Design and synthesis of potent, orally efficacious hydroxyethylamine derived β-Site amyloid precursor protein cleaving enzyme (Bace1) inhibitors

Article

pubs.acs.org/jmc

Design and Synthesis of Potent, Orally Eﬃcacious

Hydroxyethylamine Derived β-Site Amyloid Precursor Protein

Cleaving Enzyme (BACE1) Inhibitors

Thomas A. Dineen,*^,†Matthew M. Weiss,^†Toni Williamson,^‡Paul Acton,^‡Safura Babu-Khan,^§

Michael D. Bartberger,^∥James Brown,^†Kui Chen,^§Yuan Cheng,^†Martin Citron,^‡,▽

Michael D. Croghan,^†Robert T. Dunn, II,^#Joel Esmay,^⊥Russell F. Graceﬀa,^†Scott S. Harried,^†

Dean Hickman,^⊥Stephen A. Hitchcock,^†,○Daniel B. Horne,^†Hongbing Huang,^†

Ronke Imbeah-Ampiah,^†Ted Judd,^†Matthew R. Kaller,^†Charles R. Kreiman,^†Daniel S. La,^†Vivian Li,^∥

Patricia Lopez,^†Steven Louie,^⊥Holger Monenschein,^†,○Thomas T. Nguyen,^†Lewis D. P_‡ennington,^†

Tisha San Miguel,^§E. Allen Sickmier,^∥Hugo M. Vargas,^#Robert C. Wahl,^§Paul H. Wen,

Douglas A. Whittington,^∥Stephen Wood,^‡Qiufen Xue,^†Bryant H. Yang,^†Vinod F. Patel,^†,◆

and Wenge Zhong^†

^†Chemical Research and Discovery, Amgen Inc., 360 Binney Street, Cambridge, Massachusetts 02142, United States

§

∥

⊥

^‡Department of Neuroscience, Department of HTS and Molecular Pharmacology, Molecular Structure, Pharmacokinetics and

#

Drug Metabolism, Comparative Biology and Safety Sciences, Amgen Inc., One Amgen Center Drive, Thousand Oaks, California

91320, United States

S

* Supporting Information

ABSTRACT: We have previously shown that hydroxyethyl-

amines can be potent inhibitors of the BACE1 enzyme and

that the generation of BACE1 inhibitors with CYP 3A4

inhibitory activities in this scaﬀold aﬀords compounds (e.g., 1)

with suﬃcient bioavailability and pharmacokinetic proﬁles to

reduce central amyloid-β peptide (Aβ) levels in wild-type rats

following oral dosing. In this article, we describe further

modiﬁcations of the P1-phenyl ring of the hydroxyethylamine

series to aﬀord potent, dual BACE1/CYP 3A4 inhibitors which

demonstrate improved penetration into the CNS. Several of

these compounds caused robust reduction of Aβ levels in rat CSF and brain following oral dosing, and compound 37 exhibited

an improved cardiovascular safety proﬁle relative to 1.

inhibition of Aβ formation is an important and direct strategy

for therapeutic intervention in AD.⁴

INTRODUCTION

■

Alzheimer’s disease (AD) is a terminal neurodegenerative

disorder of the brain that begins with mild cognitive and short-

term memory impairment that gradually progresses to

complete loss of cognitive function and ultimately death.¹

While AD is the most common form of dementia in the elderly,

aﬀecting more that 5 million people in the U.S. and over 20

million people worldwide, current treatment options are limited

and do little to alleviate the symptoms of the disease. As a

result, a disease-modifying therapy for AD represents a large

and growing unmet medical need.²A large body of evidence

suggests that AD results from the accumulation of β-amyloid

peptides, particularly Aβ₄₀and Aβ₄₂, in the brain. These

peptides can aggregate to form toxic, soluble oligomers or

further aggregate to form insoluble senile plaques, the

histological hallmark of AD.³While it remains unclear to

what extent each of these species contributes to the overall

pathology of the disease, the available evidence suggests that

Aβ is generated via initial proteolytic cleavage of membrane-

bound amyloid precursor protein (APP) by the aspartyl

protease BACE1 (β-site APP cleaving enzyme) to give a

soluble N-terminal domain (β-sAPP) and a membrane-bound

C-terminal domain (C99). Further processing of C99 by

another aspartyl protease, γ-secretase, aﬀords soluble Aβ

peptides of varying length. Since its identiﬁcation as β-secretase

in 1999,⁵BACE1 has received considerable attention as a

therapeutic target in AD research.⁶BACE1 knockout mice do

not produce Aβ from APP in the brain but are viable,

suggesting that targeted inhibition of the BACE1 enzyme

would prevent formation of Aβ without signiﬁcant on-target

Special Issue: Alzheimer's Disease

Received: January 27, 2012

Published: April 2, 2012

9025

dx.doi.org/10.1021/jm300118s | J. Med. Chem. 2012, 55, 9025−9044

Journal of Medicinal Chemistry

Article

toxicities.⁷Given the central role of the enzyme in Aβ

formation and the viability of knockout mice, the use of a CNS-

penetrant small-molecule inhibitor of BACE1 to reduce Aβ

levels in the brain is an attractive approach toward a disease-

modifying therapy for AD.⁸

activities, some of which demonstrated signiﬁcantly improved

cardiovascular safety margins relative to 1.

