Design and synthesis of HIV-1 protease inhibitors for a long-acting injectable drug application

Article abstract of DOI:10.1016/j.bmcl.2012.10.095

The design and synthesis of novel HIV-1 protease inhibitors (PIs) (1-22), which display high potency against HIV-1 wild-type and multi-PI-resistant HIV-mutant clinical isolates, is described. Lead optimization was initiated from compound 1, a Phe-Phe hydroxyethylene peptidomimetic PI, and was directed towards the discovery of new PIs suitable for a long-acting (LA) injectable drug application. Introducing a heterocyclic 6-methoxy-3-pyridinyl or a 6-(dimethylamino)-3-pyridinyl moiety (R³) at the para-position of the P1′ benzyl fragment generated compounds with antiviral potency in the low single digit nanomolar range. Halogenation or alkylation of the metabolic hot spots on the various aromatic rings resulted in PIs with high stability against degradation in human liver microsomes and low plasma clearance in rats. Replacing the chromanolamine moiety (R¹) in the P2 protease binding site by a cyclopentanolamine or a cyclohexanolamine derivative provided a series of high clearance PIs (16-22) with EC₅₀s on wild-type HIV-1 in the range of 0.8-1.8 nM. PIs 18 and 22, formulated as nanosuspensions, showed gradual but sustained and complete release from the injection site over two months in rats, and were therefore identified as interesting candidates for a LA injectable drug application for treating HIV/AIDS.

Full text of DOI:10.1016/j.bmcl.2012.10.095

Bioorganic & Medicinal Chemistry Letters 23 (2013) 310–317
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Design and synthesis of HIV-1 protease inhibitors for a long-acting
injectable drug application
Bart Kesteleyn a, , Katie Amssoms a, Wim Schepens a, Geerwin Hache a, Wim Verschueren a,
⇑
Wim Van De Vreken a, Klara Rombauts a, Greet Meurs a, Patrick Sterkens a, Bart Stoops a, Lieven Baert a,
Nigel Austin a, Jörg Wegner a, Chantal Masungi a, Inge Dierynck a, Stina Lundgren b, Daniel Jönsson b,
Kevin Parkes b, Genadiy Kalayanov b, Hans Wallberg b, Åsa Rosenquist b, Bertil Samuelsson b,
Kristof Van Emelen a, Jan Willem Thuring a
a Janssen Infectious Diseases, A Division of Janssen Pharmaceutical Companies of Johnson & Johnson, Turnhoutseweg 30, 2340 Beerse, Belgium
b Medivir AB, P.O. Box 1086, SE-141 22 Huddinge, Sweden
a r t i c l e i n f o
a b s t r a c t
Article history:
The design and synthesis of novel HIV-1 protease inhibitors (PIs) (1–22), which display high potency
against HIV-1 wild-type and multi-PI-resistant HIV-mutant clinical isolates, is described. Lead optimiza-
tion was initiated from compound 1, a Phe–Phe hydroxyethylene peptidomimetic PI, and was directed
towards the discovery of new PIs suitable for a long-acting (LA) injectable drug application. Introducing
a heterocyclic 6-methoxy-3-pyridinyl or a 6-(dimethylamino)-3-pyridinyl moiety (R3) at the para-posi-
tion of the P10 benzyl fragment generated compounds with antiviral potency in the low single digit nano-
molar range. Halogenation or alkylation of the metabolic hot spots on the various aromatic rings resulted
in PIs with high stability against degradation in human liver microsomes and low plasma clearance in
rats. Replacing the chromanolamine moiety (R1) in the P2 protease binding site by a cyclopentanolamine
or a cyclohexanolamine derivative provided a series of high clearance PIs (16–22) with EC50s on wild-
type HIV-1 in the range of 0.8–1.8 nM. PIs 18 and 22, formulated as nanosuspensions, showed gradual
but sustained and complete release from the injection site over two months in rats, and were therefore
identified as interesting candidates for a LA injectable drug application for treating HIV/AIDS.
Ó 2012 Elsevier Ltd. All rights reserved.
Received 26 September 2012
Revised 18 October 2012
Accepted 22 October 2012
Available online 1 November 2012
Keywords:
HIV/AIDS
HIV Protease
Protease inhibitor
Long-acting formulation
Nanosuspension
Since the introduction of antiretroviral (ARV) combination ther-
apy, HIV/AIDS has become a manageable but chronic condition
that requires lifelong treatment and compliance with the pre-
scribed therapy.1 Non-compliance may lead to drug plasma con-
centrations that are insufficient to suppress viral replication and
the build-up of viral resistance, and this is one of the major factors
contributing to treatment failure.2–4
suitable matrix from which the active ingredient is slowly released
(e.g., RisperidalÒ ConstaÒ). Nanosuspensions, (i.e., suspension for-
mulations of drug particles in the nanosize range) allow the inte-
gration of poorly water- and oil-soluble drug substances in
injectable LA formulations. Nanosizing may improve the dissolu-
tion rate of poorly soluble compounds and the release rate of the
drug substance from the depot formulation.6
Long-acting (LA) formulations have recently been introduced as
an alternative to the daily oral regimens in HIV treatment.5 LA
intramuscular depot formulations generate sustained minimal
but effective inhibitory concentrations of antiviral drug in the
blood over a period of time, and offer the advantage of improved
adherence to ARV therapy. Moreover, LA formulations have the po-
tential to reduce adverse effects over time and may serve as a prac-
tical means of pre-exposure prophylaxis.
