In Vitro Pharmacokinetic Optimizations of AM2-S31N Channel Blockers Led to the Discovery of Slow-Binding Inhibitors with Potent Antiviral Activity against Drug-Resistant Influenza A Viruses

Article abstract of DOI:10.1021/acs.jmedchem.7b01536

Influenza viruses are respiratory pathogens that are responsible for both seasonal influenza epidemics and occasional influenza pandemics. The narrow therapeutic window of oseltamivir, coupled with the emergence of drug resistance, calls for the next-generation of antivirals. With our continuous interest in developing AM2-S31N inhibitors as oral influenza antivirals, we report here the progress of optimizing the in vitro pharmacokinetic (PK) properties of AM2-S31N inhibitors. Several AM2-S31N inhibitors, including compound 10b, were discovered to have potent channel blockage, single to submicromolar antiviral activity, and favorable in vitro PK properties. The antiviral efficacy of compound 10b was also synergistic with oseltamivir carboxylate. Interestingly, binding kinetic studies (K_d, K_on, and K_off) revealed several AM2-S31N inhibitors that have similar K_d values but significantly different K_on and K_off values. Overall, this study identified a potent lead compound (10b) with improved in vitro PK properties that is suitable for the in vivo mouse model studies.

Full text of DOI:10.1021/acs.jmedchem.7b01536

Article
pubs.acs.org/jmc
Cite This: J. Med. Chem. XXXX, XXX, XXX−XXX
In Vitro Pharmacokinetic Optimizations of AM2-S31N Channel
Blockers Led to the Discovery of Slow-Binding Inhibitors with Potent
Antiviral Activity against Drug-Resistant Influenza A Viruses
Yuanxiang Wang,†,⊥ Yanmei Hu,†,⊥ Shuting Xu,† Yongtao Zhang,† Rami Musharrafieh,‡
Raymond Kin Hau,† Chunlong Ma,§ and Jun Wang*,†,§
†Department of Pharmacology and Toxicology, College of Pharmacy, The University of Arizona, Tucson, Arizona 85721, United
States
‡Department of Chemistry and Biochemistry, The University of Arizona, Tucson, Arizona 85721, United States
§BIO5 Institute, The University of Arizona, Tucson, Arizona 85721, United States
S
* Supporting Information
ABSTRACT: Influenza viruses are respiratory pathogens that are responsible for both seasonal influenza epidemics and
occasional influenza pandemics. The narrow therapeutic window of oseltamivir, coupled with the emergence of drug resistance,
calls for the next-generation of antivirals. With our continuous interest in developing AM2-S31N inhibitors as oral influenza
antivirals, we report here the progress of optimizing the in vitro pharmacokinetic (PK) properties of AM2-S31N inhibitors.
Several AM2-S31N inhibitors, including compound 10b, were discovered to have potent channel blockage, single to
submicromolar antiviral activity, and favorable in vitro PK properties. The antiviral efficacy of compound 10b was also synergistic
with oseltamivir carboxylate. Interestingly, binding kinetic studies (Kd, Kon, and Koff) revealed several AM2-S31N inhibitors that
have similar Kd values but significantly different Kon and Koff values. Overall, this study identified a potent lead compound (10b)
with improved in vitro PK properties that is suitable for the in vivo mouse model studies.
There are two classes of FDA-approved influenza antivirals:
M2 channel blockers (amantadine and rimantadine),7 and
neuraminidase inhibitors (oseltamivir, zanamivir, and perami-
vir).8 Amantadine and rimantadine are no longer recommended
by the CDC for the prevention and treatment of influenza virus
infection due to the prevalent drug resistance.9 Oseltamivir is
currently the only FDA-approved oral influenza drug.
Zanamivir is administered through intranasal spray; therefore,
its use is limited in infants and critically ill patients. Recently
approved peramivir is an intravenous drug and its use is
restricted in clinical settings. Although a majority of current
circulating influenza viruses remain sensitive to oseltamivir,
oseltamivir-resistant viruses have been continuously re-
ported,10,11 and more alarmingly, the 2007−2008 seasonal
influenza H1N1 strain circulating in North America and Japan
INTRODUCTION
■
Influenza viruses are respiratory pathogens that are responsible
for both seasonal influenza epidemics and sporadic influenza
pandemics.1 Influenza A and B viruses co-circulate among
humans in each influenza season; however, the strains and
ratios change over time, and it is often difficult to predict which
strains will be circulating in the next influenza season and when
the next influenza pandemic will emerge.2 In addition, the
mortality rates for different subtypes of influenza viruses vary
significantly. For example, the mortality rates for human
infections with avian H5N1 and H7N9 viruses are 52.7% and
38.6%, respectively (WHO), compared to the 0.1−2.5%
mortality rate associated with past H1N1 influenza pandemic
viruses.3,4 Moreover, people in the high-risk groups of influenza
virus-related complications are more vulnerable for influenza
virus infection than immune-competent adults.5 Therefore,
effective antiviral drugs are undoubtedly needed for the
prevention and treatment of influenza virus infection.6
Received: October 14, 2017
© XXXX American Chemical Society
A
DOI: 10.1021/acs.jmedchem.7b01536
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Scheme 1. Synthesis of AM2-S31N Inhibitors
was completely resistant to oseltamivir due to the H275Y
mutation in the neuraminidase gene.12,13 In this regard, the next
generation of oral influenza drugs with a novel mechanism of
action is clearly needed.
the isoxazole-containing AM2-S31N inhibitor 8. The overall
yields range from 15% to 36%. Compounds 10a−10g were
synthesized by direct alkylation (Scheme 1B). The yields range
from 62% to 85%. The synthesis of compounds 11, 14, 15, and
16 was reported before.16,18,19
Hypothesis-Driven Optimization of the in Vitro PK
Properties Of AM2-S31N Inhibitors. The lead optimization
was guided by the solution NMR structure of compound 11 in
complex with AM2-S31N (19−49) (PDB 2LY0) (Figure 1).19
Compound 11 binds to the central cavity of the AM2-S31N
channel with its aryl group projecting toward the N-terminal
channel lumen. The distal thiophene substitution from
compound 11 forms hydrophobic interactions with V27 side
chain methyls, the isoxazole and the amine form bidentate
More than 95% of current circulating influenza A viruses
carry the AM2-S31N mutant in their AM2 genes,14 which
confer to amantadine and rimantadine resistance. Given the
high prevalence of the AM2-S31N mutant, we hypothesize that
AM2-S31N inhibitors will have potent antiviral activity against
both oseltamivir-sensitive and -resistant influenza A strains.
Indeed, we have shown that our rationally designed AM2-S31N
inhibitors inhibit multiple human influenza A viruses with
submicromolar efficacy, including oseltamivir-resistant
strains.15−18 Encouraged by these preliminary results, in this
study, we report our progress in optimizing the in vitro PK
properties of AM2-S31N inhibitors with the aim of prioritizing
lead compounds for the next step in vivo mice studies. Through
iterative cycles of hypothesis-driven PK optimization, we
identified one compound, 10b, with optimal in vitro PK
properties. Interestingly, unlike most other previously reported
AM2-S31N inhibitors which have fast Kon and slow Koff values,
compound 10b is a slow binding channel blocker with slow Kon
and Koff values but having a low micromolar Kd. Subsequent
kinetic studies of several representative AM2-S31N inhibitors
revealed that the antiviral efficacy of AM2-S31N inhibitors has a
better correlation with their Kd values rather than their
percentage of channel blockage at the 2 min time point.
RESULTS AND DISCUSSION
■
Chemistry. Compounds 8a−8f were synthesized according
to a previous optimized synthesis route (Scheme 1A).18 Briefly,
condensation of methyl ketone 1 and diethyl oxalate 2 using
potassium tert-butoxide as the base gave the β-ketone ester
intermediate 3, which was cyclized to isoxazole ester 4 by
heating at 50 °C in methanol with hydroxylamine hydro-
chloride. The ester 4 was reduced to alcohol 5 by sodium
borohydride and was then converted to bromide 6 by triphenyl
phosphine and carbon tetrabromide in dichloromethane. Last,
alkylation of bromide 6 with the amantadine analogue 7 gave
Figure 1. Solution NMR structure of compound 11 in complex with
the transmembrane domain of AM2-S31N (PDB 2LY0).19 The
transparency of the front helix was set as 0.5 for clarity.
