 1H), 4.02−
method A = 6.351 min and 97.53% method B = 6.923 min and
3
3
1
9
.93 (m, 3H), 3.65 (d, J = 9.5 Hz, 1H), 3.56 (s, 3H), 3.42 (s, 3H),
.03−2.87 (m, 2H), 2.67 (dt, J = 4.0, 2.0 Hz, 2H), 2.19−2.09 (m,
H), 1.41 (s, 9H), 0.94 (d, J = 7.0 Hz, 6H), 0.86 (s, 9H), 0.76 (s,
7.50%; HRMS (ESI/Orbitrap) m/z: [M + H]+ calcd for
9
C H N O , 903.5339; found, 903.5330.
4
8
70
8
9
(
5S,10S,11S,14S)-11-Benzyl-5,14-di-tert-butyl-3,6,13,16-tet-
+
H); LCMS (ES): m/z: 904.4 [M + H] ; analytical HPLC-RT and
raoxo-8-(4-(pyridin-2-yl)benzyl)-2,17-dioxa-4,7,8,12,15-pentaa-
purity: method B = 9.839 min and 99.67%.
zaoctadecan-10-yl 2-((tert-Butoxycarbonyl)amino)acetate (34). 2-
(
R)-(5S,10S,11S,14S)-11-Benzyl-5,14-di-tert-butyl-3,6,13,16-tet-
((tert-Butoxycarbonyl)amino)acetic acid (0.045 g, 0.255 mmol),
raoxo-8-(4-(pyridin-2-yl)benzyl)-2,17-dioxa-4,7,8,12,15-pentaa-
zaoctadecan-10-yl 2-Amino-3-methylbutanoate Dihydrochloride
DCC (0.088 g, 0.426 mmol), and DMAP (5.20 mg, 0.043 mmol)
were added to a solution of 1 (0.150 g, 0.213 mmol) in CH Cl (20
2
2
(
24). A solution of 4 M HCl in Et O solution (0.134 mL, 4.42 mmol)
2
mL), and the was mixture stirred for 14 h at RT. H O (1 × 50 mL)
2
was added to 22 (0.040 g, 0.044 mmol) in a 5 mL round-bottom flask
cooled to 0 °C. The reaction mixture was stirred at RT for 1 h before
being concentrated under vacuum. The residue was dissolved in a
mixture of CH CN and H O; the solution was frozen and lyophilized
was added, and the mixture was extracted with CH Cl (3 × 50 mL).
2
2
The combined organic layer was dried over anhydrous Na SO ,
2
4
filtered, and concentrated to leave an off-white solid. The crude
compound was purified by preparative HPLC (column: XBridge
3
2
for 12 h to obtain the title product as a light yellow solid (35 mg;
1
phenyl (19 × 250 mm) 5 μm; mobile phase A: 10 mM NH
4
OAc in
8
4
7
8%). H NMR (400 MHz, DMSO-d ): δ 9.06 (s, 1H), 8.72 (d, J =
6
H O; mobile phase B: CH CN; flow rate: 18 mL/min; gradient: 30−
2
3
.5 Hz, 1H), 8.60 (br s, 3H), 8.15−8.03 (m, 2H), 7.99−7.92 (m, 3H),
.54 (br s, 1H), 7.43 (d, J = 8.0 Hz, 2H), 7.26−7.12 (m, 5H), 6.75 (d,
J = 10.0 Hz, 1H), 6.68 (d, J = 9.5 Hz, 1H), 5.17 (t, J = 6.5 Hz, 1H),
6
0% mobile phase B in mobile phase A in 8 min). The pooled HPLC
fractions were concentrated under reduced pressure below 30 °C; the
residual solid was dissolved in a mixture of CH CN and H O, frozen,
3
2
4
3
1
.06−3.90 (m, 5H), 3.68−3.62 (m, 1H), 3.55 (s, 3H), 3.41 (s, 3H),
and lyophilized for 12 h to obtain the title product as an off-white
.05−2.90 (m, 2H), 2.78−2.65 (m, 2H), 2.36−2.26 (m, 1H), 1.10−
1
solid (140 mg, 76%). H NMR (400 MHz, DMSO-d ) δ 8.96 (s, 1H),
6
.05 (m, 6H), 0.83 (s, 9H), 0.76 (s, 9H); LCMS (ES) m/z: 804.4 [M
+
8.69−8.58 (m, 1H), 7.99−7.78 (m, 5H), 7.41−7.28 (m, 3H), 7.27−
+
H] ; analytical HPLC-RT and purity: method A = 6.015 min and
7
.09 (m, 6H), 6.73−6.60 (m, 2H), 5.07 (t, J = 5.8 Hz, 1H), 4.52 (br s,
9
9.04% method B = 7.144 min and 97.75%; HRMS (ESI/Orbitrap)
+
1H), 4.05−3.91 (m, 3H), 3.88 (d, J = 6.0 Hz, 1H), 3.84−3.71 (m,
m/z: [M + H] calcd for C H N O , 804.4654; found, 804.4661.
4
3
61
7
8
1
1
H), 3.66 (d, J = 9.5 Hz, 1H), 3.55 (s, 3H), 3.43 (s, 3H), 3.07 (dd, J =
3.3, 5.8 Hz, 1H), 2.91 (dd, J = 13.1, 7.5 Hz, 1H), 2.76−2.57 (m,
(
S)-(5S,10S,11S,14S)-11-Benzyl-5,14-di-tert-butyl-3,6,13,16-tet-
raoxo-8-(4-(pyridin-2-yl)benzyl)-2,17-dioxa-4,7,8,12,15-pentaa-
zaoctadecan-10-yl 2-((S)-2-((tert-Butoxycarbonyl)amino)-3-meth-
ylbutanamido)-3-methylbutanoate (25). (S)-2-((tert-
Butoxycarbonyl)amino)-3-methylbutanoic acid (0.041 g, 0.187
2H), 1.41 (s, 9H), 0.81 (s, 9H), 0.74 (s, 9H); LCMS (ES) m/z: 861.4
+
[M + H] ; analytical HPLC-RT and purity: method B = 9.011 min
and 99.83%.