CHEMISTRY

■

The compounds described in Tables 1 and 3 were prepared as

shown in Scheme 1. Sulﬁnyl imine 2 was treated with a number

of diﬀerent benzyl Grignard reagents to aﬀord sulﬁnamides 4.¹²

Simultaneous deprotection of the tert-butyl sulﬁnyl and primary

tert-butyldimethylsilyl (TBS) ether groups was accomplished by

exposure of compounds 4 to hydrogen chloride in cold ethanol,

and the resulting intermediates were reprotected to give the

Boc-carbamates 5. Dess−Martin oxidation¹³of the primary

alcohols to the corresponding aldehydes was followed by

sodium triacetoxyborohydride-mediated reductive amination

with amine 6⁹to aﬀord protected hydroxyethylamines 7.

Simultaneous removal of the N-Boc and TBS-groups in

methanolic hydrochloric acid was followed by direct exposure

of the crude amine hydrochlorides to 1-(1H-imidazol-1-yl)-2-

In previous articles, we have described the synthesis and

evaluation of a series of BACE1 inhibitors featuring a

hydroxyethylamine (HEA) scaﬀold as the transition state

isostere to bind the critical catalytic aspartic acid residues of

the BACE1 enzyme. While select compounds from these series

did demonstrate the ability to reduce central Aβ concentrations

following peripheral dosing in Sprague−Dawley (SD) rats, the

pharmacological eﬀects of these compounds were limited by

their low intrinsic stability (as measured in liver microsome

preparations) and poor penetration into the CNS as a result of

P-glycoprotein (Pgp) mediated eﬄux.⁹Consequently, signiﬁ-

cant CNS access and eﬃcacy required either high doses or

codosing with a CYP 3A4 inhibitor. In an eﬀort to obviate these

shortcomings, a strategy was adopted wherein CYP inhibitory

activities were incorporated into our HEA-derived BACE1

inhibitors, leading to the identiﬁcation of 1, a potent BACE1

inhibitor which demonstrated moderate in vivo clearance and

reduced Pgp-mediated eﬄux (Figure 1).¹⁰In the course of

methoxyethanone (8) in the presence of Hunig’s base to aﬀord

̈

the hydroxyethylamines 9. At this point, palladium-catalyzed

cross-coupling reactions were used to form the ﬁnal biaryl

products 11−30. In the case of compounds 22, 23, and 27, the

cross-coupling reactions were performed directly on inter-

mediates 7 to generate compounds 10g−i. Sequential

deprotection and acylation, as described above, aﬀorded the

corresponding ﬁnal products.

The compounds described in Table 4 were used to evaluate

the SAR around the acyl group of the inhibitors, and the

synthesis of these compounds is illustrated in Scheme 2.

Negishi coupling¹⁴of aryl bromide 5a and 2-thiazolyl zinc

bromide in 1,4-dioxane at room temperature aﬀorded biaryl 31

in 92% yield. Following oxidation of the primary alcohol unit of

31 with the Dess−Martin periodinane, reductive amination

with amine 6 in the presence of sodium triacetoxyborohydride

aﬀorded protected hydroxyethylamine 32 in 83% yield.

Deprotection of the N-Boc and TBS-groups was again

accomplished by exposure of 32 to warm methanolic

hydrochloric acid. The crude intermediate amine was directly

acylated by treatment with a preformed solution of the

appropriate acyl imidazole in dichloromethane at 0 °C to

aﬀord inhibitors 34−43.

Figure 1. Properties of BACE1 inhibitor 1.

evaluating inhibitor 1 for oﬀ-target cardiovascular (CV)

liabilities in dog, it was found that this compound exhibited a

narrow therapeutic index due to drug-induced tachycardia of an

unknown mechanism.

In this article, we report on the design and synthesis of a

series of compounds aimed at improving the CV-safety margin

of hydroxyethylamine BACE1 inhibitors exempliﬁed by 1.

Given the uncertainties of the underlying cause of the observed

oﬀ-target CV eﬀects, our strategy to improve upon this series

focused on decreasing the unbound plasma drug concentration

required to achieve signiﬁcant reduction of CNS Aβ. This

strategy thus relied on our ability to improve the functional

cellular potency of our inhibitors and/or reduce their

susceptibility to Pgp-mediated eﬄux and thereby increase

their distribution into the CNS. Our previous work had

demonstrated that the highly optimized azachroman ring

system was required to maintain potent cellular activities in

HEA BACE1 inhibitors which demonstrated suitable properties

for CNS penetration.⁹This work also revealed that the BACE1

enzyme was much more tolerant of modiﬁcation to the P1-

phenyl group, hence our eﬀorts to improve upon 1 focused on

modifying this region of the inhibitor.¹⁰In keeping with our

previously described strategy to design hydroxyethylamines

which demonstrated suitable metabolic stability, we designed

compounds incorporating N-containing heterocycles and

alkynes, functional groups commonly known to inhibit CYP

3A4.¹¹This work led to the identiﬁcation of potent, highly

orally eﬃcacious BACE1 inhibitors with CYP 3A4 inhibitory

RESULTS

■

Our eﬀorts to identify HEA-derived BACE1 inhibitors which

exhibited improved safety proﬁles relied extensively upon our

previously established SAR.^9,10As outlined above, our strategy

involved the generation of dual potent BACE1/CYP 3A4

inhibitors. Although not ideal, this strategy was deemed

necessary so as to maintain suﬃciently high drug plasma levels.