B
DOI: 10.1021/acs.jmedchem.7b01536
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Table 1. Channel Blockage, Antiviral Efficacy, Cytotoxicity, and Mouse Microsomal Stability of AM2-S31N Inhibitors
a
b
Percentage channel blockage is presented as the mean standard deviation of three independent experiments. Antiviral assay was performed in
c
plaque assay. The EC50 values are the mean standard deviation of two independent experiments. Cytotoxicity was measured using the neutral red
method with a 48 h incubation time.25 The CC50 values are the mean standard deviation of three independent experiments. T1/2 was determined
d
in mouse liver microsomes. N.T. = not tested.
C
DOI: 10.1021/acs.jmedchem.7b01536
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 2. Hypothesis-driven optimization of the microsomal stability of AM2-S31N inhibitors. (A) Two FDA-approved oral drugs that contain the
hydroxyl adamantane. (B) Strategy to increase the microsomal stability of AM2-S31N inhibitor 11.
hydrogen bonds with the N31 side chain amide, and the
adamantane cage fits in the cavity created by the G34 from
AM2-S31N.
cyclopropyl-containing analogues 10c, 10d, 10e, 10f, and 10g.
Compound 10d did not block the AM2-S31N channel (5.3%
channel blockage at 100 μM) and therefore had no antiviral
activity. All other compounds (10c, 10e, 10f, and 10g) had
potent antiviral activity (EC50 < 3.1 μM) and a high selectivity
index. Compounds 10c, 10e, and 10g were also stable in mouse
microsomes with T1/2 of 97.2, >145, and >145 min,
respectively. Despite the potent antiviral activity of compounds
10b, 10c, 10e, 10f, and 10g, they had a moderate to weak
channel blockage against the AM2-S31N channel (33.8−67.3%
at 100 μM). It is noteworthy that the value of percentage
channel blockage reported herein was the value at the 2 min
time point after applying 100 μM of compound to the oocyte
bathing medium. Therefore, this value may or may not
represent the true binding potency of the compound because
this experiment is under kinetic control. The apparent
discrepancy between these compounds’ weak to moderate
percentage channel blockage at the 2 min time point and their
potent antiviral activity suggest that they might be slow-binding
inhibitors. To confirm this hypothesis, we performed detailed
kinetic studies of these AM2-S31N inhibitors (10b, 10c, 10e,
10f, and 10g), and the results are shown in the next section.
In parallel, guided by the solution NMR structure of AM2-
S31N in complex with compound 11 (PDB 2LY0) (Figure 1),
we also designed and synthesized isoxazole adamantane
analogues with cyclobutyl (8a, 8b), cyclopentyl (8c, 8d), and
cyclohexyl (8e, 8f) substitutions. Substituting cyclopropyl with
cyclobutyl, cyclopentyl, and cyclohexyl should be accommo-
dated as this substitution forms hydrophobic interaction with
the V27 side chain methyls (Figure 1). As expected, all
compounds were found to have potent channel blockage
against the AM2-S31N channel (>84% at 100 μM) and potent
antiviral activity against the AM2-S31N-containing A/Califor-
nia/07/2009 (H1N1) virus (EC50 ≤ 1 μM). However,
replacing cyclopropyl in compound 10b with cyclobutyl (8b),
cyclopentyl (8d), and cyclohexyl (8f) led to a gradual decrease
of the mouse microsomal stability (Table 1). These results
suggest cyclopropyl substitution is more metabolically stable
than the corresponding cyclobutyl, cyclopentyl, and cyclohexyl
substitutions.
As a first step to optimize the in vitro PK properties of the
AM2-S31N inhibitors, we profiled the mouse liver microsomal
stability of compound 11, which is one of the most potent
AM2-S31N inhibitors we have discovered so far (Table 1).19 It
was found that compound 11 had an ultrashort half-life of 1.0
min in mouse liver microsomes, which excluded it from being
tested in in vivo mouse studies. Inspection of the chemical
structure of compound 11 revealed that there might be three
possible metabolic soft spots: the adamantane, the isoxazole,
and the thiophene.20 Amantadine was shown to be metabolized
to hydroxyl adamantane in vivo,21 the isoxazole ring can be
cleaved at the N−O bond,22 and the thiophene ring is known
to be easily oxidized.23 To increase the microsomal stability of
compound 11, we first searched the FDA-approved drug
database for orally bioavailable adamantane-containing drugs,
hoping to gain insights how to improve the microsomal stability
of adamantane. The search revealed two compounds,
vildagliptin (12) and saxagliptin (13) (Figure 2A), both of
which are dipeptidyl peptidase-4 (DPP-4) inhibitors and are
used as antidiabetic drugs. Vildagliptin (12) and saxagliptin
(13) each contain hydroxyl adamantane instead of adamantane.
The biological half-life T1/2 of vildagliptin (12) and saxagliptin
(13) were 2−3 and 2.5 h, respectively, following a single oral
dose in healthy volunteers.24 The presence of hydroxyl
adamantane instead of adamantane in vildagliptin (12) and
saxagliptin (13) might confer to their high microsomal stability
and oral bioavailability. We therefore applied the same strategy
in our PK optimization and converted compound 11 to the
hydroxyl adamantane-containing compound 14 (Figure 2B).
Compound 14 had similar channel blockage and antiviral
efficacy as compound 11 (Table 1) and had an improved
selectivity index over compound 11. Gratifyingly, the half-life of
compound 14 in mouse microsomes was increased to 31.5 min.
This result suggests that hydroxyl adamantane is indeed more
metabolically stable than adamantane. Following the same
rationale, we tested the mouse microsomal stability of another
hydroxyl adamantane-containing AM2-S31N inhibitor, com-
pound 15. The mouse microsomal half-life of compound 15
was 13.4 min, which was a significant improvement compared
to compound 11 but still requires further optimization. Next, to
remove another possible metabolic labile spot, the thiophene23
in compounds 11, 14, and 15, we tested the microsomal
stability of compound 10a, which contains a cyclopropyl
substitution instead of thiophene. The mouse microsomal half-
life of compound 10a was 12.8 min, which was a significant
improvement compared to compound 11 (T1/2 = 1.0 min).
Substituting adamantane in 10a with hydroxyl adamantane gave
compound 10b, which had a half-life of more than 145 min.
Encouraged by this result, we then synthesized a few other
Binding Kinetic Studies Of AM2-S31N Inhibitors. In
our previous studies, we plotted the AM2-S31N inhibitors’
percentage channel blockage at the 2 min time point with their
antiviral activity in a plaque assay.15 It was found that
compounds with a higher percentage of channel blockage
normally have more potent antiviral efficacy. However, in this
study, we found a few exceptions such as compounds 10b, 10c,
10e, 10f, and 10g, which all had moderate to weak channel
blockage ranging from 33.8% to 67.3% at the 2 min time points
when tested at 100 μM drug concentration. According to the
previous correlation, the predicted antiviral EC50 values of these
compounds should have been greater than 8 μM.15 However,
D
DOI: 10.1021/acs.jmedchem.7b01536
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 3. Kinetic studies of AM2-S31N inhibitors. (A) Binding and washing curves were recorded for two oocytes. One of the representative
recording traces is shown. During the washing period, in order to prevent prolonged acidification of oocytes, a pH 5.5 pulse, instead of continuous
application of pH 5.5 barth solution, was applied. The red bar shows when pH 5.5 pulse was applied, and the blue bar shows when 100 μM of
compound was applied at pH 5.5. Nonlabeled areas shows when pH 8.5 solution was applied. (B) The association and dissociation equations used to
fit the curves in A.
all of them had potent antiviral activity with EC50 values less
than 3.1 μM. One possible explanation is that these compounds
might be slow-binding inhibitors and equilibrium might not be
achieved at the 2 min time point; therefore, the reported
percentage channel blockage at the 2 min time point actually
underestimates the true potency of these compounds. To
validate this hypothesis, we chose compounds 10b, 10c, 10e,
10f, and 10g for detailed kinetic studies. Two additional
compounds 10a and 16 were included as controls for
comparisons.