R
DOI: 10.1021/acs.jmedchem.9b00002
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
(
5S,10S,11S,14S)-11-Benzyl-5,14-di-tert-butyl-3,6,13,16-tet-
A solution of 4 M HCl in dioxane (1.0 mL, 32.9 mmol) was added to
35 (0.2 g, 0.228 mmol) in a 5 mL round-bottom flask maintained at 0
raoxo-8-(4-(pyridin-2-yl)benzyl)-2,17-dioxa-4,7,8,12,15-pentaa-
zaoctadecan-10-yl 2-((tert-Butoxycarbonyl) (methyl)amino)-
acetate (35). 2-((tert-Butoxycarbonyl)(methyl)amino)acetic acid
°
C. The reaction mixture was warmed to RT and stirred for 2 h before
being concentrated under vacuum. The residue was triturated with
(
(
268 mg, 1.419 mmol), DMAP (87 mg, 0.709 mmol), and DCC
293 mg, 1.419 mmol) were added to a solution of 1 (500 mg, 0.709
mmol) in CH Cl (10 mL), and the mixture was stirred overnight at
RT. The mixture was concentrated under vacuum, and the residue
was purified by preparative HPLC. The pooled HPLC fractions were
concentrated under reduced pressure below 30 °C, the residual solid
was dissolved in a mixture of CH CN and H O, frozen, and
Et O and the ethereal layer decanted. The residual solid was dissolved
2
in a mixture of CH CN and H O, frozen, and lyophilized for 12 h to
3
2
1
2
2
obtain the title product as a white solid (38 mg; 19%). H NMR (400
MHz, DMSO) δ = 9.45−9.25 (m, 2H), 9.18−9.06 (m, 1H), 8.81−
8
5
.62 (m, 1H), 7.98−7.96 (m, 5H), 7.54−7.39 (m, 3H), 7.21 (br s,
H), 6.84−6.61 (m, 2H), 5.25−5.11 (m, 1H), 4.63−4.50 (m, 1H),
3
2
4.13−3.89 (m, 5H), 3.68−3.62 (m, 1H), 3.55−3.50 (m, 3H), 3.42 (s,
lyophilized for 12 h to obtain the title product as a white solid (400
3
H), 3.08−2.89 (m, 2H), 2.69 (s, 5H), 0.85 (s, 9H), 0.75 (s, 9H);
1
mg; 64%). H NMR (400 MHz, MeOD): δ 8.64−8.57 (m, 1H),
+
LCMS (ES) m/z: 776.4 [M + H] ; analytical HPLC-RT and purity:
7
7
5
2
.97−7.78 (m, 5H), 7.50 (d, J = 8.5 Hz, 2H), 7.41−7.33 (m, 1H),
.24 (s, 5H), 5.31−5.17 (m, 1H), 4.69−4.50 (m, 1H), 4.32−3.91 (m,
H), 3.77−3.49 (m, 7H), 3.12−2.99 (m, 5H), 2.97−2.84 (m, 1H),
method A = 5.886 min and 98.13% method B = 6.902 min and
+
9
8.94%. HRMS (ESI/Orbitrap) m/z: [M + H] calcd for
C H N O , 776.4341; found, 776.4343.
4
1
57
7
8
.79−2.62 (m, 1H), 1.50 (d, J = 19.6 Hz, 9H), 0.96−0.76 (m, 18H);
(
S)-(5S,10S,11S,14S)-11-Benzyl-5,14-di-tert-butyl-3,6,13,16-tet-
+
LCMS (ES) m/z: 877.4 [M + H] ; analytical HPLC-RT and purity:
raoxo-8-(4-(pyridin-2-yl)benzyl)-2,17-dioxa-4,7,8,12,15-pentaa-
zaoctadecan-10-yl 2-Amino-3-phenylpropanoate, Dihydrochloride
(53). A solution of 4 M HCl in dioxane (4.47 mL, 147 mmol) was
added to 42 (2.800 g, 2.94 mmol) in a round-bottom flask maintained
at 0 °C. The reaction mixture was warmed to RT and stirred for 1 h
before being concentrated under vacuum. The residue was dissolved
in a mixture of CH CN and H O, and the solution was frozen and
method B = 9.689 min and 99.53%.
(
raoxo-8-(4-(pyridin-2-yl)benzyl)-2,17-dioxa-4,7,8,12,15-pentaa-
zaoctadecan-10-yl 2-((tert-Butoxycarbonyl)amino)-3-phenylpro-
panoate (42). (S)-2-((tert-Butoxycarbonyl)amino)-3-phenylpropa-
noic acid (0.059 g, 0.221 mmol), DCC (0.076 g, 0.369 mmol), and
DMAP (4.51 mg, 0.037 mmol) were added to a solution of 1 (0.130
g, 0.184 mmol) in CH Cl (15 mL), and the mixture was stirred at
RT for 16 h. H O was added (1 × 50 mL), the mixture was extracted
with EtOAc (3 × 50 mL), and the combined organic layer was
washed with brine (1 × 50 mL), dried over anhydrous Na SO ,
filtered, and concentrated to leave an off-white solid. The crude
compound was purified by preparative HPLC (column: Sunfire C18
S)-(5S,10S,11S,14S)-11-Benzyl-5,14-di-tert-butyl-3,6,13,16-tet-
3
2
lyophilized for 12 h to obtain the title product as an off-white solid
1
2
2
(2.50 g; 85%). H NMR (400 MHz, DMSO-d ): δ 9.20 (s, 1H), 8.82
6
2
(br s, 3H), 8.74 (d, J = 4.5 Hz, 1H), 8.17 (br s, 1H), 8.14−8.07 (m,
1H), 7.97 (d, J = 8.5 Hz, 2H), 7.91 (d, J = 9.5 Hz, 1H), 7.60 (br s,
2
4
1H), 7.47 (d, J = 8.0 Hz, 2H), 7.40−7.32 (m, 4H), 7.32−7.25 (m,
1
H), 7.24−7.12 (m, 5H), 6.82 (d, J = 9.5 Hz, 1H), 6.72 (d, J = 10.0
Hz, 1H), 5.07 (t, J = 6.5 Hz, 1H), 4.74−4.65 (m, 1H), 4.38 (d, J = 5.5
Hz, 1H), 4.02−3.92 (m, 3H), 3.65 (d, J = 9.5 Hz, 1H), 3.53 (s, 3H),
3.41−3.32 (m, 4H), 3.29−3.21 (m, 1H), 3.01−2.88 (m, 2H), 2.67 (d,
(
19 × 150 mm) 5 μm; mobile phase A: 10 mM NH OAc in H O;
4 2
mobile phase B: CH CN; flow rate: 18 mL/min; gradient: 50−70%
3
mobile phase B in mobile phase A in 8 min). The pooled HPLC
J = 7.0 Hz, 2H), 0.83 (s, 9H), 0.74 (s, 9H).; LCMS (ES) m/z: 852.4
fractions were concentrated under reduced pressure below 30 °C, the
residual solid was dissolved in a mixture of CH CN and H O; the
solution was frozen and lyophilized for 12 h to obtain the title product
as an off-white solid (130 mg, 74%). H NMR (400 MHz, DMSO-
+
[M + H] ; analytical HPLC-RT and purity: method A = 6.879 min
3
2
and 99.67% method B = 8.304 min and 99.65%. HRMS (ESI/
+
Orbitrap) m/z: [M + H] calcd for C H N O , 852.4654; found,
4
7
61
7
8
1
8
52.4660.