In addition, to ensure adequate exposure within the CNS,

analogues with high permeability and low eﬄux ratios were

targeted. Evaluation of compounds such as 1, containing the

optimized right-hand side azachroman ring system, had

demonstrated that potent inhibitors of both BACE1 and CYP

3A4 could be generated by appropriate substitution of the P1-

phenyl ring at the 3- and 4-positions.¹⁰Furthermore,

incorporation of an ethereal oxygen at the α-position of the

acyl group had been shown to attenuate Pgp-mediated eﬄux.⁹

Upon the basis of these observations, we elected to initiate our

studies by keeping the acyl group constant as the α-methoxy

acetamide and varying the substituent attached to the P1-

9026

dx.doi.org/10.1021/jm300118s | J. Med. Chem. 2012, 55, 9025−9044

Journal of Medicinal Chemistry

Article

a

Scheme 1. Preparation of P1-Biaryl BACE1 Inhibitors

a

Reagents: (a) R-MgBr, THF, −70 °C; (b) HCl, 1,4-dioxane-EtOH, −20 °C, Et₃N, Boc₂O, CH₂Cl₂, −20 to 23 °C; (c) (i) Dess−Martin

periodinane, NaHCO₃, CH₂Cl₂, 0 to 23 °C (ii) 6, (EtO)₃CH, NaBH(OAc)₃, CH₂Cl₂, 23 °C; (d) (i) HCl, 1,4-dioxane-CH₃OH, 40 °C, (ii) 8, i-

Pr₂NEt, CH₂Cl₂, 23 °C; (e) Ar-SnBu₃or R-CC-SnBu₃, Pd₂(dba)₃, X-Phos, CsF, 1,4-dioxane, 160 °C, μ-wave; (f) Ar-SnBu₃, Pd(PPh₃)₄, 1,4-dioxane,

140 °C, μ-wave; (g) R-CC-ZnCl, Pd₂(dba)₃, X-Phos, THF, 120 °C, sealed vessel.

phenyl ring to identify inhibitors with enhanced potency and

reduced eﬄux relative to compound 1.

compound 11 was signiﬁcantly more potent on CYP 3A4 (IC₅₀

= 160 nM) than the 2-pyridyl regioisomer 12 (IC₅₀= 400 nM),

and this diﬀerence is reﬂected in greater turnover of 12 in

human liver microsomes. Further, compounds with estimated

IC₅₀s on CYP 3A4 below 50 nM demonstrated consistently low

microsomal turnover. Overall, inhibitors substituted with a

pyridyl group were less metabolically stable than compounds

containing other heterocycles examined. While compounds

bearing 2-thiazolyl (15 and 21) or 5-thiazolyl (16) groups

demonstrated good intrinsic metabolic stability, 4-thiazolyl

substitution (17) generated a compound with high microsomal

turnover. Introduction of a methyl group on the thiazole ring,

as in 4-methyl-2-thiazolyl analogue 18, signiﬁcantly decreased

metabolic stability. Similarly, an internal alkyne (20) was much

less stable relative to terminal aryl acetylenes (19 and 22).

Given that a primary focus of this work was to identify a

series of BACE1 inhibitors which demonstrated improved

penetration into the CNS, potent inhibitors were also assessed

for their passive permeability and potential for Pgp-mediated

eﬄux. Interestingly, as a whole, this series of compounds

displayed excellent passive permeability and did not show a

tendency for signiﬁcant Pgp-mediated eﬄux. Most of the

After extensive screening, we identiﬁed a series of biaryl

inhibitors which functioned as potent dual BACE1/CYP 3A4

inhibitors. More speciﬁcally, it was found that the incorporation

of a variety of N-containing heterocycles provided compounds

which compared favorably with our previously identiﬁed HEA-

containing BACE1 inhibitors (i.e., 1). As illustrated in Table 1,

a range of hetereocycles were tolerated at this position, and

incorporation of pyridyl (11−12), oxazolyl (13), or thiazolyl

(15−18) groups generated potent BACE1 inhibitors as

measured in both the enzymatic and cellular assays. A notable

exception is N-methylimidazole 14 with reduced potency in the

enzymatic assay but inexplicably good potency in the cellular

assay. Alkynyl-substitution was also well tolerated, with the

terminal aryl acetylene 19 being roughly 3−5-fold more potent

than internal alkyne 20. Substitution at the 4-position of the

P1-phenyl ring with either the 2-thiazolyl group (21) or

terminal alkyne (22) generated less potent inhibitors.