The HIV-1 protease enzyme is an attractive antiretroviral drug
target for the treatment of HIV/AIDS, as it plays an essential role
in the viral assembly and maturation process required for the pro-
duction of infectious virus particles.7 With 10 protease inhibitors
(PIs) approved for clinical use by the FDA since 1995, these com-
pounds have an established role as a major component in ARV
combination therapy.8 HIV-1 PIs typically occur as peptidomimetic
competitive inhibitors of HIV-1 aspartic protease.9 Examples of
marketed HIV-1 PIs include lopinavir, atazanavir and darunavir
(Fig. 1).
Typically, LA formulations consist of an active ingredient that is
suspended in an oil-based solution or suspension (e.g., haloperidol
depot), a microfine watery suspension (e.g., steroid depot), or a
The drug properties required for a new HIV-1 PI for a LA drug
application differ substantially from those for an oral application.
Slow release of the drug from the depot formulation is expected
to provide sustained but low plasma concentrations. At these low
⇑
Corresponding author. Tel.: +32 14641652; fax: +32 14605737.
E-mail address: bkestele@its.jnj.com (B. Kesteleyn).
0960-894X/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmcl.2012.10.095

B. Kesteleyn et al. / Bioorg. Med. Chem. Lett. 23 (2013) 310–317
311
N
O
O
O
OH
O
H
N
H
N
H
O
N
N
O
HN
N
N
H
O
N
H
N
H
O
OH
O
O
lopinavir
atazanavir
NH2
O
O
O
O
S
O
O
N
H
N
OH
darunavir
Figure 1. Structures of the marketed HIV-1 protease inhibitors lopinavir, atazanavir and darunavir.
concentrations, the drug must still suppress the viral replication of
HIV-1 wild-type and mutant viruses. It is therefore believed that
important requirements for a LA application include: (1) ultra-high
potency against wild-type HIV, (2) high potency against multi-
PI-resistant HIV-mutant clinical isolates, and (3) low functional
plasma protein binding (PPB), as determined by the shift in EC50
value in the presence of 50% human serum in the antiviral assay.
Also, with respect to the pharmacokinetic properties of the drug,
(4) a high stability against microsomal degradation, resulting in
low plasma clearance, might increase circulating drug levels in
the plasma. Finally, (5) a low aqueous solubility of the drug is nec-
essary, to limit the dissolution rate of the drug from the depot
formulation.
Compound 1 is an example of the so-called Phe-Phe hydroxy-
ethylene peptidomimetic scaffold for HIV-1 PIs, decorated with a
chromanolamine moiety in the P2 position and a t-butylglycine
fragment, similar to atazanavir, in P20.10,11 The history of this scaf-
fold goes back to the early days of HIV-1 PI research at Merck and
the former Ciba-Geigy, which led to several preclinical and clinical
candidates that have been evaluated up to Phase
I clinical
trials.12–16 Poor oral bioavailability seems to have hampered the
development of this class of inhibitors.
As shown in Table 1, compound 1 was found to be a potent
inhibitor of wild-type HIV-1, with an EC50 of 10 nM. Remarkably,
it also displayed equal potency against a panel of mutant isolates
(M1, M2 and M3) and in the presence of 50% human serum. The
multi-PI-resistant mutants M1–3 contain combinations of known
PI primary resistance-associated mutations (RAMs) reported by
IAS,17 and each induces—to a certain extent—resistance against
the marketed PIs lopinavir, atazanavir and darunavir, as shown in
Table 1. Hence, these data indicate a low sensitivity of compound 1
to reported primary RAMs for PIs combined with low binding to
plasma proteins; characteristics that match our postulated drug
requirements (1), (2) and (3) for a LA PI drug application. The turn-
over of compound 1 by human liver microsomes (HLM) was fast
(99% metabolized by HLM in 15 min, see Table 2). Also, in vivo in
rats, compound 1 was rapidly cleared from the plasma following
intravenous dosing (Clrat = 4.0 L/h/kg, see Table 2).
In a lead-finding exercise, compound 1 was identified as an
interesting PI scaffold to initiate a lead optimization program for
a LA HIV-1 PI application (Fig. 2).
P2
P1
P1'
P2'
O
O
O
O
H
N
O
N
H
N
H
OH
OH
1
Detailed analysis of the metabolites generated by HLM degrada-
tion of compound 1 indicated that the aromatic moieties in posi-
tions P1 and P10, and the chromanolamine moiety in position P2,
Figure 2. Structure of lead compound 1 with indication of the HIV-protease binding
sites (P-sites).
Table 1
Antiviral activity against wild-type HIV-1IIIB and HIV-mutant clinical isolates M1–3
a
b
c
c
c
Compound
HIV-1IIIB EC50, nM
HIV-1IIIB + 50%HS EC50, nM
M1 EC50, nM (FC)
M2 EC50, nM (FC)
M3 EC50, nM (FC)
Lopinavir
Atazanavir
Darunavir
1
11
15
18
22
13
63
13
17
19
6
6.8
2
3.7
110 (9)
57 (8)
12 (2)
310 (23)
3.2 (<1)
51 (8)
6 (<1)
0.9 (<1)
3.3 (2)
32 (3)
61 (9)
6.8
6.6
10
1.7
1.8
0.8
1.1
2.4 (<1)
7 (<1)
0.8 (<1)
1.5 (<1)
0.5 (<1)
0.4 (<1)
19 (2)
1.3 (<1)
5.4 (3)
0.8 (<1)
0.9 (<1)
0.3 (<1)
0.6 (<1)
a
The EC50 values were determined using MT4 cells. All EC50 values are reported as averages of at least two independent experiments (typically three to four) in which each
assay was conducted in quadruplicate. Toxicity was measured in MT4 cells. CC50 values were >10 M for all compounds.
l
b
The antiviral EC50 value for HIV-1, measured in the presence of 50% human serum (50% HS) was used to determine functional plasma protein binding.