d6): δ 8.89 (s, 1H), 8.65 (dd, J = 6.5, 1.0 Hz, 1H), 7.93 (d, J = 8.0 Hz,
(5S,10S,11S,14S)-11-Benzyl-5,14-di-tert-butyl-3,6,13,16-tet-
2
H), 7.91−7.82 (m, 2H), 7.70 (d, J = 10.0 Hz, 1H), 7.38 (d, J = 8.0
raoxo-8-(4-(pyridin-2-yl)benzyl)-2,17-dioxa-4,7,8,12,15-pentaa-
z a o c t a d e c a n - 1 0 - y l 2 - ( ( S ) - 2 - A m i n o - N - m e t h y l - 3 -
phenylpropanamido)acetate, Dihydrochloride (74). A solution of
4 M HCl in dioxane (1 mL, 32.9 mmol) was added to 79 (65 mg,
0.064 mmol) in a 5 mL flask maintained at 0 °C. The reaction mixture
was warmed to RT and stirred for 1 h before being concentrated
Hz, 2H), 7.35−7.26 (m, 6H), 7.25−7.12 (m, 6H), 6.72 (d, J = 10.0
Hz, 1H), 6.66 (d, J = 9.0 Hz, 1H), 5.02 (br s, 1H), 4.61 (br s, 1H),
4
3
1
9
.35 (br s, 1H), 4.00 (t, J = 11.0 Hz, 3H), 3.67 (d, J = 9.0 Hz, 1H),
.54 (s, 3H), 3.43 (s, 3H), 3.26 (d, J = 4.5 Hz, 1H), 3.12−3.03 (m,
H), 3.01−2.85 (m, 2H), 2.73−2.58 (m, 2H), 1.32 (s, 9H), 0.87 (s,
+
H), 0.74 (s, 9H).; LCMS (ES) m/z: 952.9 [M + H] . Analytical
under vacuum. The residue was triturated with Et
layer was decanted, and the residual solid was dried under vacuum.
The solid was dissolved in a mixture of CH CN and H O, frozen, and
O, the ethereal
2
HPLC-RT and purity: method A = 17.030 min and 99.76% method B
18.399 min and 99.83%.
5S,10S,11S,14S)-11-Benzyl-5,14-di-tert-butyl-3,6,13,16-tet-
=
3
2
(
lyophilized for 12 h to obtain the title product as white solid (63.95
1
raoxo-8-(4-(pyridin-2-yl)benzyl)-2,17-dioxa-4,7,8,12,15-pentaa-
zaoctadecan-10-yl 2-Aminoacetate Dihydrochloride (44). A
solution of 4 M HCl in dioxane (0.441 mL, 14.50 mmol) was
added to 34 (0.125 g, 0.145 mmol) in a 5 mL round-bottom flask
maintained at 0 °C. The mixture was stirred for 1 h at RT before
being concentrated under vacuum. The residue was dissolved in a
mixture of CH CN and H O; the solution was frozen and lyophilized
mg; 98%). H NMR (400 MHz, CH OH-d ): δ 8.86−8.79 (m, 1H),
3
6
8.68−8.56 (m, 1H), 8.38−8.28 (m, 1H), 8.21−7.88 (m, 3H), 7.75−
7.65 (m, 2H), 7.42−7.30 (m, 5H), 7.20−7.10 (m, 5H), 5.29−5.17
(m, 1H), 4.81−4.71 (m, 2H), 4.52−4.36 (m, 1H), 4.28−4.01 (m,
3H), 3.94 (s, 1H), 3.81−3.57 (m, 5H), 3.52 (s, 2H), 3.28−3.12 (m,
2H), 3.11−2.96 (m, 5H), 2.92−2.64 (m, 2H), 0.91−0.73 (m, 18H);
+
3
2
LCMS (ES) m/z: 923.4 [M + H] ; HPLC-RT and purity: (a) method
for 12 h to obtain the title product as a light yellow solid (110 mg;
9%). H NMR (400 MHz, DMSO-d ): δ 9.19 (s, 1H), 8.77 (d, J =
A = 7.199 min and 96.7%; (b) method B = 7.519 min and 96.9%.
1
+
8
4
1
7
1
3
3
HRMS (ESI/Orbitrap) m/z: [M + H] calcd for C H N O ,
6
50 66
8
9
.5 Hz, 1H), 8.59 (br s, 3H), 8.25 (br s, 1H), 8.16 (d, J = 8.0 Hz,
H), 8.02−7.93 (m, 3H), 7.66 (br s, 1H), 7.48 (d, J = 8.0 Hz, 2H),
.28−7.11 (m, 5H), 6.71 (d, J = 9.5 Hz, 1H), 6.75 (d, J = 9.5 Hz,
H), 5.16 (t, J = 6.5 Hz, 1H), 4.65 (br s, 1H), 4.06−3.88 (m, 2H),
.75−3.62 (m, 2H), 3.56 (s, 3H), 3.53−3.44 (m, 2H), 3.40 (s, 3H),
923.5026; found, 923.5032.