The Kd values of compounds 10a, 10b, 10c, 10e, 10f, 10g,
and 16 were obtained by fitting the inhibition and washing
curves (Figure 3A) with the association and dissociation
equations (Figure 3B). All compounds were tested at 100 μM.
The best-fit results are shown in Table 2. As shown in Figure
3A and Table 2, despite the moderate to low channel blockage
at the 2 min time point for compounds 10b, 10c, 10e, and 10f
E
DOI: 10.1021/acs.jmedchem.7b01536
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
a
Table 2. Fitting Results from Kinetic Studies of AM2-S31N Inhibitors
a
The best fit values for the inhibition by compounds 10a, 10b, 10c, 10e, 10f, 10g, and 16. Results were obtained in Prism 5.0 using the equations
shown in Figure 3B. The compound concentration was 100 μM. The results are the mean standard deviation of two independent experiments.
b
cLogP was calculated using the QikProp in Schrodinger Glide.
̈
(44.8−67.3%), they had comparable Kd values (4.5−11.5 μM)
as the potent channel blocker compound 10a (7.7 μM).
Consistent with our hypothesis, compounds 10b, 10c, 10e, and
10f are indeed slow-binding inhibitors as shown by their slow
Kon (<1 × 104 M−1 min−1) and slow Koff (<0.1 min−1) values.
Slow Koff is defined as the value being 10-fold lower than the
fast Koff value. In contrast, compound 10a has fast Kon (>1 ×
104 M−1 min−1) and slow Koff (<0.1 min−1) values. Compound
time (>30 min), which often results in cell death and
membrane leakage. On average, we could only obtain one
quality binding and washing curve out of five measurements
from different oocytes. Therefore, instead of simply filtering
compounds based on their percentage channel blockage at the
2 min time point in the primary electrophysiological assays, all
compounds should be tested in secondary antiviral plaque
assays in parallel. In this way, potent compounds that show
slow-binding kinetics in inhibiting AM2-S31N proton channel
will not be mistakenly filtered.
Antiviral Activity of Compound 10b against Drug-
Resistant Influenza A Viruses. The antiviral activity of
compound 10b was further tested against five additional human
influenza A viruses, among which A/Denmark/528/2009
(H1N1), A/Washington/29/2009 (H1N1), A/Texas/04/
2009 (H1N1), and A/North Carolina/29/2009 (H1N1) are
resistant to both amantadine and oseltamivir carboxylate, while
A/Switzerland/9715293/2013 (H3N2) is resistant to amanta-
dine but sensitive to oseltamivir. These five strains, together
with A/California/07/2009 (H1N1), all contain the AM2-
S31N mutant. It was found that compound 10b inhibited all six
strains with EC50 values less than 1 μM (Figure 4). In addition,
compound 10b also showed potent inhibition against the AM2-
S31N-containg A/WSN/33 (H1N1) virus with an EC50 value
of 0.5 0.2 μM. The cytotoxicity CC50 values of compound
10b were greater than 300 μM for both MDCK and A549 cells,
which corresponds to a selectivity index of greater than 375.
Overall, AM2-S31N inhibitors such as 10b were shown to have
potent antiviral activity against AM2-S31N-containing influenza
A viruses with diverse genetic backgrounds.
10g has the lowest Kon of 1745
102 M−1min−1 and the
highest Kd of 22.7 2.0 μM, yet it had potent antiviral activity
with an EC50 of 2.6 0.7 μM. In contrast, compound 16 has
both fast Kon (>1 × 104 M−1 min−1) and fast Koff (>0.1 min−1),
with a Kd of 17.3 2.8 μM.
In summary, on the basis of channel binding kinetics, AM2-
S31N inhibitors can be grouped into three categories: (1)
Compounds with fast Kon and slow Koff, such as compound 10a.
This group of compounds normally has potent antiviral activity.
In this case, the percentage channel blockage at the 2 min time
point is an accurate predictor for the antiviral activity. (2)
Compounds with slow Kon and slow Koff, such as the newly
discovered slow-binding inhibitors 10b, 10c, 10e, 10f, and 10g.
Compounds in this category may or may not have potent
antiviral activity, which depends on their Kd values. Because of
their slow Kon values, the percentage channel blockage at the 2
min time point cannot accurately predict the antiviral potency.
(3) Compounds with fast Kon and fast Koff such as compound
16. Similar to the slow-binding inhibitors, the antiviral potency
of compounds in this category depends on their Kd values not
their percentage inhibition at the 2 min time point.
It is noteworthy that although Kd values are more accurate
predictors for the compounds’ antiviral activity, in practice, it is
not feasible to measure the Kd value for every compound. The
reason for this is that Kd measurements require bathing AM2-
S31N-expressing oocytes at pH 5.5 for an extended period of
in Vitro PK Profiling of Compound 10b. Given
compound 10b’s potent antiviral activity, high selectivity
index, and long half-life in mouse microsomes, we further
profiled other in vitro PK properties of compound 10b, which
F
DOI: 10.1021/acs.jmedchem.7b01536
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 4. Plaque assays of compound 10b in inhibiting human influenza A viruses. The EC50 values are the mean standard deviation of two
independent experiments. Representative plaque assay images from the one of the two independent experiments are shown.
include human liver microsome stability, cell permeability, and
cytochrome P450 inhibition (Table 3). It was found that
compound 10b has high membrane permeability in Caco-2
cells, and it did not show inhibition against the five isoforms of
CYP enzymes: 1A2, 2C9, 2C19, 2D6, and 3A4. Compound 10b
has a similar high stability in the human liver microsomes, with
a half-life greater than 145 min. Compound 10b also did not
violate the Lipinski rule-of-five. In summary, the in vitro PK
profiling predicts that compound 10b will likely have high oral
bioavailability.
Combination Therapy Potential of Compound 10b
with Oseltamivir Carboxylate. Combination therapy has
been proven effective in decreasing the incidence of drug
resistance as well as in lowering cytotoxicity.26 As such, we are
interested in exploring the combination therapy potential of
AM2-S31N inhibitors such as 10b with oseltamivir carboxylate.
As AM2-S31N inhibitor 10b targets a different viral replication
stage than oseltamivir carboxylate, it is expected that a
combination of compound 10b with oseltamivir carboxylate
will have a synergistic antiviral effect. To test this hypothesis,
we applied the standard fractional inhibitory concentration
G
DOI: 10.1021/acs.jmedchem.7b01536
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Table 3. In Vitro PK Properties of AM2-S31N Inhibitor 10b
a
b
Lipinski rule-of-five: MW < 500, HBD < 5, HBA < 10, clogP < 5; values were calculated using the QikProp in Schrodinger Glide. Bidirectional
̈
c
permeability in Caco-2 cells was assessed over a 120 min incubation at 37 °C and 5% CO2 with saturated humidity; CYP inhibition was tested at 10
μM, and IC50 was derived from the standard curve.
Table 4. Combination Therapy of 10b with Oseltamivir Carboxylate
a
combination ratio (EC50)
EC50 in combination
EC50 alone
oseltamivir(nM)
EC50 equivalent
b
10b:oseltamivir
10b (μM)
oseltamivir (nM)
10b (μM)
10b
oseltamivir
FICI
10:1
5:1
0.13 0.014
0.14 0.031
0.092 0.014
0.078 0.021
0.039 0.0044
0.06 0.007
0.11 0.016
0.42 0.027
1.91 0.18
1.97 0.22
0.394
0.424
0.279
0.236
0.118
0.005
0.009
0.035
0.159
0.164
0.40
0.43
0.31
0.40
0.28
1:1
0.33 0.01
12.0 2.5
1:5
1:10
a
Concentration in EC50 equivalent was the normalized concentration that was calculated by dividing the EC50 of drug in combination with its EC50
b
alone. FICI was the sum of 10b and oseltamivir EC50-equivalent concentrations used in each combination.
index (FICI) method to evaluate the combination therapy
effect of compound 10b with oseltamivir carboxylate.27−30
A
FICI value of lower or higher than 1 indicates synergy or
antagonism, respectively. It was found that all five sets of
combination gave FICI less than 0.5 (Table 4), suggesting
strong synergism by combing 10b with oseltamivir carboxylate.