Mutant virus isolates M1, M2 and M3 contain the following PI resistance-associated mutations (bold figures represent primary resistance mutations, according to IAS17),
c
M1: L10I, M46I, I64V, I84V, L90M, I93L; M2: L10I, I13V, M46I, I50V, L63P, L76V; M3: L10I, K20R, M36I, G48V, I62V, A71V, V82A, I93L. The numbers in parentheses represent
the fold change (FC) in EC50 value for each mutant virus relative to wild-type HIV-1.

312
B. Kesteleyn et al. / Bioorg. Med. Chem. Lett. 23 (2013) 310–317
are metabolic hot spots (data not shown). Based on these observa-
tions, part of our lead optimization strategy was aimed at blocking
all metabolic sites, mainly by halogenation at various positions on
the aromatic moieties in the molecule. In addition, to improve the
potency of the lead compound 1, we envisioned introducing a het-
erocyclic aromatic moiety R3 at the para-position of the P10 benzyl
fragment, following the example of atazanavir.18
moiety in P2. Also, replacing the R3-fragment MP by a DMAP group
(compound 19) did not improve any drug properties with respect
to those of compound 18.
In conclusion, our SAR exploration of the Phe–Phe hydroxyeth-
ylene peptidomimetic PI scaffold afforded two compound series
with distinct compound properties. Compounds 11 and 15 are
the most potent (antiviral activity on HIV-1IIIB in the presence of
50%HS) ‘low clearance’ PIs. Compounds 18 and 22, on the other
hand, represent the most potent ‘high clearance’ PIs. These four
compounds, with the R1-substituent as the major structural differ-
entiator, were taken forward in a comparison of their resistance
profiles and LA pharmacokinetic properties.
All new HIV-1 PIs (1–22) were synthesized according to the
reaction sequence outlined in Scheme 1. The synthesis of com-
pound 1 was started from lactone 28a (R = H). For compounds
2–22, the introduction of an aromatic moiety (R3), typically via a
Suzuki coupling reaction, required the presence of a bromo-substi-
tuent at the para-position of the P1 benzyl group of the lactone,
and thus required the synthesis of the halogenated lactone 28b
(R = Br). The synthesis of lactone 28a (R = H) has previously been
described elsewhere.19,20 For the synthesis of the bromo-lactone
28b (R = Br), a new enantioselective methodology was developed.
As shown in Scheme 2, the synthesis started from commercially
available N-Boc-protected (S)-4-bromophenylalanine 23. After
conversion of the carboxylic acid group to the Weinreb amide 24,
substitution with the Grignard reagent 3-butenylmagnesium bro-
mide provided ketone 25. Oxidation of the terminal double bond
As outlined in Table 2, following the proposed lead optimization
strategy to block metabolic hot spots, the benzyl moiety was dec-
orated with an ortho-F, ortho-Cl or meta-OCF3 substituent (R2) in all
new PIs. The introduction of a heterocyclic aromatic group (R3) al-
lowed for a substantial increase of antiviral potency, as exemplified
by compounds 2 and 3, which hold a methoxypyridine (MP) or
dimethylaminopyridine (DMAP) moiety, respectively. These MP
and DMAP fragments were selected from an extensive screen of al-
kyl-, alkoxy- and dialkylamino-substituted pyridines, pyrimidines,
pyrazines and thiazoles as R3 moieties. The MP and DMAP frag-
ments provided the best overall compound profiles in terms of
antiviral potency, effect on functional plasma protein binding,
and microsomal stability. However, introducing a fluorine atom
in P1 and a heterocyclic fragment R3 induced only a minor reduc-
tion in metabolic turnover by HLM for compounds 2 and 3. Block-
ing the remaining metabolic hot spot at the chromanolamine
fragment, by introducing a para-chloro atom in compound 4, was
required to obtain a serious reduction of microsomal turnover
and a reduction of plasma clearance in rats. The microsomal turn-
over by HLM could be reduced to virtually zero by replacing the
ortho-F atom in P1 by an ortho-Cl (compound 5) or meta-OCF3
(compound 6) substituent, but no further reduction in plasma
clearance in rats was observed for these compounds. Also, these
P1 modifications did not benefit the antiviral potency of com-
pounds 5 and 6, particularly when measured in the presence of
50%HS.
with catalytic RuCl3.3H2O and excess NaIO4 afforded the c-keto-
carboxylic acid 26. After conversion of compound 26 to the corre-
sponding methyl ester 27 with methyl iodide, and using potassium
bicarbonate as a mild base, reductive cyclization with N-selectride
in THF at low temperature afforded the desired bromo-lactone 28b
in high enantiomeric excess (ee >95%).