(5S,10S,11S,14S)-11-Benzyl-5,14-di-tert-butyl-3,6,13,16-tet-
raoxo-8-(4-(pyridin-2-yl)benzyl)-2,17-dioxa-4,7,8,12,15-pentaa-
zaoctadecan-10-yl 2-((S)-2-((tert-Butoxycarbonyl)amino)-N-meth-
yl-3-phenylpropanamido)acetate (79). To a stirred solution of 45
(
0.2 g, 0.236 mmol) in DMF (8 mL), DIPEA (0.206 mL, 1.178
.07−2.90 (m, 2H), 2.79−2.64 (m, 2H), 0.82 (s, 9H), 0.75 (s, 9H).;
LCMS (ES) m/z: 762.4 [M + H] ; analytical HPLC-RT and purity:
+
mmol) was added followed by HATU (0.134 g, 0.353 mmol) and
Boc-Phe-OH (0.125 g, 0.471 mmol). The reaction mixture was stirred
method A = 5.954 min and 99.11% method B = 6.730 min and
at RT for 8 h before being partitioned between H O and EtOAc. The
9
9.08%. HRMS (ESI/Orbitrap) m/z: [M + H]+ calcd for
2
organic layer was separated and washed with brine solution (1 × 25
C H N O , 762.4185; found, 762.4193.
4
0
(
55
7
8
mL), dried over Na SO , and concentrated under vacuum at 50 °C to
5S,10S,11S,14S)-11-Benzyl-5,14-di-tert-butyl-3,6,13,16-tet-
2
4
raoxo-8-(4-(pyridin-2-yl)benzyl)-2,17-dioxa-4,7,8,12,15-pentaa-
leave the crude product which was purified by preparative HPLC. The
zaoctadecan-10-yl 2-(Methylamino)acetate Dihydrochloride (45).
pooled HPLC fractions were concentrated under reduced pressure
S
DOI: 10.1021/acs.jmedchem.9b00002
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
below 30 °C. The residual solid was dissolved in a mixture of CH CN
simulated intestinal fluid; HBA, hydrogen bond acceptor;
HBD, hydrogen bond donor; HPßCD, hydroxypropyl ß-
cyclodextrin; LPV, lopinavir; MRP, multi-drug resistant
protein; NNRTI, non-nucleoside reverse transcriptase inhib-
itor; NRTI, nucleoside reverse transcriptase inhibitor; OATP,
organic anion transporting polypeptide; OCT, organic cation
transporter; PepT1, peptidase transporter 1; P-gp, P-
glycoprotein; PI, protease inhibitor; PMSF, phenylmethylsul-
fonyl fluoride; POM, phosphonooxymethyl; PPI, proton pump
inhibitor; PSA, polar surface area; QD, quaque die; rIALP,
recombinant intestinal alkaline phosphatase; RLM, rat liver
microsome; RTV, ritonavir; UGT1A1, uridine diphosphate
glucuronosyltransferase 1A1; WHO, World Health Organ-
ization
3
and H O, frozen, and lyophilized for 12 h to obtain the title product
2
1
as a white solid (0.2 g, 81%). H NMR (400 MHz, DMSO): δ 9.06−
8
(
1
4
3
(
9
.94 (m, 1H), 8.68−8.60 (m, 1H), 8.01−7.79 (m, 5H), 7.42−7.30
m, 4H), 7.30−7.22 (m, 3H), 7.20−7.13 (m, 6H), 7.09−6.97 (m,
H), 6.73−6.62 (m, 2H), 5.28−5.05 (m, 1H), 4.89−4.64 (m, 1H),
.56−4.40 (m, 2H), 4.32−4.08 (m, 2H), 4.01−3.86 (m, 3H), 3.70−
.52 (m, 3H), 3.45−3.40 (m, 4H), 3.13−3.05 (m, 2H), 3.00−2.96
m, 1H), 2.91−2.70 (m, 4H), 2.68−2.60 (m, 1H), 1.33−1.22 (m,
+
H), 0.84−0.68 (m, 18H) m/z: 1023.7 [M + H] ; HPLC-RT: (a)
method B = 10.351 min.
ASSOCIATED CONTENT
Supporting Information
■
*
S
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.jmed-
chem.9b00002.
REFERENCES
■
Molecular formula strings (CSV)
(
1) Cihlar, T.; Fordyce, M. Current status and prospects of HIV
treatment. Curr. Opin. Virol. 2016, 18, 50−56.
Figures 3−11 and Table 13; experimental section
(
(
7
synthesis) = experimental procedures of prodrugs
23, 30, 31, 32, 46, 47, 48, 49, 51, 52, 54, 59, 64, 66,
1, 78, 81, and 82); details of in vitro and in vivo
(2) Samji, H.; Cescon, A.; Hogg, R. S.; Modur, S. P.; Althoff, K. N.;
Buchacz, K.; Burchell, A. N.; Cohen, M.; Gebo, K. A.; Gill, M. J.;
Justice, A.; Kirk, G.; Klein, M. B.; Korthuis, P. T.; Martin, J.;
Napravnik, S.; Rourke, S. B.; Sterling, T. R.; Silverberg, M. J.; Deeks,
S.; Jacobson, L. P.; Bosch, R. J.; Kitahata, M. M.; Goedert, J. J.;
Moore, R.; Gange, S. J. The North American AIDS Collaboration on
Research and Design (NA-ACCORD) of leDEA. Closing the Gap:
Increases in Life Expectancy Among Treated HIV-Positive Individuals
in the United States and Canada. PLoS One 2013, 8, No. e81355.