Among the five sets of results, a combination of 0.039 μM 10b
with 1.97 nM oseltamivir (combination ratio 1:10) exerted the
highest synergistic efficacy with a FICI value of 0.28.
Docking Model of Compound 10b in AM2-S31N. To
gain insight into how compound 10b binds to the AM2-S31N
channel, we performed molecular modeling using the
Schrodinger Glide standard precision docking. The solution
̈
NMR structure of AM2-S31N (PDB 2LY0) was used as the
protein template. It was found that compound 10b can bind to
the AM2-S31N channel in two different conformations (Figure
5A,B). In one conformation (Figure 5A), the isoxazole nitrogen
from 10b forms a hydrogen bond with the N31 side chain
amide NH2 from one of the monomers, while the ammonium
from 10b forms another hydrogen bond with the N31 side
chain amide carbonyl from the opposite monomer. In the
another binding confirmation (Figure 5B), the isoxazole
nitrogen and ammonium from 10b form a bidentate hydrogen
bond with the N31 side chain amide from one monomer. In
both conformations (Figure 5A,B), the cyclopropyl substitution
from 10b form hydrophobic interactions with the V27 side
chain methyls, and the hydroxyl substitution on adamantane
forms a hydrogen bond with the A30 backbone amide carbonyl.
Figure 5. Docking models of compound 10b in the transmembrane
domain of AM2-S31N (PDB 2LY0).19 The transparency of the front
helix was set as 0.5 for clarity. Docking was performed using
Schrodinger Glide standard precision.
̈
selective AM2-S31N channel blockers that inhibit multiple
strains of human influenza A viruses, including viruses that are
resistant to amantadine, oseltamivir, or both. However, the first
generation of AM2-S31N inhibitors such as compound 11 was
found to have low stability in mouse microsomes, precluding
them from further development. Guided by a hypothesis-driven
approach, we were able to significantly improve the microsomal
stability of AM2-S31N inhibitors. The lead compound 10b had
a T1/2 of more than 145 min in both mouse and human liver
microsomes, and it was also highly permeable in Caco-2 cells
and did not inhibit five isoforms of CYP. Interestingly, albeit
the moderate percentage channel blockage (47.9 1.0%) of
compound 10b was at the 2 min time point when tested at 100
μM, 10b had similar antiviral activity as compound 10a, which
showed 85.4 2.2% channel blockage at the 2 min time point
when tested at 100 μM. The potent antiviral activity of
compound 10b was unexpected: according to previous
CONCLUSIONS
■
Although novel drugs are urgently needed to address the unmet
medical needs, drug discovery is unfortunately a lengthy and
expensive venture.31 In the case of AM2-S31N drug discovery,
no progress was made for decades until the report of the first-
generation AM2-S31N inhibitors in 2013.19,32 Thereafter, lead
optimization led to the development of more potent and
H
DOI: 10.1021/acs.jmedchem.7b01536
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
dropwise. The resulting solution was stirred at room temperature
overnight. The reaction was quenched with 1 N HCl and extracted
with ethyl acetate (3×). The combined organic layers were dried over
MgSO4, filtered, and concentrated under reduced pressure. The crude
product was dissolved in methanol, and hydroxylamine hydrochloride
(1.04g, 1.5 equiv) was added. The solution was heated to 50 °C for 6
h. The solvent was removed under reduced pressure, and the resulting
isoxazole ester was purified by flash column chromatography (10−20%
EtOAc/hexane) to give the final product as a white solid. Yield: 56%.
1H NMR (400 MHz, CDCl3) δ 6.37 (s, 1H), 4.40 (q, J = 7.2 Hz, 2H),
correlation study results, the predicted antiviral EC50 value of
compound 10b would be more than 10 μM. However, in
reality, compound 10b inhibits multiple AM2-S31N-containing
influenza A viruses with EC50 values ranging from 0.2 to 0.8 μM
(Figure 4). The apparent discrepancy led us to examine the
binding kinetics of these AM2-S31N inhibitors. By fitting the
binding and washing curves with the association and
dissociation equations, we calculated the Kon, Koff, and Kd
values of compounds 10a and 10b, along with a few other
compounds (10c, 10e, 10f, 10g, and 16). The results showed
that the newly discovered compounds 10b, 10c, 10e, 10f, and
10g are slow-binding inhibitors and the binding equilibrium
was not achieved at the 2 min time point when tested at 100
μM. Therefore, the reported percentage inhibition at the 2 min
time point underestimates the true potency of these
compounds (10b, 10c, 10e, 10f, and 10g). As such, although
compound 10b had significantly less percentage channel
blockage than 10a at the 2 min time point (47.9% versus
85.4%), they had similar Kd values (4.5 versus 7.7 μM). These
results suggest that the Kd value is a more accurate predictor of
the compound’s antiviral activity rather than the percentage
channel blockage at the 2 min time point.
Overall, the results presented herein may have important
implications for other ion channel drug discovery program. As
the electrophysiological assay is time-consuming, it is generally
used as a primary assay to test compounds at a single drug
concentration. However, as this process is under kinetic
control, slow-binding inhibitors might be missed from this
primary screening. To ensure slow-binding inhibitors are not
being mistakenly filtered, functional assays should be performed
in parallel.
3.26−3.18 (m, 1H), 2.11−2.05 (m, 2H), 1.79−1.62 (m, 6H), 1.37 (t, J
= 7.2 Hz, 3H). C11H15NO3 EI-MS: m/z (M + H+): 210.2 (calculated),
210.0 (found).
3-(Bromomethyl)-5-cyclopentyl-1,2-oxazole (6b). The ester 4b
(0.21g, 1 equiv) was dissolved in methanol and cooled down to 0 °C.
NaBH4 (0.15g, 4 equiv) was added in small portions to the solution
over 10 min. The mixture was warmed to room temperature and
stirred for 4 h. The reaction was quenched by adding diluted HCl, and
the organic solvent was removed under reduced pressure. The
resulting aqueous layer was extracted with ethyl acetate (3×), and the
organic layers were combined and dried over MgSO4, and the solvent
was removed under reduced pressure. This hydroxyl intermediate 5b
was used for the next step without further purification. Hydroxyl
intermediate (0.17 g, 1 equiv) was dissolved in DCM, and the resulting
solution was cooled down to 0 °C. CBr4 (0.50g, 1.5 equiv) and PPh3
(0.39 g, 1.5 equiv) were added sequentially. The solution was stirred at
0 °C for 20 min and gradually warmed up to room temperature. The
solvent was removed under reduced pressure, and the residue was
purified by flash column chromatography (20% hexane/DCM) to give
the desired intermediate 6b. Yield: 62%. 1H NMR (400 MHz, CDCl3)
δ 6.06 (s, 1H), 4.39 (s, 2H), 3.24−3.15 (m, 1H), 2.14−2.05 (m, 2H),
1.83−1.65 (m, 6H). C9H12BrNO EI-MS: m/z (M + H+): 231.1
(calculated), 231.0 (found).
3-(Bromomethyl)-5-cyclobutyl-1,2-oxazole (6a). The synthesis
and characterization of bromide 6a was reported.18
In summary, this study led to the development of a
promising lead compound (10b) with potent antiviral activity,
high selectivity index, and favorable in vitro PK properties, and
combination therapy of 10b with oseltamivir carboxylate also
showed strong synergism.
3-(Bromomethyl)-5-cyclohexyl-1,2-oxazole (6c). The synthesis
and characterization of bromide 6c was reported.18
General Procedure of Alkylations. The bromide (1 equiv) and
amantadine or 1-amino-3-hydroxyadamantane or 5-aminoadamantan-
2-one (1.5 equiv) were dissolved in 2-propanol; CsI (0.1 equiv) and
triethyl amine (2 equiv) were then added. The reaction mixture was
heated to reflux overnight. The solvent was removed under reduced
pressure, and the resulting residue was extracted with ethyl acetate and
water. The organic layer was separated, dried over anhydrous MgSO4,
filtered, and concentrated under reduced pressure. The mixture was
then purified by silica gel flash column chromatography (5−10%
CH3OH/CH2Cl2) to give the final product.