In a further effort to fine-tune the compound properties, a series
of chromanolamine derivatives with fluoro-, chloro- and methyl-
substituents at positions 6 and 8 (R00 and R0 respectively, in com-
pounds 7–15) were synthesized. As an overall trend, blocking the
R00-position with a halogen was found to be beneficial for metabolic
stability and to provide lower plasma clearance in rats. The combi-
nation of a fluorine atom at R00 and a chloro- or methyl-substituent
at R0 (compounds 11 and 15, respectively), provided the best over-
all profile; this included potent antiviral activity (also in the pres-
ence of 50%HS) and low functional plasma clearance in rats. In the
case of compound 11, the presence of a DMAP substituent (R3) was
found to be more favorable than MP (compound 10), in terms of
antiviral potency and plasma clearance in rats.
As mentioned above, for a LA drug application, the plasma con-
centrations that can be obtained by slow sustained release from a
nanosuspension depot formulation are expected to be very low in
practice. It is crucially important that the drug is able to suppress
the virus (wild-type HIV-1, as well as the mutant clinical isolates)
at these concentrations. Ultra-potent (preferably sub-nanomolar)
inhibitors of wild-type HIV-1, with good coverage of resistant mu-
tants, are therefore required.
To this end, we noticed that replacing the chromanolamine moi-
ety in P2 by small lipophilic cyclopentanolamines or cyclohexanol-
amines provided a series of compounds (16–22) with potent
antiviral activity against wild-type HIV-1 (EC50 range of 0.8 –
1.1 nM). The effect was most pronounced for compounds 18 and
22, which have a (R)-3-methyl-cyclohexanolamine and a thiophene-
cyclohexanolamine moiety, respectively, in the P2 position. How-
ever, all compounds in this series had high microsomal turnover
and were rapidly cleared from the plasma in rats. Replacing the
ortho-F R2-substituent by ortho-Cl (compound 20), or meta-OCF3
(compound 21), did not reduce the microsomal turnover, as was
the case for compounds 5 and 6 with the 6-chloro-chromanolamine
Alkylation of lactones 28a or 28b with substituted benzyl bro-
mides (29), using lithium hexamethyldisilazane as a base at low
temperature, afforded trans-alkylation products 30.21 Hydrolysis
of compound 30 with NaOH or LiOH in methanol-water, and subse-
quent quenching by acidification to pH = 3 with citric acid, provided
the carboxylic acid derivative 31. Spontaneous re-lactonization of
compound 31 was prevented by silylation of the alcohol group with
TBDMSCl to intermediate 32. In order to create a diversity of exam-
ples of PIs 1–22, the introduction of the P2 amine and the R3 substit-
uents was accomplished by the HATU-mediated coupling of amines
33 and subsequent palladium-assisted Suzuki coupling of boronic
acids 35 (except for compound 1). Finally, N-Boc deprotection
under acid conditions, or using a combination of TMSCl and NaI
in acetonitrile, was followed by HATU-mediated coupling of N-
(methoxycarbonyl)-L-tert-leucine (38), to provide the desired
HIV-1 PIs 1-22 in high overall yields and enantioselective purity
(ee >95%). In some cases it proved advantageous to postpone depro-
tection of the silyl-protecting group until after the Suzuki-coupling
reaction, to avoid degradation by the elimination of water. This deg-
radation reaction was observed particularly in the case of Suzuki-
coupling reactions that required longer reaction times. Also, the
introduction of the R3 aromatic moiety can be performed as early
as the first step in the reaction scheme—at the stage of compound
28b—or be postponed until the final step in the reaction scheme,
using virtually the same Suzuki-coupling reaction conditions as de-
scribed in the general scheme.
All halogenated and alkylated chromanolamine P2 derivatives
45b–j were prepared by adaptation of a procedure formerly
reported for the synthesis of chromanolamine 45a.22 As shown in
Scheme 3, ortho- and/or para-substituted phenols 39 were
converted to the corresponding chromanon derivatives 41 by
O-alkylation with 3-bromopropionic acid and subsequent
Friedel–Crafts cyclization of intermediates 40. In several cases, a

B. Kesteleyn et al. / Bioorg. Med. Chem. Lett. 23 (2013) 310–317
313
Table 2
SAR overview for HIV-1 protease inhibitors 1–22
P1'
R3
R3-fragments
P2
O
O
O
N
H
N
N
N
R1
O
N
H
N
H
OH
O
R2
DMAP
MP
P2'
P1
R1-amine
R0
R00
R2
R3
H
HIV-1IIIB EC50, nM
HIV-1IIIB + 50%HS EC50, nM
HLMc %Metab.
Clrat L/h/Kg
a
b
d
Compound
1
2
H
o–F
10
2.5
19
6.6
99
72
4.0
5.8
MP
3
o–F
DMAP
1.5
10
90
9.1
O
N
H
OH
OH
4
5
Cl
o–F
o–Cl
MP
MP
2.7
3.9
14
13
23
5
0.47
1.4
6
m-OCF3
MP
4.2
24
0
-
O
N
H
7
8
R'
R''
Cl
H
H
F
o–F
o–F
MP
MP
2.4
1.2
5.3
5.8
47
48
3.9
2.1
9
F
H
F
F
Cl
Me
Me
F
o–F
o–F
o–F
o–F
o–F
o–F
o–F
MP
MP
DMAP
MP
MP
2.8
2.6
1.7
4.1
0.9
2.8
1.8
6.9
8.4
6
98
8
5.3
1.9
0.8
—
2.8
5
O
10
11
12
13
14
15
Cl
Cl
F
H
F
22
23
63
48
17
N
H
16
OH
7.7
9.8
6.8
MP
MP
Me
1.6
16
17
o–F
o–F
MP
MP
1.1
1.8
3.3
5.9
99
99
4.7
5.2
N
HO
H
N
H
OH
18
19
o–F
o–F
MP
DMAP
0.8
1.1
2.0
4.9
99
99
5.4
5.6
20
21
o–Cl
m–OCF3
MP
MP
1.1
0.9
7.1
3.2
99
94
4.2
5.4
N
H
OH
S
22
o–F
MP
1.1
3.7
91
4.8
N
H
OH
a
The EC50 values were determined using MT4 cells. All EC50 values are reported as averages of at least two independent experiments (typically three to four), in which each
assay was conducted in quadruplicate. Toxicity was measured in MT4 cells. CC50 values were >10 M for all compounds.