(3) Zhan, P.; Pannecouque, C.; De Clercq, E.; Liu, X. Anti-HIV
Drug Discovery and Development: Current Innovations and Future
Trends. J. Med. Chem. 2016, 59, 2849−2878.
experiments = stability evaluation in aqueous buffer;
stability evaluation in bio-relevant media; solubility
evaluation in aqueous buffer; pharmacokinetic evalua-
tion; ex vivo stability in whole blood; alkaline
phosphatase assay: determination of prodrug cleavage
rate in vitro; rat hepatocyte study with and without
PMSF/ABT; procedure for UGT1A1 assay; antiviral
assays (PDF)
(4) Ghosh, A. K.; Osswald, H. L.; Prato, G. Recent Progress in the
Development of HIV-1 Protease Inhibitors for the Treatment of
HIV/AIDS. J. Med. Chem. 2016, 59, 5172−5208.
AUTHOR INFORMATION
Corresponding Author
E-mail: murugaiah.andappan@syngeneintl.com. Phone: +91-
731600213.
■
(5) Midde, N. M.; Patters, B. J.; Rao, P.; Cory, T. J.; Kumar, S.
Investigational Protease Inhibitors as Antiretroviral Therapies. Expert
Opin. Invest. Drugs 2016, 25, 1189−1200.
*
9
(6) Farajallah, A.; Bunch, R. T.; Meanwell, N. A. Discovery and
ORCID
Development of Atazanavir. Antiviral Drugs; John Wiley & Sons, Inc.:
New Jersey, 2011; pp 1−17.
̈
(7) Gunthard, H. F.; Saag, M. S.; Benson, C. A.; del Rio, C.; Eron, J.
Nicholas A. Meanwell: 0000-0002-8857-1515
Present Addresses
ViiV Healthcare, 36 East Industrial Road, Branford, CT
6405, USA.
Formulation Division, R&D, Dr. Reddy’s Laboratories Ltd.,
Bachupally, Qutubullapur, Hyderabad, Telangana, 500090,
India.
¶
J.; Gallant, J. E.; Hoy, J. F.; Mugavero, M. J.; Sax, P. E.; Thompson, M.
A.; Gandhi, R. T.; Landovitz, R. J.; Smith, D. M.; Jacobsen, D. M.;
Volberding, P. A. Antiretroviral Drugs for Treatment and Prevention
of HIV Infection in Adults: 2016 Recommendations of the
International Antiviral Society-USA Panel. JAMA 2016, 316, 191−
0
∇
2
(
10.
Notes
8) Appendix B, Table 3. Characteristics of Protease Inhibitors.
The authors declare no competing financial interest.
https://aidsinfo.nih.gov/guidelines/htmltables/1/5482 (accessed
March 23, 2018).
ACKNOWLEDGMENTS
(9) Kis, O.; Zastre, J. A.; Hoque, M. T.; Walmsley, S. L.; Bendayan,
R. Role of Drug Efflux and Uptake Transporters in Atazanavir
Intestinal Permeability and Drug-Drug Interactions. Pharm. Res. 2013,
■
We thank Drs. Ramakanth Sarabu, Arvind Mathur, and Percy
Carter for their support and inspiration. We thank Drs. Punit
Marathe, Michael Hageman, and Bruce Car for productive
discussions. We acknowledge the Discovery Analytical Sciences
team for the analytical support.
3
0, 1050−1064.
10) Zhu, L.; Persson, A.; Mahnke, L.; Eley, T.; Li, T.; Xu, X.;
Agarwala, S.; Dragone, J.; Bertz, R. Effect of Low-Dose Omeprazole
20 mg daily) on the Pharmacokinetics of Multiple-Dose Atazanavir
(
(
with Ritonavir in Healthy Subjects. J. Clin. Pharmacol. 2011, 51, 368−
ABBREVIATIONS
377.
■
(11) Zhang, L.; Wu, F.; Lee, S. C.; Zhao, H.; Zhang, L. pH-
ABT, 1-aminobenzotriazole; ADME, absorption distribution
metabolism and excretion; ALP, alkaline phosphatase; APV,
amprenavir; ATV, Atazanavir; BCRP, breast cancer resistance
protein; BID, bis in die; BLQ, below limit of quantification;
cART, combination anti-retroviral therapy; CYP, cytochrome
P450; DMAc, dimethylacetamide; N,N-dimethyl-Gly, N,N-
dimethylglycine; FaSSIF, fasted state simulated intestinal fluid;
FDA, Federal Drug Administration; FeSSIF, fed state
Dependent Drug-Drug Interactions for Weak Base Drugs: Potential
Implications for New Drug Development. Clin. Pharmacol. Ther.
2
(
014, 96, 266−277.
12) Klein, C. E.; Chiu, Y.-L.; Cai, Y.; Beck, K.; King, K. R.;
Causemaker, S. J.; Doan, T.; Esslinger, H.-U.; Podsadecki, T. J.;
Hanna, G. J. Effects of Acid-Reducing Agents on the Pharmacoki-
netics of Lopinavir/Ritonavir and Ritonavir-Boosted Atazanavir. J.
Clin. Pharmacol. 2008, 48, 553−562.
T
DOI: 10.1021/acs.jmedchem.9b00002
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
(
13) Bertz, R. J.; Persson, A.; Chung, E.; Zhu, L.; Zhang, J.;
S.; Krishnamurthy, P.; Ramarao, M. Crystal Structure of Rat Intestinal
Alkaline Phosphatase - Role of Crown Domain in Mammalian
Alkaline Phosphatases. J. Struct. Biol. 2013, 184, 182−192.
(31) Subramanian, M.; Paruchury, S.; Pratap Singh, S.; Arla, R.;
Pahwa, S.; Jana, S.; Katnapally, P.; Yoganand, V.; Lakshmaiah, B.;
Mazumder Tagore, D.; Ghosh, K.; Marathe, P.; Mandlekar, S.
Characterization of Recombinantly Expressed Rat and Monkey
Intestinal Alkaline Phosphatases: In Vitro Studies and In Vivo
Correlations. Drug Metab. Dispos. 2013, 41, 1425−1432.
McGrath, D.; Grasela, D. Pharmacokinetics and Pharmacodynamics
of Atazanavir-Containing Antiretroviral Regimens, With or Without
Ritonavir, in Patients Who Are HIV-Positive and Treatment-Naive.
Pharmacotherapy 2013, 33, 284−294.