Generally, the relative intrinsic metabolic stability within the

series could be correlated to the estimated potency of the

inhibitors on the CYP 3A4 enzyme. For example, 4-pyridyl

9027

dx.doi.org/10.1021/jm300118s | J. Med. Chem. 2012, 55, 9025−9044

Journal of Medicinal Chemistry

Article

a

Scheme 2. Preparation of Acetamide Analogues

a

Reagents: (a) Pd(Pt-Bu₃)₂(5 mol %), ZnCl₂(2 equiv), 2-thiazolyl zinc bromide (1.2 equiv), dioxane, 23 °C; (b) (i) Dess−Martin periodinane,

NaHCO₃, CH₂Cl₂, 0 to 23 °C, (ii) 6, (EtO)₃CH, NaBH(OAc)₃, CH₂Cl₂, 23 °C; (c) HCl, dioxane-CH₃OH, 40 °C; (d) carboxylic acid, carbonyl

diimidazole, i-Pr₂NEt, CH₂Cl₂, 0 °C.

compounds showed higher eﬄux by rat Pgp relative to the

human transporter; however, in most cases, these values were

within 3-fold of one another. The three compounds which were

recognized and transported by Pgp were the 4-pyridyl, 2-

oxazolyl, and N-methylimidazole derivatives (11, 13, and 14).

Compound 17 showed a moderate potential for rat Pgp-

mediated eﬄux but was less prone to transport by the human

transporter.

considerably reduced exposure relative to its in vitro potency,

which likely explains the moderate eﬃcacy observed with this

inhibitor. Similarly, the modest Aβ reduction observed

following oral administration of 16 can be explained by the

low drug levels in the CSF (2 nM). While it is unclear why the

concentration of 5-thiazolyl isomer 16 was low in CSF at the 4

h time point, the lower passive permeability (9.9 × 10⁻⁶cm/s)

and higher Pgp-mediated eﬄux (eﬄux ratio_rat= 8.5) of oxazole

13 explain the low levels of drug observed in the CSF for this

inhibitor.

With the knowledge that compounds 15 and 19, featuring 3-

(2-thiazolyl)phenyl and 3-(ethynyl)phenyl P1-groups, respec-

tively, produced robust eﬀects in our PD model, we further

explored the SAR of dual BACE1/CYP 3A4 inhibitors bearing

these groups. We anticipated that by identifying additional

compounds with more potent in vivo Aβ reduction in our PD

model, we would increase our chances of identifying BACE1

inhibitors with signiﬁcantly improved cardiovascular safety

margins relative to inhibitors like 1.

Previous work had demonstrated that the incorporation of a

ﬂuorine atom on the P1-phenyl ring can result in signiﬁcant

improvements on the potencies and pharmacokinetic proﬁles of

hydroxyethylamine BACE1 inhibitors,⁹so we next examined

ﬂuorinated analogues of compounds 15 and 19 (Table 3).

Within the series of compounds containing the 2-thiazolyl

moiety, incorporation of a ﬂuorine at the 2 position of the

phenyl ring led to a compound (23) with diminished BACE1

potency. Alternatively, compounds containing a ﬂuorine at the

4, 5, or 6 position (24−26) demonstrated similar levels of

Compounds 13, 15, 16, and 19 oﬀered the best balance of

functional potency and intrinsic metabolic stability, so these

compounds were proﬁled in an acute rat pharmacodynamic

model measuring levels of Aβ₄₀in rat CSF and brain. Each

compound was dosed by oral gavage at 30 mg/kg to male

Sprague−Dawley rats (Table 2). After 4 h, 2-thiazolyl

substituted compound 15 and terminal alkyne 19 demonstrated

robust reduction of Aβ₄₀levels in the CSF (75% and 76%,

respectively) while oxazole 13 and 5-thiazolyl isomer 16 led to

a much less pronounced eﬀect of CSF Aβ₄₀levels (44% and

27%, respectively). At the 4 h time point, reductions of Aβ₄₀

levels in the brain were similar to that observed in the CSF for

compounds 13 and 19 (34% and 68%, respectively), but

surprisingly compound 15 only aﬀorded a 49% reduction in

brain Aβ₄₀levels. The magnitude of these eﬀects could be

correlated with the concentration of inhibitor in the CSF, a

surrogate for brain free fraction.¹⁵Although all compounds

exhibited good total plasma levels (2.5−7 μM) at the 4 h time

point, compounds 15 and 19 achieved CSF concentrations of

greater than 8-fold at their respective in vitro cellular potencies.

On the other hand, CSF levels of 13 were 14 nM, a

9028

dx.doi.org/10.1021/jm300118s | J. Med. Chem. 2012, 55, 9025−9044

Journal of Medicinal Chemistry

Article

Table 1. Substituted P1-Phenyl BACE1 Inhibitors

a

b

c

IC₅₀values were averaged values determined by at least two independent experiments. Human embryonic kidney cells. IC₅₀values were estimated

d

based on a single-concentration experiment. Compound concentration = 3 μM Rat liver microsomal (RLM) and human liver microsomal (HLM)

e

clearance. Compound concentration = 1 μM. Microsomal protein concentration = 250 μg/mL. Apparent permeability measured in parental LLC-

f

PK1 cells. Values are an average of apical to basolateral (A to B) and basolateral to apical (B to A) velocities and are reported as 10⁻⁶cm/s. Eﬄux

measured in LLC-PK1 cells transfected with either rat MDR1A/1B or human MDR1 and are reported as a ratio of (B to A)/(A to B).