The antiviral EC50 value for HIV-1, measured in the presence of 50% human serum (50% HS), was used to determine functional plasma protein binding.
l
b
c
Metabolic turnover by HLM (human liver microsomes) is represented as the %compound metabolized at a concentration of 1
Plasma clearance in rats (Clrat) was determined after intravenous dosing of 2 mg/kg compound in a suitable formulation, typically 1 mg/mL in a mixture of PEG400/
l
M after 15 min, incubated at 37 °C.
d
ethanol/saline at a ratio of 70/20/10. Reported clearance figures are averages taken from two or three animals. Clearance from plasma in rats higher than 70% of the liver blood
flow (Clrat > 3 L/h/kg) is ranked as ‘high’. Compound extraction from plasma in rats below 30% of the liver blood flow (Clrat < 1.3 L/h/kg) is ranked as ‘low’.
two-step alternative procedure was used to prepare intermediates
40: NaH-mediated alkylation with 3-chloropropanol, followed by
oxidation of the primary alcohol with TEMPO and bis(acet-
oxy)iodobenzene (BAIB), providing higher overall yields of inter-
mediates 40. Reaction of chromanons 41 with bromine in
dichloromethane afforded a mixture of the mono- and di-bromi-
nated reaction products 42 and 43. Upon treatment of this mixture
with sodium sulfite in acetic acid at 70 °C, the di-brominated inter-
mediates 42 were selectively reduced to the mono-brominated
intermediates 43. After reduction with sodium borohydride, bro-
mohydrines 44 were submitted to a Ritter reaction using sulfuric
acid in acetonitrile to provide, after hydrolysis of the cyclic inter-
mediate oxazolidines, the cis-substituted chromanolamines 45 as
racemic mixtures. The enantiomers 45 were separated via chiral
SFC. The absolute stereochemistry was assigned by vibrational cir-
cular dichroism (VCD), by comparing the recorded VCD spectra of
both enantiomers with those obtained by computational ab initio
calculation.23
(1R,2S,3R)-2-amino-3-methylcyclopentanol, the R1-fragment of
compound 16, was prepared as previously described elsewhere.24

314
B. Kesteleyn et al. / Bioorg. Med. Chem. Lett. 23 (2013) 310–317
R
1. LHMDS
2.
R
R
Br
29
1. NaOH or LiOH
CH3OH / H2O, rt
O
R2
THF, - 78°C
HO
NHBoc
NHBoc
OH
NHBoc
2. citric acid
O
O
O
R2
O
31
30
28a (R = H)
28b (R = Br)
R2
Br
R
1. TBDMSCl
R3B(OH)2 (35)
O
O
1. HATU, Et3N, rt
R1
N
Imidazole
DMF, rt
R1NH2 (33)
Pd(PPh3)4, Na2CO3
NHBoc
HO
NHBoc
OTBDMS
H
OH
dioxane / H2O, 110°C
2. TBAF, 50°C
DMF
2. MeOH
R2
R2
34
32
R3
R3
R3
O
H
N
O
HO
O
O
O
O
O
HCl, iPrOH, rt
H
N
R1
38
R1
R1
O
N
H
NH2
N
H
N
H
N
NHBoc
H
or: TFA, DCM, rt
or: TMSCl, NaI
CH3CN, rt
OH
OH
O
OH
HATU, Et3N
DMF, rt
R2
R2
1- 22
R2
37
36
Scheme 1. Synthesis of protease inhibitors 1–22.
Br
Br
Br
Br
NH
.HCl
Mg, I2
O
N
THF, 0°C -> rt
OH
BocHN
HATU
Et3N, DCM, rt
BocHN
O
BocHN
23
O
O
O
25
24
RuCl3.3H2O, NaIO4
acetone / H2O, rt
Br
Br
O
Br
N-selectride
MeI, KHCO3
DMF, rt
O
THF, - 65°C
BocHN
BocHN
BocHN
O
OH
O
O
O
O
26
27
28b
Scheme 2. Synthesis of bromo-lactone (28b).