(
14) Furfine, E. S.; Baker, C. T.; Hale, M. R.; Reynolds, D. J.;
Salisbury, J. A.; Searle, A. D.; Studenberg, S. D.; Todd, D.; Tung, R.
D.; Spaltenstein, A. Preclinical Pharmacology and Pharmacokinetics
of GW433908, a Water-Soluble Prodrug of the Human Immunode-
ficiency Virus Protease Inhibitor Amprenavir. Antimicrob. Agents
Chemother. 2004, 48, 791−798.
(32) Yang, Y.-h.; Aloysius, H.; Inoyama, D.; Chen, Y.; Hu, L.-q.
Enzyme-Mediated Hydrolytic Activation of Prodrugs. Acta Pharm.
Sin. B 2011, 1, 143−159.
(
15) DeGoey, D. A.; Grampovnik, D. J.; Flosi, W. J.; Marsh, K. C.;
Wang, X. C.; Klein, L. L.; McDaniel, K. F.; Liu, Y.; Long, M. A.; Kati,
W. M.; Molla, A.; Kempf, D. J. Water-Soluble Prodrugs of the Human
Immunodeficiency Virus Protease Inhibitors Lopinavir and Ritonavir.
J. Med. Chem. 2009, 52, 2964−2970.
(33) Riggs-Sauthier, J. Water Soluble Non-Peptidic Oligomer-
Lipophilic Moieties-Protease Inhibitor Conjugates, Especially Con-
taining Atazanavir and Tipranavir as the Protease Inhibitors, Their
Preparation, Pharmacokinetics and Anti-HIV-1 Agents. WO
(
16) Subbaiah, M. A. M.; Meanwell, N. A.; Kadow, J. F.; Subramani,
2
(
010144869 A2, 2010.
34) Dhareshwar, S. S.; Stella, V. J. Your Prodrug Releases
Formaldehyde: Should You be Concerned? No! J. Pharm. Sci. 2008,
7, 4184−4193.
35) Yamada, K.; Kato, K.; Nagase, H.; Hirata, Y. Protection of
L.; Annadurai, M.; Ramar, T.; Desai, S. D.; Sinha, S.; Subramanian,
M.; Mandlekar, S.; Sridhar, S.; Padmanabhan, S.; Bhutani, P.; Arla, R.;
Jenkins, S. M.; Krystal, M. R.; Wang, C.; Sarabu, R. Coupling of an
Acyl Migration Prodrug Strategy with Bio-activation To Improve Oral
Delivery of the HIV-1 Protease Inhibitor Atazanavir. J. Med. Chem.
9
(
Tertiary Hydroxyl Groups as Methylthiomethyl Ethers. Tetrahedron
Lett. 1976, 17, 65−66.
2
(
018, 61, 4176−4188.
17) FDA. Website: Atazanavir Documents. https://www.
(36) Huttunen, K. M.; Rautio, J. Prodrugs - an Efficient Way to
accessdata.fda.gov/drugsatfda_docs/nda/2003/21-567_Reyataz_
BioPharmr_P1.pdf (accessed June 09, 2018).
(
Breach Delivery and Targeting Barriers. Curr. Top. Med. Chem. 2011,
1, 2265−2287.
37) Larson, K. B.; Wang, K.; Delille, C.; Otofokun, I.; Acosta, E. P.
1
(
18) Patel, M.; Mandava, N.; Gokulgandhi, M.; Pal, D.; Mitra, A.
Amino Acid Prodrugs: An Approach to Improve the Absorption of
Pharmacokinetic Enhancers in HIV Therapeutics. Clin. Pharmacoki-
net. 2014, 53, 865−872.
HIV-1 Protease Inhibitor, Lopinavir. Pharmaceuticals 2014, 7, 433−
4
52.
(38) Vig, B. S.; Huttunen, K. M.; Laine, K.; Rautio, J. Amino Acids
(
19) Jain, R.; Agarwal, S.; Mandava, N. K.; Sheng, Y.; Mitra, A. K.
as Promoieties in Prodrug Design and Development. Adv. Drug
Delivery Rev. 2013, 65, 1370−1385.
Interaction of Dipeptide Prodrugs of Saquinavir with Multidrug
Resistance Protein-2 (MRP-2): Evasion of MRP-2 Mediated Efflux.
Int. J. Pharm. 2008, 362, 44−51.
(
39) Krecmerova, M. Amino Acid Ester Prodrugs of Nucleoside and
̌ ́
Nucleotide Antivirals. Mini-Rev. Med. Chem. 2017, 17, 818−833.
(
20) Jain, R.; Duvvuri, S.; Kansara, V.; Mandava, N. K.; Mitra, A. K.
(40) Murakami, T. A Mini-Review: Usefulness of Transporter-
Intestinal Absorption of Novel-Dipeptide Prodrugs of Saquinavir in
Rats. Int. J. Pharm. 2007, 336, 233−240.
(
Hayashi, Y.; Kiso, Y. Water-Soluble Prodrugs of Dipeptide HIV
Protease Inhibitors Based on O→N Intramolecular Acyl Migration:
Design, Synthesis and Kinetic Study. Bioorg. Med. Chem. 2004, 12,
Targeted Prodrugs in Enhancing Membrane Permeability. J. Pharm.
Sci. 2016, 105, 2515−2526.
21) Hamada, Y.; Matsumoto, H.; Yamaguchi, S.; Kimura, T.;
(41) Roche, D.; Greiner, J.; Aubertin, A.-M.; Vierling, P. Synthesis
and in vitro Biological Evaluation of Valine-Containing Prodrugs
Derived From Clinically Used HIV-Protease Inhibitors. Eur. J. Med.
Chem. 2008, 43, 1506−1518.
1
59−170.
22) Corbett, A. H.; Kashuba, A. D. Fosamprenavir. Vertex
Pharmaceuticals/GlaxoSmithKline. Curr. Opin. Investig. Drugs 2002,
, 384−390.
23) Subbaiah, M. A. M.; Meanwell, N. A.; Kadow, J. F. Design
(42) Wang, Z.; Pal, D.; Mitra, A. K. Stereoselective Evasion of P-
(
Glycoprotein, Cytochrome P450 3A, and Hydrolases by Peptide
Prodrug Modification of Saquinavir. J. Pharm. Sci. 2012, 101, 3199−
3
(
3
(
213.