potency in the enzymatic assay as that shown by the parent

compound (15). Of note was compound 25, which

demonstrated a roughly 3-fold improvement in our enzymatic

assay. Similar trends were observed within the acetylene series

9029

dx.doi.org/10.1021/jm300118s | J. Med. Chem. 2012, 55, 9025−9044

Journal of Medicinal Chemistry

Article

Table 2. Reduction of CSF Aβ₄₀in Wild-Type Rats

CSF Aβ

brain Aβ

b ⁴⁰

CSF compd

c

plasma compd

d

cell

IC₅₀(nM)

a

b⁴⁰

compd

reduction (%)

conc (μM)

13

15

16

19

44

75

27

76

34

49

0.014 0.016

0.052 0.032

0.002 0.003

0.078 0.032

7.08 4.11

3.27 0.43

2.51 1.30

3.54 0.48

12

6.5

8.2

7.1

ND

68

a

b

Compounds were dosed orally at 30 mg/kg in male Sprague−Dawley rats. n = 5 animals per group. CSF and brain Aβ₄₀measured at t = 4 h. ND =

c

d

not determined. CSF drug concentration measured at t = 4 h. Plasma drug concentration measured at t = 4 h.

Table 3. P1-Fluorophenyl BACE1 Inhibitors

a

b

c

IC₅₀values were averaged values determined by at least two independent experiments. Human embryonic kidney cells. Rat liver microsomal

(RLM) and human liver microsomal (HLM) clearance. Compound concentration = 1 μM. Microsomal protein concentration = 250 μg/mL.

d

Apparent permeability measured in parental LLC-PK1 cells. Values are an average of apical to basolateral (A to B) and basolateral to apical (B to A)

e

velocities and are reported as 10⁻⁶cm/s. Eﬄux measured in LLC-PK1 cells transfected with either rat MDR1A/1B or human MDR1 and are

reported as a ratio of (B to A)/(A to B).

9030

dx.doi.org/10.1021/jm300118s | J. Med. Chem. 2012, 55, 9025−9044

Journal of Medicinal Chemistry

Article

Table 4. SAR of the N-Acyl Unit of P1-3-(2-Thiazolyl)phenyl BACE1 Inhibitors

a

b

c

IC₅₀values were averaged values determined by at least two independent experiments. Human embryonic kidney cells. Rat liver microsomal

(RLM) and human liver microsomal (HLM) clearance. Compound concentration = 1 μM. Microsomal protein concentration = 250 μg/mL.

d

Apparent permeability measured in parental LLC-PK1 cells. Values are an average of apical to basolateral (A to B) and basolateral to apical (B to A)

e

velocities and are reported as 10⁻⁶cm/s. Eﬄux measured in LLC-PK1 cells transfected with either rat MDR1A/1B or human MDR1 and are

reported as a ratio of (B to A)/(A to B).

of compounds. Notably, compound 27, which contains a

ﬂuorine at the 2-position, showed a modest loss in potency,

while the 3-ﬂuoro derivative 29 showed a small improvement in

BACE1 potency. In general, the enzyme-to-cell potency shifts

were marginally lower in the ﬂuorinated derivatives (celluar

IC₅₀/enzyme IC₅₀= 1.4−3.8) compared to parent compound

15 (cellular IC₅₀/enzyme IC₅₀= 5.5). As was observed with the

aryl thiazoles, 4-ﬂuoro substitution in the acetylene series (28)

did not lead to appreciable changes in potency. It should be

noted that within both series of compounds, introduction of

ﬂuorine had no discernible eﬀect on either metabolic stability

or eﬄux potential. It is worth pointing out, however, that

compounds substituted with a ﬂuorine at the 6-position (26

and 30) showed a tendency toward reduced passive

permeability relative to their des-ﬂuoro analogues.

Having established that incorporation of a 2-thiazolyl group

onto the 3-position of the P1-phenyl ring of hydroxyethylamine

compounds produced potent BACE1 inhibitors with moderate

intrinsic metabolic stability, high permeability, and low Pgp-

mediated eﬄux, we next examined systematic variation of the

N-acyl group for further optimization. As shown in Table 4, the

N-acetyl analogue (34) displayed increased BACE1 potency

but suﬀered from signiﬁcantly increased eﬄux ratios. This

supports our previous observations that the Pgp-mediated

eﬄux could be attenuated by incorporating an ethereal oxygen

in the amide tether.^9,10Increasing the chain length of the acyl

group (35) resulted in a diminished human eﬄux ratio, but the

9031

dx.doi.org/10.1021/jm300118s | J. Med. Chem. 2012, 55, 9025−9044

Journal of Medicinal Chemistry

Article

Figure 2. Cocrystal structure of BACE1 complexed with 37. (a) Interactions of 37 with the BACE1 enzyme and extension of the meta-thiazole unit

into S3. (b) Interactions of the methoxyacetamide group with BACE1.

a

Table 5. Reduction of CSF Aβ in Sprague−Dawley Rats

CSF Aβ

brain Aβ

CSF compd

c

plasma compd

d

rat Pgp-eﬄux

b⁴⁰

b ⁴⁰

compd

reduction (%)

conc (μM)

ratio

15

19

25

34

37

41

75

76

64

0

49

68

0.052 0.032

0.078 0.032

0.035 0.018

0.002 0.003

0.085 0.022

0.046 0.046

3.27 0.43

3.54 0.48

11.7 1.53

5.06 1.06

3.11 0.18

4.18 1.99

2.7

2.3

1.1

4.9

1.7

2.1

45

ND

59

69

61

20

a

b

Compounds were dosed orally at 30 mg/kg in male Sprague−Dawley rats. n = 5 animals per group. CSF and brain Aβ₄₀measured at t = 4 h. ND =

c

d

not determined. CSF drug concentration measured at t = 4 h. Plasma drug concentration measured at t = 4 h.