Enantioselective synthesis of the chiral P2 methylcyclohexanolamine
fragment 51 was found to be rather challenging, especially with re-
gard to stereo control of the amino and alcohol substituents on the
cyclohexane ring. Several reaction steps in the sequence provided
mixtures that required tedious chromatographic separation. As
shown in Scheme 4, the synthesis of compound 51 was started from
commercially available 3R-methylcyclohexanone (46). Acid-
catalyzed O-acetylation of compound 46 with isopropenyl acetate
afforded a non-separable mixture of compounds 47 and 48 in a ratio
of 10:7. Treatment of the mixture with concentrated nitric acid in
acetic anhydride afforded a mixture of all four possible nitrated reac-
tion products. The desired 2S-nitro-3R-methylcyclohexanone (49),
the major component in the mixture, was separated by chromatogra-
phy over silica gel, and isolated in an overall yield of 11% from
compound 46. Reduction of the carbonyl group of compound 49 with
sodium borohydride afforded a 1:1 epimeric mixture of alcohols, and
required chromatographic separation to provide the desired cis-
nitro-alcohol intermediate 50 in a yield of 22%. Finally, after reductive

B. Kesteleyn et al. / Bioorg. Med. Chem. Lett. 23 (2013) 310–317
315
O
O
O
O
1. (CO)2Cl2, 0°C
R"
R"
R"
Br
OH
2. AlCl3, rt
OH
NaOH, H2O, reflux
OH
DCM
O
R'
39
R'
R'
1. NaH, DMF, 60°C
40
41
Cl
OH
2. BAIB, TEMPO
CH3CN, H2O, rt
Br2
DCM, reflux
OH
O
O
O
R"
Br
Br
Br
R"
Br
R"
NaBH4
MeOH, rt
+
O
O
R'
R'
R'
43
44
42
Na2SO3
H2SO4
CH3CN, 45°C
HOAc, 70°C
NH2
O
NH2
N
R"
OH
R"
OH
O
R"
H2O / CH3CN
reflux
chiral separation
O
O
R'
R'
R'
(+) or (-)-45
(rac)
(rac)-45
45a R' = H, R" = H
45b R' = H, R" = Cl
45c R' = Cl, R" = H
45d R' = H, R" = F
45e R' = F, R" = H
45f R' = Cl, R" = F
45g R' = F, R" = Cl
45h R' = H, R" = Me
45i R' = F, R" = Me
45j R' = Me, R" = F
Scheme 3. Synthesis of chromanolamine derivatives (45a–j).
O
OBn
1. NHMDS
O
O
N
BnO
O
.HCl
NH2
pyridine, rt
OH
OH
THF, -78°C
pTSA.H2O
HNO3
Ac2O, rt
+
S
S
S
2.
NO2
49
reflux
Cl
OAc
OAc
48
O
54
55
O
46
52
Cl
47
N
BH3.DMS
THF, reflux
S
O
NaBH4
MeOH, rt
O
53
NH2
O
NH2
1. Raney Ni, H2
ethyl acetate, 5°C
1. crystallization with
(+)-S-mandelic acid
OH
OH
S
2. HCl / dioxane
S
NH2
.HCl
(-)-51
2. NaOH
NO2
OH
(+/-)-56
OH
50
(+)-56
Scheme 5. Synthesis of thiophenecyclohexanolamine (56).
Scheme 4. Synthesis of methylcyclohexanolamine (51).
with borane afforded an enantiomerically enriched mixture
(ee = 60%) of thiophenecyclohexanolamines 56. The pure enantio-
mer (+)ꢀ56 (ee >95%) was obtained by recrystallization of its S-
mandelic acid salt from methanol. The absolute stereochemistry
was confirmed by VCD.
PIs 11, 15, 18 and 22 were tested against a panel of HIV-mutant
clinical isolates (M1, M2 and M3) (Table 1). Compared to the mar-
keted drugs lopinavir, atazanavir and darunavir, compounds 11,
15, 18 and 22 performed better overall against wild-type HIV-1
and the resistant mutants, in terms of both absolute potency and
fold changes (FC) of the EC50 value relative to the activity on
wild-type HIV-1. Virtually no resistance to any of the mutant
hydrogenation of the nitro group, methylcyclohexanolamine 51 was
isolated as the hydrochloric acid salt with an ee >95% (as determined
by SFC analysis after conversion to the corresponding Mosher amide).
For the synthesis of the thiophenecyclohexanolamine P2 frag-
ment 56, an enantioselective route was developed as depicted in
Scheme 5. The procedure uses the chiral oxidation reagent camp-
horsulfonyloxaziridine (53) for the asymmetric alpha-oxidation
of commercially available 6,7-dihydro-5H-benzo[b]thiophen-4-
one (52), as reported earlier for the alpha-hydroxylation of
chromanon enolates.25 Condensation of compound 54 with O-ben-
zyl-hydroxyamine and subsequent diastereoselective reduction

316
B. Kesteleyn et al. / Bioorg. Med. Chem. Lett. 23 (2013) 310–317
Table 3
Nanosuspension composition, particle size, dose and bioavailability in rats of compounds 11, 18 and 22
a
a
a
b
Compound Nanosuspension composition
d10
(nm)
d50
d90
Dose (mg/ FIM
(nm)
(nm)
kg)
(%)
18
22
100 mg/mL of Compound 18 and 25 mg/mL polysorbate 20 (tween20) and 9 mg/mL NaCl in water 75
50 mg/mL of Compound 22 and 12.5 mg/mL poloxamer 338 (Pluronic F108) and 50 mg/mL sucrose 78
128
125
211
219
37.5
41
>100
>100
in water
50 mg/mL of Compound 22 and 15 mg/mL Vit-E TPGS in water
72
123
213
40
100
11
100 mg/mL of Compound 11 and 25 mg/mL polysorbate 20 (tween20) and 15 mg/mL NaCl in water 67
50 mg/mL of Compound 11 and 25 mg/mL poloxamer 338 (Pluronic F108) and 50 mg/mL sucrose in 64
123
120
219
213
50
30
3
5
water
50 mg/mL of Compound 11 and 25 mg/mL Vit-E TPGS and 1 mg/mL DSPG and 50 mg/mL sucrose (in 72
122
207
30
12
0.15 M NaCl) in water
a
d10–50–90: the size of particle for which 10%, 50% and 90% of the sample is below this size. Particle sizes of the nanosuspensions were measured by laser diffraction with a
Beckman Coulter LS230 diffractometer (Beckman Coulter, Inc., Fullerton, CA, USA).
b
FIM: absolute bioavailability of the intramuscular (IM) route. Bioavailability was assessed after 1 month of sampling, and was determined as ((AUCIM ꢁ doseIV)/
(AUCIV ꢁ doseIM)) ꢁ 100.