43) de Oliveira, M. P.; Olivier, J.-C.; Pariat, C.; Roche, D.; Greiner,
Strategies in the Prodrugs of HIV-1 Protease Inhibitors to Improve
J.; Vierling, P.; Couet, W. Investigation of Oral Bioavailability and
Brain Distribution of the Ind(8)-Val Conjugate of Indinavir in
Rodents. J. Pharm. Pharmacol. 2005, 57, 453−458.
the Pharmaceutical Properties. Eur. J. Med. Chem. 2017, 139, 865−
8
83.
(
24) Harbeson, S. L.; Tung, R. D. Azapeptide Derivatives. U.S.
Patent 8,158,805 B2, Apr 17, 2012.
25) Concert Pharmaceuticals, Inc.’s Patent Owner Preliminary
Response. https://www.concertpharma.com/wp-content/uploads/
017/07/Concert-IPR2017-01256-Patent-Owners-Preliminary-
Response.pdf (accessed Oct 6, 2018).
26) Clas, S.-D.; Sanchez, R. I.; Nofsinger, R. Chemistry-enabled
(44) Agarwal, S.; Boddu, S. H. S.; Jain, R.; Samanta, S.; Pal, D.;
Mitra, A. K. Peptide Prodrugs: Improved Oral Absorption of
(
Lopinavir, a HIV Protease Inhibitor. Int. J. Pharm. 2008, 359, 7−14.
(45) Lai, L.; Xu, Z.; Zhou, J.; Lee, K.-D.; Amidon, G. L. Molecular
2
Basis of Prodrug Activation by Human Valacyclovirase, an α-Amino
Acid Ester Hydrolase. J. Biol. Chem. 2008, 283, 9318−9327.
(
(46) Gaucher, B.; Rouquayrol, M.; Roche, D.; Greiner, J.; Aubertin,
drug delivery (prodrugs): recent progress and challenges. Drug
A.-M.; Vierling, P. Prodrugs of HIV Protease Inhibitors-Saquinavir,
Indinavir and Nelfinavir-Derived from Diglycerides or Amino Acids:
Synthesis, Stability and Anti-HIV Activity. Org. Biomol. Chem. 2004,
2, 345−357.
(47) Rouquayrol, M.; Gaucher, B.; Roche, D.; Greiner, J.; Vierling,
P. Transepithelial Transport of Prodrugs of the HIV Protease
Inhibitors Saquinavir, Indinavir, and Nelfinavir Across Caco-2 Cell
Monolayers. Pharm. Res. 2002, 19, 1704−1712.
(48) Usansky, H. H.; Hu, P.; Sinko, P. J. Differential Roles of P-
Glycoprotein, Multidrug Resistance-Associated Protein 2, and CYP3A
on Saquinavir Oral Absorption in Sprague-Dawley Rats. Drug Metab.
Dispos. 2008, 36, 863−869.
Discovery Today 2014, 19, 79−87.
(
27) Rautio, J.; Karkkainen, J.; Sloan, K. B. Prodrugs - Recent
̈ ̈
Approvals and a Glimpse of the Pipeline. Eur. J. Pharm. Sci. 2017, 109,
46−161.
28) Walther, R.; Rautio, J.; Zelikin, A. N. Prodrugs in Medicinal
Chemistry and Enzyme Prodrug Therapies. Adv. Drug Delivery Rev.
017, 118, 65−77.
29) Rautio, J.; Meanwell, N. A.; Di, L.; Hageman, M. J. The
1
(
2
(
Expanding Role of Prodrugs in Contemporary Drug Design and
Development. Nat. Rev. Drug Discovery 2018, 17, 559−587.
(
30) Ghosh, K.; Mazumder Tagore, D.; Anumula, R.; Lakshmaiah,
B.; Kumar, P. P. B. S.; Singaram, S.; Matan, T.; Kallipatti, S.; Selvam,
U
DOI: 10.1021/acs.jmedchem.9b00002
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
(
49) Gong, J.; Gan, J.; Caceres-Cortes, J.; Christopher, L. J.; Arora,
Reactivity and Toxicity of Dipeptide Esters of Paracetamol. Bioorg.
Med. Chem. Lett. 2005, 15, 1595−1598.
V.; Masson, E.; Williams, D.; Pursley, J.; Allentoff, A.; Lago, M.; Tran,
S. B.; Iyer, R. A. Metabolism and Disposition of [14C]Brivanib
Alaninate after Oral Administration to Rats, Monkeys, and Humans.
Drug Metab. Dispos. 2011, 39, 891−903.
(66) Meibom, D.; Albrecht-Ku
Kast, R.; Kram
P. G.; Sussmeier, F.; Vakalopoulos, A.; Zimmermann, K. Neladenoson
pper, B.; Diedrichs, N.; Hubsch, W.;
̈ ̈
̈
er, T.; Krenz, U.; Lerchen, H.-G.; Mittendorf, J.; Nell,
̈
(
50) Cai, Z.-w.; Zhang, Y.; Borzilleri, R. M.; Qian, L.; Barbosa, S.;
Bialanate Hydrochloride: A Prodrug of a Partial Adenosine A1
Receptor Agonist for the Chronic Treatment of Heart Diseases.
ChemMedChem 2017, 12, 728−737.
Wei, D.; Zheng, X.; Wu, L.; Fan, J.; Shi, Z.; Wautlet, B. S.; Mortillo,
S.; Jeyaseelan, R.; Kukral, D. W.; Kamath, A.; Marathe, P.; D’Arienzo,
C.; Derbin, G.; Barrish, J. C.; Robl, J. A.; Hunt, J. T.; Lombardo, L. J.;
Fargnoli, J.; Bhide, R. S. Discovery of Brivanib Alaninate ((S)-((R)-1-
4-(4-Fluoro-2-methyl-1H-indol-5-yloxy)-5-methylpyrrolo[2,1-f ]-
1,2,4]triazin-6-yloxy)propan-2-yl)2-aminopropanoate), A Novel Pro-
(67) Santos, C. R.; Capela, R.; Pereira, C. S. G. P.; Valente, E.;
Gouveia, L.; Pannecouque, C.; De Clercq, E.; Moreira, R.; Gomes, P.