modestly increased lipophilicity dramatically increased the

observed enzyme to cell shift. Alkyl-substitution at the α-

position, as in α-methoxypropanamide 37 and tetrahydrofuran-

2-carboxamide epimers 41 and 42, generated compounds with

similar potency, stability, and permeability properties as

methoxyacetamide 15, although α-methoxypropanamide epi-

mer 38 and α-ethoxypropanamide epimer 39 were slightly less

potent and 39 was slightly less stable in rat liver microsomes as

well. Incorporation of additional oxygenation, as in 2,2-

dimethoxyacetamide 40, was also tolerated. Addition of a free

alcohol to the acetamide (43) resulted in a compound with

good potency on the BACE1 enzyme, but this compound

suﬀered from an increased enzyme to cell shift, reduced passive

permeability, and elevated eﬄux ratios.

An X-ray cocrystal structure of BACE1 with compound 37 is

represented in Figure 2.¹⁶As expected,^9,10the azachroman and

neopentyl groups of the molecule occupy the S1′ and S2′

regions of BACE1, respectively, and the hydroxyethylamine

unit engages in hydrogen bonding interactions with the

catalytic aspartic acid residues, Asp32 and Asp228. While the

phenyl group of the inhibitor occupies the S1 region of BACE1

as observed previously, the crystal structure indicates that the

preferred meta-orientation of the 2-thiazolyl group allows it to

ﬁll the S3 pocket. This observation is consistent with the higher

BACE1 potency of P1 meta-substituted inhibitors (e.g., 15 and

19) relative to their para-substituted counterparts (21 and 22,

respectively). Additionally, the methoxyacetamide unit of 37

appears to bind diﬀerently than the corresponding group of

inhibitor 1. In the latter compound, a bidentate interaction

occurs between the methoxy and carbonyl groups of the

acetamide with Thr72.¹⁰In the case of 37, only the carbonyl of

the acetamide engages Thr72, while the methoxy group

hydrogen bonds to Thr232 via a water molecule. The diﬀerent

orientation of the methoxy group observed in the crystal

structure of 37 and BACE1 may occur to accommodate the

binding of the 2-thiazolyl group in the S3 pocket.

In Vivo Model for Reduction of CNS Aβ Levels. Studies

with potent BACE1 inhibitors in rats had demonstrated that

the extent of Aβ-reduction in the brain and CSF could be

correlated with the concentrations of inhibitor in the CSF, a

surrogate measurement for unbound drug in the brain. Within

an acute setting, the dosing of peripherally restricted BACE1

inhibitors resulted in reduced plasma Aβ concentrations;

however, the levels of Aβ in the brain and CSF remained

unperturbed.⁹Although distribution of a compound into the

CNS is governed by a number of complex factors, compounds

with moderate-to-high passive permeability (≥10 × 10⁻⁶cm/s)

and low transporter-mediated (e.g., Pgp) eﬄux (eﬄux ratios

<3) generally have a better likelihood of achieving good CNS

exposures.¹⁷Most of the compounds featuring either a 2-

thiazolyl-phenyl or ethynylphenyl P1 group that we examined

exhibited low (<3) rat and human eﬄux ratios. Exceptions to

this trend are straight-chain alkyl amides 34 and 35 and

hydroxypropanamide 43. These analogues displayed higher

human and rat eﬄux ratios (generally >3) that are likely to

decrease CNS penetration. Of the methoxyacetamides

examined, only the oxazole 13 and the 6-ﬂuorophenyl

derivatives 26 and 30 exhibited low passive permeabilities

(<15 × 10⁻⁶cm/s), and all but one of the acyl analogues

examined in Table 4 (hydroxypropanamide 43) showed good

passive permeability. On the basis of these considerations, a

select group of compounds from Tables 1−4 were evaluated

further in the acute rat pharmacodynamic model (Table 5). In

general it was found that compounds that achieved CSF

concentrations at or over their functional IC₅₀s resulted in

robust reduction (60−76%) of CSF Aβ₄₀levels. As mentioned

above, Pgp-mediated eﬄux can limit the exposure of

compounds in the CNS, and this trend is exempliﬁed by

9032

dx.doi.org/10.1021/jm300118s | J. Med. Chem. 2012, 55, 9025−9044

Journal of Medicinal Chemistry

Article

Table 6. Plasma Protein Binding and Pharmacokinetic Parameters in Male Sprague−Dawley Rats

a

b

iv

po

t_1/2

(h)

AUC_0→∞

(ng·h/mL)

C_max

(ng/mL)

t_max

(h)

F

(%)

plasma protein

c

compd

CL (L/h/kg)

V_ss(L/kg)

binding (%)