EXPERIMENTAL SECTION
■
Chemistry. Chemicals were ordered from commercial sources and
were used without further purification. Synthesis procedures for
reactions described in Scheme 1 were show below. All final
1
compounds were purified by flash column chromatography. H and
13C NMR spectra were recorded on a Bruker-400 NMR spectrometer.
N-[(5-Cyclobutyl-1,2-oxazol-3-yl)methyl]adamantan-1-amine
(8a). Compound 8a was synthesized according to the above-described
alkylation procedure starting with bromide 6a. The characterization of
compound 8a was reported before.18
Chemical shifts are reported in parts per million referenced with
respect to residual solvent (CD3OD) 3.31 ppm, (DMSO-d6) 2.50
ppm, and (CDCl3) 7.24 ppm or from internal standard tetramethylsi-
lane (TMS) 0.00 ppm. The following abbreviations were used in
reporting spectra: s, singlet; d, doublet; t, triplet; q, quartet; m,
multiplet; dd, doublet of doublets; ddd, doublet of doublet of doublets.
All reactions were carried out under N2 atmosphere unless otherwise
stated. HPLC-grade solvents were used for all reactions. Flash column
chromatography was performed using silica gel (230−400 mesh,
Merck). Low-resolution mass spectra were obtained using an ESI
technique on a 3200 Q Trap LC/MS/MS system (Applied
Biosystems). The purity was assessed by using a Shimadzu LC-MS
with a Waters XTerra MS C-18 column (part no. 186000538), 50 mm
× 2.1 mm, at a flow rate of 0.3 mL/min; λ = 250 and 220 nm; mobile
phase A, 0.1% formic acid in H2O, and mobile phase B′, 0.1% formic in
60% 2-propanol, 30% CH3CN, and 9.9% H2O. All compounds
submitted for testing in TEVC assay, plaque assay, cytotoxicity assay,
and PK studies were confirmed to be >95.0% purity by LC-MS traces.
Synthesis Procedures. Ethyl 5-Cyclopentyl-1,2-oxazole-3-car-
boxylate (4b). To a stirred solution of diethyl oxalate (1.46 g, 1.1
equiv) and 1-cyclopentylethanone (1.12g, 1 equiv) in toluene was
added a solution of potassium tert-butoxide in THF (1.2 equiv)
3-{[(5-Cyclobutyl-1,2-oxazol-3-yl)methyl]amino}adamantan-1-ol
(8b). Compound 8b was synthesized according to the above-described
1
alkylation procedure starting with bromide 6a. Yield: 78%. H NMR
(400 MHz, CDCl3): δ 6.03 (s, 1H), 3.83 (s, 2H), 3.63−3.55 (m, 1H),
2.41−2.31 (m, 2H), 2.31−2.21 (m, 4H), 2.10−1.90 (m, 4H), 1.72−
1.64 (m, 6H), 1.64−1.59 (m, 4H), 1.55−1.50 (m, 2H). 13C NMR
(100 MHz, CDCl3): δ 176.51, 163.24, 99.27, 69.64, 54.51, 50.00,
44.34, 41.07, 37.01, 35.09, 32.07, 30.73, 28.00, 18.74. C18H26N2O2 EI-
MS: m/z (M + H+): 303.4 (calculated), 303.0 (found).
N-[(5-Cyclopentyl-1,2-oxazol-3-yl)methyl]adamantan-1-amine
(8c). Compound 8c was synthesized according to the above-described
1
alkylation procedure starting with bromide 6b. Yield: 75%. H NMR
(400 MHz, CDCl3 + CD3OD): δ 6.85 (s, 1H), 4.22−4.15 (m, 2H),
3.20−3.08 (m, 1H), 2.17−2.10 (m, 3H), 2.08−2.02 (m, 6H), 2.02−
1.95 (m, 2H), 1.75−1.61 (m, 12H). 13C NMR (100 MHz, CDCl3): δ
178.91, 157.08, 100.94, 59.01, 38.73, 37.48, 35.85, 35.38, 31.84, 29.14,
25.24. C19N28N2O EI-MS: m/z (M + H+): 301.4 (calculated), 301.0
(found).
I
DOI: 10.1021/acs.jmedchem.7b01536
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
3-{[(5-Cyclopentyl-1,2-oxazol-3-yl)methyl]amino}adamantan-1-
ol (8d). Compound 8d was synthesized according to the above-
described alkylation procedure starting with bromide 6b. Yield: 82%.
1H NMR (400 MHz, CDCl3 + CD3OD): δ 6.71 (s, 1H), 4.20−4.11
(m, 2H), 3.23−3.12 (m, 1H), 2.40−2.32 (m, 2H), 2.14−1.99 (m, 4H),
1.98−1.85 (m, 4H), 1.81−1.61 (m, 10H), 1.59−1.50 (m, 2H). 13C
NMR (100 MHz, CDCl3): δ 180.05, 155.89, 100.68, 68.50, 60.21,
42.73, 37.44, 36.66, 34.17, 31.80, 30.19, 25.20. C19H28N2O2 EI-MS: m/
z (M + H+): 317.4 (calculated), 317.0 (found).
N-[(5-Cyclohexyl-1,2-oxazol-3-yl)methyl]adamantan-1-amine
(8e). Compound 8e was synthesized according to the above-described
alkylation procedure starting with bromide 6c. The characterization of
compound 8e was reported before.18
67.82, 59.95, 54.82, 46.05, 43.87, 37.93, 36.93, 34.30, 30.13, 12.04,
11.42. C16H23N3OS EI-MS: m/z (M + H+): 306.4 (calculated), 306.0
(found).
Characterization of compounds 11, 14, 15, and 16 was reported
before.16,18,19
Two-Electrode Voltage Clamp (TEVC) Assay. The compounds
were tested in a two-electrode voltage clamp assay using Xenopus laevis
frog oocytes microinjected with RNA expressing either the AM2-WT
or the AM2-S31N mutant of the AM2 protein, as previously
reported.33 The M2 sequences used for this study were identical to
the A/California/07/2009 (H1N1) M2 sequence (AM2-S31N) and
the corresponding AM2-N31S mutant (AM2-WT). The M2 sequences
from A/WSN/33 (H1N1) and A/California/07/2009 (H1N1) differ
only by one amino acid at the transmembrane region at position 28: in
A/WSN/33 (H1N1) is valine and in A/California/07/2009 (H1N1)
is isoleucine. The potency of the inhibitors was expressed as
percentage inhibition of AM2 current observed after 2 min of
incubation with 100 μM of compounds at pH 5.5. All measurements
were repeated three times with different oocytes. For kinetic studies,
during the binding period, the oocyte was fluxed with pH 5.5 buffer
containing 100 μM of compound until binding equilibrium was
achieved. During the washing period, in order to prevent prolonged
acidification of oocytes, a pH 5.5 pulse, instead of continuous
application of pH 5.5 barth solution, was applied until equilibrium was
achieved.
3-{[(5-Cyclohexyl-1,2-oxazol-3-yl)methyl]amino}adamantan-1-ol
(8f). Compound 8f was synthesized according to the above-described
1
alkylation procedure starting with bromide 6c. Yield: 85%. H NMR
(400 MHz, CD3Cl + CD3OD): δ 6.57 (s, 1H), 4.19−4.02 (m, 2H),
2.80−2.63 (m, 1H), 2.43−2.24 (m, 2H), 2.06−1.79 (m, 8H), 1.79−
1.58 (m, 7H), 1.59−1.45 (m, 2H), 1.45−1.16 (m, 5H). 13C NMR
(100 MHz, CDCl3+CD3OD): δ 180.06, 155.59, 100.11, 68.79, 60.48,
42.68, 36.70, 36.31, 35.30, 34.10, 30.89, 30.11, 25.55, 25.47.
C20H30N2O2 EI-MS: m/z (M + H+): 331.5 (calculated), 332.0
(found).
Compounds 10a−10g were synthesized using the general alkylation
procedure described above starting from commercial available
bromides.