Figure 3. Mean plasma concentration profiles (ng/mL) of compound 11, 18 and 22 long-acting nanosuspensions in rats after IM dosing.
isolates was observed for compounds 11, 18 and 22, and LA phar-
macokinetic profiles were generated for these three most interest-
ing PIs.
Figure 3 shows the LA pharmacokinetic profiles for the low
clearance (11) and two high clearance (18 and 22) PIs. Overall,
the plasma concentration profiles for the three compounds were
found to be quite similar, providing low but sustained drug con-
centration in the plasma over a period of two months. As an overall
observation regarding the effect of the surfactant on compound re-
lease profiles, the nanosuspension formulations stabilized with
surfactants Tween20 (for compounds 11 and 18) and Vit-E TPGS
(for compound 22) provided a more gradual compound release
profile than the formulations based on Pluronic F108. Such LA PK
For an injectable nanosuspension depot formulation, the phar-
macokinetic compound release profiles typically depend partly
upon the nanocrystal particle size distribution and the surfactant
used to stabilize the formulation, but mainly upon the intrinsic
physico-chemical properties of the compound, such as the aqueous
solubility of the crystal form (polymorph). Therefore, the low clear-
ance (11) and the two high clearance (18, 22) PIs were recrystal-
lized and isolated as stable crystalline polymorphic forms. The
aqueous solubility of all polymorphs was found to be less than
profiles, characterized by a lower initial burst and low peak (Cmax
)
to trough (Cmin) ratio, may offer the advantage of higher efficacious
compound concentrations in the plasma, less frequent dosing, and
reduced safety issues.
100 l
g/mL. Using Elan’s NanoCrystal technology,26 sterile nanosus-
pensions were prepared by continuous wet milling of the crystals
in an aqueous carrier containing a hydrophilic surfactant, resulting
in nanocrystals with a particle size of less than 200 nm. As shown
in Table 3, several non-ionic surfactants were evaluated as addi-
tives for the LA nanosuspension formulations: (1) poloxamer 338
(Pluronic F108), (2) D-alpha-tocopheryl polyethylene glycol 1000
succinate (Vit-E TPGS), and (3) polysorbate 20 (tween20). The
nanosuspension LA depot formulations were injected intramuscu-
larly (IM) in rats. Blood samples were taken at regular intervals
over a period of two months after dosing, and the compound re-
lease profiles in plasma and absolute bioavailability (FIM) were
determined.
With regard to bioavailability (FIM), compounds 18 and 22 both
reached a bioavailability of 100% as indicated in Table 3, suggesting
complete dissolution of the nanosuspensions from the injection
site. The bioavailability of the low clearance compound 11, how-
ever, remained low (FIM <12%) after applying several different
types of additive to the nanosuspension formulation. As a result
of the limited bioavailability of compound 11, the postulated
hypotheses of the potential beneficial effect of reduced clearance
on the concentrations of circulating drug could not be evaluated.
As a possible explanation, limited solubility and/or permeability
may have hampered the release of compound 11 from the injection

B. Kesteleyn et al. / Bioorg. Med. Chem. Lett. 23 (2013) 310–317
317
11. Lyle, T. A.; Wiscount, C. M.; Guare, J. P.; Thompson, W. J.; Anderson, P. S.; Darke,
P. L.; Zugay, J. A.; Emini, E. A.; Schleif, W. A.; Quintero, J. C.; Dixon, R. A. F.; Sigal,
I. S.; Huff, J. R. J. Med. Chem. 1991, 34, 1228.
12. Thompson, W. J.; Fitzgerald, P. M. D.; Holloway, M. K.; Emini, E. A.; Darke, P. L.;
McKeever, B. M.; Schleif, W. A.; Quintero, J. C.; Zugay, J. A.; Tucker, T. J.;
Schwering, J. E.; Homnick, C. F.; Nunberg, J.; Springer, J. P.; Huff, J. R. J. Med.
Chem. 1992, 35, 1685.
13. Lin, J. H.; Chen, I. W.; King, J. J. Pharmacol. Exp. Ther. 1992, 263, 105.
14. Alteri, E.; Bold, G.; Cozens, R.; Faessler, A.; Klimkait, T.; Lang, M.; Lazdins, J.;
Poncioni, B.; Roesel, J. L.; Schneider, P.; Walker, M.; Woods-Cook, K. Antimicrob.
Agents Chemother. 1993, 37, 2087.
15. Capraro, H.-G.; Bold, G.; Frässler, A.; Cozens, R.; Klimkait, T.; Lazdins, J.; Mestan,
J.; Poncioni, B.; Rösel, J. L.; Stover, D.; Lang, M. Arch. Pharm. Pharm. Med. Chem.
1996, 329, 273.
16. Cozens, R. M.; Bold, G.; Capraro, H.-G.; Fässler, A.; Mestan, J.; Lang, M.;
Poncioni, B.; Stover, D.; Rösel, J. L. Antiviral Chem. Chemother. 1996, 7,
294.