Structure-Activity Relationships for Dipeptide Prodrugs of Acyclovir:
Implications for Prodrug Design. Eur. J. Med. Chem. 2009, 44, 2339−
2346.
(
[
drug of Dual Vascular Endothelial Growth Factor Receptor-2 and
Fibroblast Growth Factor Receptor-1 Kinase Inhibitor (BMS-
(68) Perez-Picaso, L.; Escalante, J.; Olivo, H.; Rios, M. Y. Efficient
́
Microwave Assisted Syntheses of 2,5-Diketopiperazines in Aqueous
5
(
40215). J. Med. Chem. 2008, 51, 1976−1980.
51) Bou-Chacra, N.; Melo, K. J. C.; Morales, I. A. C.; Stippler, E. S.;
Kesisoglou, F.; Yazdanian, M.; Lobenberg, R. Evolution of Choice of
Media. Molecules 2009, 14, 2836−2849.
̈
(69) Robillard, K. R.; Chan, G. N. Y.; Zhang, G.; la Porte, W.;
Bendayan, R. Role of P-Glycoprotein in the Distribution of the HIV
Protease Inhibitor Atazanavir in the Brain and Male Genital Tract.
Antimicrob. Agents Chemother. 2014, 58, 1713−1722.
Solubility and Dissolution Media After Two Decades of Biopharma-
ceutical Classification System. AAPS J 2017, 19, 989−1001.
(
52) Wang, J.; Yadav, V.; Smart, A. L.; Tajiri, S.; Basit, A. W. Toward
(70) Kis, O.; Walmsley, S. L.; Bendayan, R. In vitro and In situ
Oral Delivery of Biopharmaceuticals: an Assessment of the Gastro-
Evaluation of pH-Dependence of Atazanavir Intestinal Permeability
and Interactions with Acid-Reducing Agents. Pharm. Res. 2014, 31,
intestinal Stability of 17 Peptide Drugs. Mol. Pharm. 2015, 12, 966−
9
73.
2404−2419.
(
53) Parrish, K. E.; Mao, J.; Chen, J.; Jaochico, A.; Ly, J.; Ho, Q.;
Mukadam, S.; Wright, M. In Vitro and in Vivo Characterization of
CYP Inhibition by 1-Aminobenzotriazole in Rats. Biopharm. Drug
Dispos. 2016, 37, 200−211.
(
54) Vockley, J.; Andersson, H. C.; Antshel, K. M.; Braverman, N.
E.; Burton, B. K.; Frazier, D. M.; Mitchell, J.; Smith, W. E.;
Thompson, B. H.; Berry, S. A. Phenylalanine Hydroxylase Deficiency:
Diagnosis and Management Guideline. Genet. Med. 2014, 16, 188−
2
(
00.
55) Stappaerts, J.; Brouwers, J.; Annaert, P.; Augustijns, P. In Situ
Perfusion in Rodents to Explore Intestinal Drug Absorption:
Challenges and Opportunities. Int. J. Pharm. 2015, 478, 665−681.
(
56) Zhang, D.; Chando, T. J.; Everett, D. W.; Patten, C. J.; Dehal, S.
S.; Humphreys, W. G. In vitro Inhibition of UDP Glucuronosyl-
transferases by Atazanavir and Other HIV Protease Inhibitors and the
Relationship of This Property to In Vivo Bilirubin Glucuronidation.
Drug Metab. Dispos. 2005, 33, 1729−1739.
(
57) Gammal, R.; Court, M.; Haidar, C.; Iwuchukwu, O.; Gaur, A.;
Alvarellos, M.; Guillemette, C.; Lennox, J.; Whirl-Carrillo, M.;
Brummel, S.; Ratain, M.; Klein, T.; Schackman, B.; Caudle, K.;
Haas, D. Clinical Pharmacogenetics Implementation Consortium
(
CPIC) Guideline forUGT1A1and Atazanavir Prescribing. Clin.
Pharmacol. Ther. 2016, 99, 363−369.
58) Wang, B.-C.; Wang, L.-J.; Jiang, B.; Wang, S.-Y.; Wu, N.; Li, X.-
Q.; Shi, D.-Y. Application of Fluorine in Drug Design During 2010-
(
2
6
015 Years: A Mini-Review. Mini-Rev. Med. Chem. 2017, 17, 683−
92.
(
59) Ojima, I. Strategic Incorporation of Fluorine into Taxoid
Anticancer Agents for Medicinal Chemistry and Chemical Biology
Studies. J. Fluorine Chem. 2017, 198, 10−23.
(
60) Gillis, E. P.; Eastman, K. J.; Hill, M. D.; Donnelly, D. J.;
Meanwell, N. A. Applications of Fluorine in Medicinal Chemistry. J.
Med. Chem. 2015, 58, 8315−8359.
(
61) Meanwell, N. A. Fluorine and Fluorinated Motifs in the Design
and Application of Bioisosteres for Drug Design. J. Med. Chem. 2018,
1, 5822−5880.
62) Ritchie, T. J.; Macdonald, S. J. F. Physicochemical Descriptors
6
(
of Aromatic Character and Their Use in Drug Discovery. J. Med.
Chem. 2014, 57, 7206−7215.
(
63) Pelletier, J. C.; Chen, S.; Bian, H.; Shah, R.; Smith, G. R.;
Wrobel, J. E.; Reitz, A. B. Dipeptide Prodrugs of the Glutamate
Modulator Riluzole. ACS Med. Chem. Lett. 2018, 9, 752−756.
(
64) Gomes, P.; Vale, N.; Moreira, R. Cyclization-Activated
Prodrugs. Molecules 2007, 12, 2484−2506.
65) Santos, C.; Mateus, M. L.; Santos, A. P.; Moreira, R.; de
Oliveira, E.; Gomes, P. Cyclization-Activated Prodrugs. Synthesis,
(
V
DOI: 10.1021/acs.jmedchem.9b00002
J. Med. Chem. XXXX, XXX, XXX−XXX