Plaque Assay. The plaque assay was performed as previously
reported29,30 except MDCK cells expressing ST6Gal I were used
instead of regular MDCK cells.34 Briefly, a confluent monolayer of
ST6Gal I MDCK cells was incubated with ∼100 pfu virus/well in
DMEM with 0.5% BSA for 1 h at 4 °C then at 37 °C for 1 h. The
inoculums were removed, and the cells were washed with phosphate
buffered saline (PBS). The cells were then overlaid with DMEM
containing 1.2% Avicel microcrystalline cellulose (FMC BioPolymer,
Philadelphia, PA) and NAT (2.0 μg/mL). To examine the effect of the
compounds on plaque formation, the overlay media was supplemented
with compounds at testing concentrations. After 2 days of incubation
at 37 °C with 5% of CO2 in the cell culture incubator, the overlay was
removed and the cell monolayers were fixed and stained with crystal
violet dye solution (0.2% crystal violet, 20% methanol). Influenza A
viruses A/Switzerland/9715293/2013 X-247 (H3N2), FR-1366, A/
Washington/29/2009 (H1N1), FR-460, A/North Carolina/29/2009
(H1N1), FR-488, and A/California/07/2009 (H1N1), FR-201, were
obtained through the Influenza Reagent Resource, Influenza Division,
WHO Collaborating Center for Surveillance, Epidemiology and
Control of Influenza, Centers for Disease Control and Prevention,
Atlanta, GA, USA. Influenza virus A/Denmark/528/2009 (H1N1) was
obtained from Dr. Elena Govorkova at St. Jude Children’s Research
Hospital. Influenza virus A/Texas/04/2009 (H1N1) was obtained
from Dr. James Noah at the Southern Research Institute.
Cytotoxicity Assay. Evaluation of the cytotoxicity of compounds
and the efficacy of compounds against influenza-induced cytopathic
effect was carried out using the neutral red uptake assay.25 Briefly,
80000 cells/mL of MDCK or A549 cells in DMEM medium
supplemented with 10% FBS and 100 U/mL penicillin−streptomycin
were dispensed into 96-well cell culture plates at 100 μL/well. Then 24
h later, the growth medium was removed and washed with 100 μL of
PBS buffer, then for the cytotoxicity assay, 200 μL of fresh DMEM (no
FBS) medium containing serial diluted compounds was added to each
well. After incubating for 48 h at 37 °C with 5% CO2 in a CO2
incubator, the medium was removed, and then 100 μL of DMEM
medium containing 40 μg/mL neutral red was added and the plate was
incubated for 4 h at 37 °C. The amount of uptaken neutral red was
determined at absorbance 540 nm using a Multiskan FC microplate
photometer (Fisher Scientific). The CC50 values were calculated from
best-fit dose response curves with variable slope in GraphPad Prism
version 5.
N-[(5-Cyclopropyl-1,2-oxazol-3-yl)methyl]adamantan-1-amine
(10a). The characterization of 10a was reported before.19
3-{[(5-Cyclopropyl-1,2-oxazol-3-yl)methyl]amino}adamantan-1-
1
ol (10b). Yield: 72%. H NMR (400 MHz, CDCl3): δ 5.92 (s, 1H),
3.83 (s, 2H), 2.29−2.23 (m, 2H), 2.02−1.91 (m, 2H), 1.68−1.63 (m,
5H), 1.61−1.57 (m, 4H), 1.53−1.48 (m, 2H), 1.05−0.98 (m, 2H),
0.95−0.89 (m, 2H). 13C NMR (100 MHz, CDCl3): δ 174.76, 164.18,
98.63, 70.25, 54.02, 50.16, 44.26, 41.31, 36.88, 35.41, 30.98, 8.29, 8.08.
C17H24N2O2 EI-MS: m/z (M + H+): 289.4 (calculated), 289.0
(found).
5-{[(5-Cyclopropyl-1,2-oxazol-3-yl)methyl]amino}adamantan-2-
1
one (10c). Yield: 80%. H NMR (400 MHz, CDCl3): δ 5.88 (s, 1H),
3.76 (s, 2H), 2.60−2.53 (m, 2H), 2.28−2.21 (m, 1H), 2.00−1.87 (m,
11H), 1.04−0.97 (m, 2H), 0.93−0.87 (m, 2H). 13C NMR (100 MHz,
CDCl3): δ 217.3, 175.03, 163.72, 98.46, 50.75, 46.64, 43.00, 41.69,
38.62, 37.42, 28.90, 8.67, 8.62. C17N22N2O2 EI-MS: m/z (M + H+):
287.4 (calculated), 287.0 (found).
3-{[(5-Cyclopropyl-1,3,4-oxadiazol-2-yl)methyl]amino}-
1
adamantan-1-ol (10d). Yield: 71%. H NMR (400 MHz, CDCl3): δ
3.97 (s, 2H), 2.32−2.23 (m, 2H), 2.16−2.05 (m, 1H), 1.68−1.61 (m,
6H), 1.60−1.55 (m, 4H), 1.53−1.47 (m, 2H), 1.13−1.11 (m, 2H),
1.11−1.09 (m, 2H). 13C NMR (100 MHz, CDCl3): δ 168.50, 164.95,
69.48, 53.78, 49.61, 43.95, 40.74, 35.99, 34.60, 30.55, 7.87, 5.70.
C16H23N3O2 EI-MS: m/z (M + H+): 290.4 (calculated), 290.0
(found).
3-{[(5-Cyclopropyl-1,2,4-oxadiazol-3-yl)methyl]amino}-
1
adamantan-1-ol (10e). Yield: 84%. H NMR (400 MHz, CDCl3 +
CD3OD): δ 4.15 (s, 2H), 2.35−2.28 (m, 2H), 2.22−2.12 (m, 1H),
1.96−1.90 (m, 2H), 1.89−1.78 (m, 4H), 1.70−1.61 (m, 4H), 1.54−
1.47 (m, 2H), 1.26−1.16 (m, 4H). 13C NMR (100 MHz, CDCl3 +
CD3OD): δ 183.64, 162.35, 68.08, 60.75, 45.82, 42.45, 36.75, 35.34,
33.85, 30.28, 10.53, 7.18. C16H23N3O2 EI-MS: m/z (M + H+): 290.4
(calculated), 290.0 (found).
N-[(5-Cyclopropyl-1,3,4-thiadiazol-2-yl)methyl]adamantan-1-
1
amine (10f). Yield: 78%. H NMR (400 MHz, CDCl3): δ 4.18 (s,
2H), 2.42−2.34 (m, 1H), 2.14−2.08 (m, 3H), 1.74−1.67 (m, 9H),
1.67−1.60 (m, 3H), 1.25−1.18 (m, 2H), 1.12−1.06 (m, 2H). 13C
NMR (100 MHz, CDCl3): δ 174.15, 42.42, 40.17, 36.45, 29.48, 11.65,
11.38. C16H23N3S EI-MS: m/z (M + H+): 290.4 (calculated), 290.0
(found).
3-{[(5-Cyclopropyl-1,3,4-thiadiazol-2-yl)methyl]amino}-
adamantan-1-ol (10g). Yield: 78%. 1H NMR (400 MHz, DMSO-d6):
δ 4.66 (s, 2H), 2.63−2.55 (m, 1H), 2.30−2.23 (m, 2H), 1.82−1.76 (m,
6H), 1.61−1.50 (m, 4H), 1.50−1.43 (m, 2H), 1.29−1.22 (m, 2H),
1.06−1.00 (m, 2H). 13C NMR (100 MHz, DMSO-d6): δ 175.97,
Combination Therapy of Compound 10b with Oseltamivir
Carboxylate. The synergistic antiviral effect of compound 10b and
oseltamivir carboxylate was evaluated in cell cultures as described
previously using CPE assay.29,30 Five combinations of compound 10b
J
DOI: 10.1021/acs.jmedchem.7b01536
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
and oseltamivir carboxylate, at the fixed EC50 ratios of 10:1, 5:1, 1:1,
1:5, and 1:10, were included. In each combination, a stock solution
with the designated EC50 ratio of compound 10b and oseltamivir
carboxylate was first made, then six 3-fold serial dilutions of the stock
solution were made and tested to obtain the dose−response curve,
based on which EC50 of individual 10b or oseltamivir carboxylate was
determined. Subsequently, fractional inhibitory concentration index
(FICI) was calculated using the following formula: FICI = ((EC50 of
10b in combination)/(EC50 of 10b alone)) + ((EC50 of oseltamivir
carboxylate in combination)/(EC50 of oseltamivir carboxylate alone)).