15

19

25

37

1.46

0.71

1.03

0.87

2.19

1.54

1.92

1.22

8.96

5.30

8.81

5.33

9030

24250

5335

1887

4863

4440

5290

3.33

5.33

4.00

5.33

120

100

210

297

96.3

ND

98.4

34770

98.6

a

b

c

Dosed iv at 2 mg/kg as a solution in DMSO. Dosed po at 10 mg/kg as an aqueous suspension in 1% Tween 80 in 2% HMPC. ND = not

determined.

acetamide 34, a compound with good passive permability but

which is susceptible to active transport by Pgp. Dosed orally at

30 mg/kg, this compound achieved only low levels (0.002 μM)

in the CSF at the 4 h time point despite plasma concentrations

of ca. 5.06 μM. As a result, 34 produced no measurable change

in CSF Aβ₄₀levels. In contrast, (R)-2-methoxypropanamide 37

displayed similar passive permeability but did not appear to be a

substrate for Pgp-mediated eﬄux. Despite lower plasma

concentrations (3.11 μM), 37 achieved 42-fold higher CSF

levels (0.085 μM) than 34, producing a robust 69% reduction

in CSF Aβ 4 h after oral dosing.

Pharmacokinetic Proﬁles in Rat. The pharmacokinetic

properties of a group of these compounds were assessed in rat

following both intravenous (iv) and oral (po) dosing. As

summarized in Table 6, thiazole-containing compounds 15, 25,

and 37 and aryl acetylene 19 displayed moderate clearance

(0.71−1.46 L/h/kg) and half-lives (ca. 5−9 h) in male

Sprague−Dawley rats when dosed iv. As a consequence of

their CYP 3A4 inhibitory activities, these compounds each

demonstrated high plasma exposure following a 10 mg/kg oral

dose in addition to excellent bioavailabilities (>100%).

Compound 37 had a particularly high AUC_0→∞of 34470

ng·h/mL, and the thiazole-containing compounds showed

similar plasma protein binding in the range of 96.3−98.6%.

Determination of In Vivo Plasma EC₅₀for Central Aβ₄₀

Reduction. On the basis of their favorable pharmacokinetic

and pharmacodynamic properties, compounds 15 and 37 were

further evaluated in dose−response studies at a 4 h time point.

As shown in Table 7, the total plasma concentrations required

concentrations than those required for compounds such as 1

(EC_{50, unbound}= 176 nM), likely as a result of reduced Pgp-

mediated eﬄux and consequently higher CNS exposures.

Therefore, this series represented a substantial advantage over

earlier HEA BACE1 inhibitors in terms of eﬃcacy in our rat

pharmacodynamic model.

Pharmacokinetic Proﬁle and Cardiovascular Safety

Evaluation of Inhibitor 37 in Male Beagle Dogs. To study

the cardiovascular proﬁle of 37 in an anesthetized dog model,

we ﬁrst determined the pharmacokinetic parameters of this

compound in dogs. Upon intravenous administration to male

beagles, 37 displayed moderate clearance (0.23 L/h/kg) and V_ss

(0.586 L/kg), resulting in a short half-life of 2.8 h (Table 8).

Table 8. Plasma Protein Binding and Pharmacokinetic

a

Parameters of 37 in Male Beagle Dogs

AUC_0→∞

(ng·h/mL)

CL

(L/h/kg)

V_ss

(L/kg)

t_1/2

(h)

plasma protein

binding (%)

compd

37

11970

0.23

0.58

2.82

98.6

a

Dosed iv at 2 mg/kg as a solution in 5% propylene glycol, 95% D5W

at pH = 4.0.

On the basis of this information, compound 37 was

administered intravenously as a series of three 60 min infusions

at doses of 1.5, 5, and 20 mg/kg to two chloralose-anesthetized

dogs; blood samples were taken throughout the infusion period

to quantify exposure. At the end of each infusion period, the

unbound plasma concentrations increased in a dose-dependent

manner and were determined (average) to be 37.5, 219, and

793 nM, respectively. No signiﬁcant eﬀects on mean arterial

pressure, heart rate, cardiac output, or systemic vascular

resistance were observed by the end of the second infusion,

at which point unbound plasma concentrations of 37 were ca.

16-fold over the eﬀective unbound plasma concentration for a

50% reduction in CSF Aβ in the acute rat pharmacodynamic

model. Upon infusion of the high dose, concentrations of 37

were ca. 60-fold over the EC_50,unbound, and reductions in mean

arterial pressure (15%) and systemic vascular resistance (43%)

were observed along with reﬂex tachycardia (50%). The wider

safety margin represented a signiﬁcant improvement in CV

safety for 37 compared to lead compound 1.