17. Johnson, V. A.; Calvez, V.; Günthard, H. F.; Paredes, R.; Pillay, D.; Shafer, R.;
Wensing, A. M.; Richman, D. D. Top Antiviral Med. 2011, 19, 156.
18. Bold, G.; Faessler, A.; Capraro, H.-G.; Cozens, R.; Klimkait, T.; Lazdins, J.;
Mestan, J.; Poncioni, B.; Rösel, J.; Stover, D.; Tintelnot-Blomley, M.; Acemoglu,
F.; Beck, W.; Boss, E.; Eschbach, M.; Huerlimann, T.; Masso, E.; Roussel, S.; Ucci-
Stoll, K.; Wyss, D.; Lang, M. J. Med. Chem. 1998, 41, 3387.
site. Despite the high plasma clearance properties of compounds
18 and 20, sustained plasma concentrations well above EC50 values
were obtained for both compounds following single dosing.
With respect to synthetic efficiency, the synthesis of PI 22 (out-
lined in Schemes 1, 2 and 5) requires a total of 15 reaction steps but
can nevertheless be accomplished with an overall yield of 9%. On
the contrary, the current low-yielding procedure for the synthesis
of the methylcyclopentanolamine building block 51 (Scheme 4),
leads to a poor overall yield for the synthesis of PI 18. Nevertheless,
taking into account the high intrinsic antiviral potency against HIV
wild-type, the high activity against a panel of multi-PI-resistant
HIV-mutant clinical isolates and the favorable long-acting
pharmacokinetic properties, PIs 18 and 22, formulated as nanosus-
pensions, are interesting candidates for a LA injectable drug appli-
cation for the treatment of HIV/AIDS.
References and notes
19. Evans, B. E.; Rittle, K. E.; Homnick, C. F.; Springer, J. P.; Hirshfield, J.; Veber, D. F.
J. Org. Chem. 1985, 50, 4615.
20. Degoey, D. A.; Flentge, C. A.; Flosi, W. J.; Grampovnik, D. J.; Kempf, D. J.; Klein, L.
L.; Yeung, M. C.; Randolph, J. T.; Wang, X. C.; Yu, S. U.S. Patent Appl. Publ. 2005,
U.S. 20,050,148,623.
21. Ghosh, A. K.; McKee, S. P.; Thompson, W. J. J. Org. Chem. 1991, 56, 6500.
22. Cheng, Y.; Zhang, F.; Rano, T. A.; Lu, Z.; Schleif, W. A.; Gabryelski, L.; Olsen, D.
B.; Stahlhut, M.; Rutkowski, C. A.; Lin, J. H.; Jin, L.; Emini, E. A.; Chapman, K. T.;
Tata, J. R. Bioorg. Med. Chem. Lett. 2002, 12, 2419.
23. Kuppens, T.; Bultinck, P.; Langenaeker, W. Drug Discovery Today Technol. 2004,
1, 269.
24. Hungate, R. W.; Chen, J. L.; Starbuck, K. E. Tetrahedron Lett. 1991, 32, 6851.
25. Davis, F. A.; Weismuller, M. C. J. Org. Chem. 1990, 55, 3715.
26. Elan Drug Technologies, Nanocrystal Technology, Technology Focus Brochure,
2008. Available from: <http://www.elan.com/EDT/nanocrystal_technology/>
[Last accessed on 9 October, 2011].
1. Bartlett, J. A.; DeMasi, R.; Quinn, J.; Moxham, C.; Rousseau, F. AIDS 2001, 15,
1369.
2. Fogarty, L.; Roter, D.; Larson, S.; Burke, J.; Gillespie, J.; Levy, R. Patient Educ.
Couns. 2002, 2, 93.
3. Rathbun, C.; Farmer, K.; Stephens, J. R.; Lackhart, S. M. Clin. Ther. 2005, 27, 199.
4. Sethi, A. K.; Celentano, D. D.; Gange, S. J.; Moore, R. D.; Gallant, J. E. Clin. Infect.
Dis. 2003, 37, 1112.
5. Baert, L.; Van’t Klooster, G.; Dries, W.; François, M.; Wouters, A.; Basstanie, E.;
Iterbeke, K.; Stappers, F.; Stevens, P.; Schueller, L.; Van Remoortere, P.; Kraus,
G.; Wigerinck, P.; Rosier, J. Eur. J. Pharm. Biopharm. 2009, 72, 502.
6. Rabinow, B. E. Nat. Rev. Drug Disc. 2004, 3, 785.
7. Kohl, N. E.; Emini, E. A.; Schleif, W. A.; Davis, L. J.; Heimbach, J. C.; Dixon, R. A. F.;
Scolnick, E. M.; Sigal, I. S. PNAS 1988, 85, 4686.
8. Martinez-Cajas, J. L.; Wainberg, M. A. Antiviral Res. 2007, 76, 203.
9. Tsantrizos, Y. S. Acc. Chem. Res. 2008, 41, 1252.
10. Vacca, J. P.; Guare, J. P.; deSolms, S. J.; Sanders, W. M.; Giuliani, E. A.; Young, S.
D.; Darke, P. L.; Zugay, J.; Sigal, I. S.; Schleif, W. A.; Quintero, J. C.; Emini, E. A.;
Anderson, P. S.; Huff, J. R. J. Med. Chem. 1991, 34, 1225.