FICI < 0.5 was interpreted as a significant synergistic antiviral effect.
Microsome Stability. Test compounds were incubated at 37 °C
with liver microsomes (mouse or human) (Xenotech) at 1 μM drug
concentration in the presence of a NADPH regenerating system at 0.5
mg/mL microsomal protein. Testosterone (3A4 substrate), prop-
afenone (2D6), and diclofenac (2C9) were included as positive
controls. Compounds were incubated with microsomes in the
presence of a NADPH regenerating system (Sigma, catalogue no.
N0505, lot SLBH3107V). Aliquots at 0, 5, 10, 20, 30, and 60 min post
incubation were collected and immediately mixed with cold
acetonitrile containing internal standard (IS). Compounds incubated
with microsomes without NADPH regenerating system for 60 min
were also included. All the samples were analyzed by LC/MS/MS;
disappearance of test compound were assessed based on peak area
ratios of analyte/IS. Following equations were applied to calculate the
microsome clearance:
on the basolateral side), A is the surface area for the transport, i.e.,
0.0804 cm2 for the area of the monolayer, and C0 is the initial
concentration in the donor chamber (μM).
CYP Inhibition. Compound was incubated at 10 μM with human
liver microsomes with protein concentration at 0.253 mg/mL. α-
Naphthoflavone, sulfaphenazole, (+)-N-3-benzylnirvanol, quinidine,
and ketoconazole were used as positive controls. The mixture of test
compound or positive control were prewarmed with human liver
microsomes at 37 °C for 10 min. Phenacetin, diclofenac, S-
mephenytoin, dextromethorphan, and midazolam were mixed and
were used as substrates of CYP1A2, 2C9, 2C19, 2D6, and 3A4,
respectively. After addition of substrate mixture and cofactor NADPH,
the solution was incubated for another 10 min. The samples were then
quenched with stop solution containing internal standard. Next, the
solution was centrifuged at 4000 rpm (Centrifuge, 5810R Eppendorf)
at room temperature for 20 min. The supernatant was diluted with
purified water (v/v 2:1), which was shaken at 1000 rpm (Titer plate
shaker, Thermo) for 10 min before LC-MS/MS injection. The
formation of the metabolites acetaminophen, 4′-hydroxy diclofenac,
4′-hydroxy mephenytoin, dextrorphan, and 1′-hydroxy midazolam
were determined by LC-MS/MS and IC50 values were calculated.
SigmaPlot (V.11) was used to plot the mean CYP activity (% VC)
versus the test compound concentrations with nonlinear regression
analysis. IC50 values were reported as “>100 μM” when % inhibition at
highest concentration (100 μM) was less than 50%.
The IC50 values were extrapolated using the following equation:
IC50 = x(100 − %inhibition at x)/%inhibition
elimination rate constant (k) = −gradient;
0.693
suppose Hill slope = 1, where x = concentration of test compounds.
half life (T1/2) (min) =
k
ASSOCIATED CONTENT
CLint(mic) = 0.693/half-life/mg microsome protein per mL. Liver wt:
40, 30, 32, 20, and 88 g/kg for rat, monkey, dog, human, and mouse.
Using CLint(mic) to calculate the whole the liver clearance: mg
microsomal protein/g liver weight, 45 mg/g for five species. CLint(liver)
= CLint(mic) × mg microsomal protein/g liver weight × g liver weight/
kg body weight.
■
S
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.jmed-
chem.7b01536.
Caco-2 Cell Permeability Assay. Caco-2 cells (ATCC) were
cultured in MEM with 10% FBS. Cells at the passage number from 30
to 50 are seeded onto PET membranes of 96-well insert plates at 1 ×
105 cells/cm2 for 21−28 days to reach confluent cell monolayer. The
integrity of the monolayer was verified by performing Lucifer yellow
rejection assay. The quality of the monolayer were verified by
measuring the unidirectional (A → B) permeability of fenoterol (low
permeability marker), propranolol (high permeability marker), and
bidirectional permeability of digoxin (a P-glycoprotein substrate
marker) in duplicate wells. Assay conditions for test compounds
listed are as follows:
Molecular formula strings (CSV)
Docking model for 10b (Figure 5A) (PDB)
Docking model for 10b (Figure 5B) (PDB)
AUTHOR INFORMATION
■
Corresponding Author
*Phone: 520-626-1366. Fax: 520-626-0749. E-mail: junwang@
pharmacy.arizona.edu.
ORCID
• Test concentration: 2 μM (DMSO ≤ 1%)
• Replicates: n = 2
Jun Wang: 0000-0002-4845-4621
Author Contributions
• Directions: bidirectional transport including A → B and B → A
• Incubation time: single time point, 2 h
⊥Yuanxiang Wang and Yanmei Hu contributed equally to this
work
• Transport buffer: HBSS containing 10 mM HEPES, pH 7.4
• Incubation condition: 37 °C, 5% CO2, 95% relative humidity
Notes
Testing compounds were diluted in transport buffer and acetonitrile
(containing internal standard), and the resulting solution is designated
as the T0 sample. At the end of incubation, sample solutions from
both donor and receiver wells were immediately mixed with cold
acetonitrile containing internal standard (IS). All samples, including
T0 samples, donor samples, and receiver samples, were analyzed using
LC/MS/MS. Concentrations of test compound were expressed as
peak area ratio of analytes versus IS without a standard curve.
The apparent permeability coefficient Papp (cm/s) was calculated
using the equation:
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This research is supported by startup funding from the
University of Arizona and NIH grant AI119187 to J.W. R.M.
was supported by the NIH training grant T32 GM008804. We
thank Dr. David Bishop for proofreading and editing the
manuscript.
P
= (dCr/dt) × V /(A × C0)
ABBREVIATIONS USED
app
r
■
WT, wild type; DMEM, Dulbecco’s Modified Eagle Medium;
MDCK, Madin−Darby canine kidney; TEVC, two-electrode
voltage clamps
Where dCr/dt is the cumulative concentration of compound in the
receiver chamber as a function of time (μM/s), Vr is the solution
volume in the receiver chamber (0.075 mL on the apical side, 0.25 mL
K
DOI: 10.1021/acs.jmedchem.7b01536
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
the drug-resistant S31N mutant of the M2 ion channel of influenza A
virus. Proc. Natl. Acad. Sci. U. S. A. 2013, 110, 1315−1320.
(20) Kerns, E.; Di, L. Drug-like Properties: Concepts, Structure Design
and Methods: from ADME to Toxicity Optimization. Academic Press:
Boston, 2016.
(21) Wanka, L.; Iqbal, K.; Schreiner, P. The lipophilic bullet hits the
targets: medicinal chemistry of adamantane derivatives. Chem. Rev.
2013, 113, 3516−3604.
(22) Kalgutkar, A. S.; Nguyen, H. T.; Vaz, A. D.; Doan, A.; Dalvie, D.
K.; McLeod, D. G.; Murray, J. C. In vitro metabolism studies on the
isoxazole ring scission in the anti-inflammatory agent lefluonomide to
its active alpha-cyanoenol metabolite A771726: mechanistic similarities
with the cytochrome P450-catalyzed dehydration of aldoximes. Drug
Metab. Dispos. 2003, 31, 1240−1250.
(23) Gramec, D.; Peterlin Masic, L.; Sollner Dolenc, M. Bioactivation
potential of thiophene-containing drugs. Chem. Res. Toxicol. 2014, 27,
1344−1358.
(24) White, J. R. Dipeptidyl peptidase-IV inhibitors: pharmacological
profile and clinical use. Clin Diabetes 2008, 26, 53−57.
(25) Repetto, G.; del Peso, A.; Zurita, J. L. Neutral red uptake assay
for the estimation of cell viability/cytotoxicity. Nat. Protoc. 2008, 3,
1125−1131.
(26) Hayden, F. Newer influenza antivirals, biotherapeutics and
combinations. Influenza Other Respir. Viruses 2013, 7, 63−75.
(27) Huggins, J. W.; Robins, R. K.; Canonico, P. G. Synergistic
antiviral effects of ribavirin and the C-nucleoside analogs tiazofurin and
selenazofurin against togaviruses, bunyaviruses, and arenaviruses.
Antimicrob. Agents Chemother. 1984, 26, 476−480.
REFERENCES
■
(1) Webster, R. G.; Monto, A. S.; Braciale, T. J.; Lamb, R. A.
Textbook of Influenza; John Wiley & Sons, Ltd: West Sussex, UK,
2013.
(2) Zhang, W. Q.; Webster, R. G. Can we beat influenza? Science
2017, 357, 111−111.
(3) Zhu, H.; Lam, T. T.; Smith, D. K.; Guan, Y. Emergence and
development of H7N9 influenza viruses in China. Curr. Opin. Virol.
2016, 16, 106−113.
(4) Lai, S.; Qin, Y.; Cowling, B. J.; Ren, X.; Wardrop, N. A.; Gilbert,
M.; Tsang, T. K.; Wu, P.; Feng, L.; Jiang, H.; Peng, Z.; Zheng, J.; Liao,
Q.; Li, S.; Horby, P. W.; Farrar, J. J.; Gao, G. F.; Tatem, A. J.; Yu, H.
Global epidemiology of avian influenza A H5N1 virus infection in
humans, 1997−2015: a systematic review of individual case data.
Lancet Infect. Dis. 2016, 16, e108−118.
(5) People at High Risk of Developing Flu−Related Complications;
Centers for Disease Control and Prevention: Atlanta, GA, 2018;
https://www.cdc.gov/flu/about/disease/high_risk.htm (accessed on
November 20, 2017.
(6) Shaw, M. L. The next wave of influenza drugs. ACS Infect. Dis.
2017, 3, 691−694.
(7) Wang, J.; Li, F.; Ma, C. Recent progress in designing inhibitors
that target the drug-resistant M2 proton channels from the influenza A
viruses. Biopolymers 2015, 104, 291−309.
(8) Loregian, A.; Mercorelli, B.; Nannetti, G.; Compagnin, C.; Palu,
G. Antiviral strategies against influenza virus: towards new therapeutic
approaches. Cell. Mol. Life Sci. 2014, 71, 3659−3683.
(9) Wang, J.; Qiu, J. X.; Soto, C.; DeGrado, W. F. Structural and
dynamic mechanisms for the function and inhibition of the M2 proton
channel from influenza A virus. Curr. Opin. Struct. Biol. 2011, 21, 68−
80.
(28) Meletiadis, J.; Pournaras, S.; Roilides, E.; Walsh, T. J. Defining
fractional inhibitory concentration index cutoffs for additive
interactions based on self-drug additive combinations, Monte Carlo
simulation analysis, and in vitro-in vivo correlation data for antifungal
drug combinations against Aspergillus fumigatus. Antimicrob. Agents
Chemother. 2010, 54, 602−609.
(29) Hu, Y.; Zhang, J.; Musharrafieh, R. G.; Ma, C.; Hau, R.; Wang, J.
Discovery of dapivirine, a nonnucleoside HIV-1 reverse transcriptase
inhibitor, as a broad-spectrum antiviral against both influenza A and B
viruses. Antiviral Res. 2017, 145, 103−113.
(30) Hu, Y.; Zhang, J.; Musharrafieh, R.; Hau, R.; Ma, C.; Wang, J.
Chemical genomics approach leads to the identification of hesperadin,
an aurora B kinase inhibitor, as a broad-spectrum influenza antiviral.
Int. J. Mol. Sci. 2017, 18, 1929.
(31) DiMasi, J. A.; Grabowski, H. G.; Hansen, R. W. Innovation in
the pharmaceutical industry: New estimates of R&D costs. J. Health
Econ 2016, 47, 20−33.
(32) Wang, J.; Ma, C.; Wang, J.; Jo, H.; Canturk, B.; Fiorin, G.; Pinto,
L. H.; Lamb, R. A.; Klein, M. L.; DeGrado, W. F. Discovery of novel
dual inhibitors of the wild-type and the most prevalent drug-resistant
mutant, S31N, of the M2 proton channel from influenza A virus. J.
Med. Chem. 2013, 56, 2804−2812.
(33) Balannik, V.; Wang, J.; Ohigashi, Y.; Jing, X.; Magavern, E.;
Lamb, R. A.; DeGrado, W. F.; Pinto, L. H. Design and pharmacological
characterization of inhibitors of amantadine-resistant mutants of the
M2 ion channel of influenza A virus. Biochemistry 2009, 48, 11872−
11882.
(34) Hatakeyama, S.; Sakai-Tagawa, Y.; Kiso, M.; Goto, H.;
Kawakami, C.; Mitamura, K.; Sugaya, N.; Suzuki, Y.; Kawaoka, Y.
Enhanced expression of an α2,6-linked sialic acid on MDCK cells
improves isolation of human influenza viruses and evaluation of their
sensitivity to a neuraminidase inhibitor. J. Clin Microbiol 2005, 43,
4139−4146.
(10) Gubareva, L. V.; Besselaar, T. G.; Daniels, R. S.; Fry, A.;
Gregory, V.; Huang, W.; Hurt, A. C.; Jorquera, P. A.; Lackenby, A.;
Leang, S. K.; Lo, J.; Pereyaslov, D.; Rebelo-de-Andrade, H.; Siqueira,
M. M.; Takashita, E.; Odagiri, T.; Wang, D.; Zhang, W.; Meijer, A.
Global update on the susceptibility of human influenza viruses to
neuraminidase inhibitors, 2015−2016. Antiviral Res. 2017, 146, 12−20.
(11) Webster, R. G.; Govorkova, E. A. Continuing challenges in
influenza. Ann. N. Y. Acad. Sci. 2014, 1323, 115−139.
(12) Hurt, A. C. The epidemiology and spread of drug resistant
human influenza viruses. Curr. Opin. Virol. 2014, 8, 22−29.
(13) Matsuzaki, Y.; Mizuta, K.; Aoki, Y.; Suto, A.; Abiko, C.; Sanjoh,
K.; Sugawara, K.; Takashita, E.; Itagaki, T.; Katsushima, Y.; Ujike, M.;
Obuchi, M.; Odagiri, T.; Tashiro, M. A two-year survey of the
oseltamivir-resistant influenza A(H1N1) virus in Yamagata, Japan and
the clinical effectiveness of oseltamivir and zanamivir. Virol. J. 2010, 7,
53.
(14) Dong, G.; Peng, C.; Luo, J.; Wang, C.; Han, L.; Wu, B.; Ji, G.;
He, H. Adamantane-resistant influenza a viruses in the world (1902−
2013): frequency and distribution of M2 gene mutations. PLoS One
2015, 10, e0119115.
(15) Li, F.; Hu, Y.; Wang, Y.; Ma, C.; Wang, J. Expeditious lead
optimization of isoxazole-containing influenza A virus M2-S31N
inhibitors using the suzuki-miyaura cross-coupling reaction. J. Med.
Chem. 2017, 60, 1580−1590.
(16) Hu, Y.; Wang, Y.; Li, F.; Ma, C.; Wang, J. Design and
expeditious synthesis of organosilanes as potent antivirals targeting
multidrug-resistant influenza A viruses. Eur. J. Med. Chem. 2017, 135,
70−76.
(17) Li, F.; Ma, C.; Hu, Y.; Wang, Y.; Wang, J. Discovery of potent
antivirals against amantadine-resistant influenza A viruses by targeting
the M2-S31N proton channel. ACS Infect. Dis. 2016, 2, 726−733.
(18) Li, F.; Ma, C.; DeGrado, W. F.; Wang, J. Discovery of highly
potent inhibitors targeting the predominant drug-resistant S31N
mutant of the influenza A virus M2 proton channel. J. Med. Chem.
2016, 59, 1207−1216.
(19) Wang, J.; Wu, Y.; Ma, C.; Fiorin, G.; Wang, J.; Pinto, L. H.;
Lamb, R. A.; Klein, M. L.; Degrado, W. F. Structure and inhibition of
L
DOI: 10.1021/acs.jmedchem.7b01536
J. Med. Chem. XXXX, XXX, XXX−XXX