Table 7. Eﬀective Concentrations in Plasma for CSF Aβ₄₀

Reduction in Male Sprague−Dawley Rats 4 h after Oral

a

Dosing

EC_{50, total}for CSF

Aβ₄₀reduction (μM)

EC_{50, unbound}for CSF

Aβ₄₀reduction (μM)

compd

15

37

1.85

1.62

0.0066

0.0135

a

Each compound was dosed at 10, 30, and 100 mg/kg, po as an

aqueous suspension in 1% Tween 80 in 2% HMPC. Plasma drug

concentrations and CSF Aβ₄₀concentrations measured at t = 4 h. For

each group, n = 5.

to produce a 50% reduction in CSF Aβ₄₀for both compounds

were determined to be very similar and in the low single-digit

micromolar range (EC_{50, total}= 1.85 μM for 15; EC_{50, total}= 1.62

μM for 37). When corrected for plasma protein binding at

these concentrations, the eﬀective unbound plasma concen-

trations were found to be in the low nanomolar range

CONCLUSIONS

■

Dual BACE1/CYP 3A4 inhibitors were prepared via

modiﬁcation of the hydroxyethylamine scaﬀold to incorporate

N-heterocyclic or alkynyl groups at the 3-position of a P1-

phenyl ring. α-Alkoxy acetamides bearing 2-thiazolyl or ethynyl

groups on the P1-phenyl ring provided the best balance of

functional potency, intrinsic metabolic stability, high passive

permeability, and low eﬄux ratios. By virtue of their CYP 3A4-

(EC_50,

= 6.6 nM and 13.5 nM for 15 and 37,

unbound

respectively). Compounds 15 and 37 produced robust

reductions in CSF Aβ₄₀at signiﬁcantly lower unbound plasma

9033

dx.doi.org/10.1021/jm300118s | J. Med. Chem. 2012, 55, 9025−9044

Journal of Medicinal Chemistry

Article

Prior to the transport experiment,²¹culture medium was aspirated

from both apical and basolateral wells and cells were rinsed with

warmed (37 °C) Hank’s balanced salt solution supplemented with 10

mM Hepes at pH 7.4 (HHBSS, Invitrogen, Grand Island, NY).

HHBSS was removed from wells prior to dosing with test drugs at 5

μM in transport buﬀer (HHBSS containing 0.1% bovine serum

albumin). Then 150 μL of transport buﬀer were added to receiver

chambers prior to dosing in triplicate to apical or basolateral chambers.

The dosed transwell plates containing the cell monolayers were

incubated for 2 h at 37 °C on a shaking platform. At the end of the

incubation period, 100 μL samples were collected from receiver

reservoirs and analyzed by LC-MS/MS on an API4000 (Applied

Biosystem, Foster City, CA) triple quadruple mass spectrometer

interfaced with turbo IonSpray operated in positive mode using

Analyst 1.4.2 software.

inhibiting properties, these compounds exhibited good

pharmacokinetic proﬁles in both rat and dog. When dosed

orally in an acute rat pharmacodynamic model, several

compounds were shown to produce robust decreases in CSF

and brain Aβ₄₀levels following a single 30 mg/kg dose. More

rigorous dose−response eﬃcacy studies demostrated that

compounds 15 and 37 reduced CSF Aβ₄₀levels by 50% at

eﬀective unbound plasma concentrations in the range of 6−13

nM. Compound 37 was further proﬁled in an anesthetized dog

model for cardiovascular hazard identiﬁcation and was found to

produce no signiﬁcant changes in cardiovascular function at

concentrations exceeding 16-fold the eﬀective unbound plasma

concentration required for 50% reduction of CSF Aβ₄₀levels in

rats, a signiﬁcant improvement over BACE1 inhibitors

exempliﬁed by 1. Further studies describing pharmacodynamic

and toxicology studies of these compounds will be reported in

due course.

The apparent permeability coeﬃcient (P_app) of all tested agents was

estimated from the slope of a plot of cumulative amount of the agent

versus time based on the following equation:

P

= (dQ /dt)/(AC₀)

app

EXPERIMENTAL SECTION

■

where dQ/dt is the penetration rate of the agent (ng/s), A is the

surface area of the cell layer on the Transwell (0.11 cm²), and C₀is the

initial concentration of the test compound (ng/mL). Eﬄux ratio (ER)

was calculated from the basolateral-to-apical permeability divided by

the apical-to-basolateral permeability: ER = P_appB > A/P_appA > B.

X-Ray Crystal Structure. The catalytic domain of human BACE1

(residues 14−453) was overexpressed in Escherichia coli and refolded

from inclusion bodies following a procedure described previously.²²

Puriﬁed protein was concentrated to ∼8 mg/mL and 0.5 mM

compound 37 was added to the sample, which was allowed to incubate

on ice for 2 h. BACE1 was crystallized by hanging drop vapor diﬀusion

at room temperature by mixing 1.0 uL of protein solution with 1.0 uL

of reservoir solution containing 16% PEG 8000 and 0.1 M sodium

citrate (pH 5.0). Diﬀraction data were collected at the Advanced Light

Source, beamline 5.0.2, using λ = 1.0000 Å and an ADSC Q315R CCD

detector. Data were reduced using HKL2000.²³The crystals belong to

the space group P2₁with approximate unit cell dimensions of a = 82.7

Å, b = 104.8 Å, and c = 103.9 Å. The structure was solved by molecular

replacement using PHASER²⁴and employing a human BACE1

structure with the same space group (PDB 3DUY) as a search model.

The structure was reﬁned with REFMAC5, and model building was

performed with COOT.^25,26Data collection and reﬁnement statistics

appear in the Supporting Information.

Pharmacodynamic Assay. Male Sprague−Dawley rats (175−200

g) were purchased from Harlan and were maintained on a 12 h light/

dark cycle with unrestricted access to food and water until use. Rats

were administered compound by oral gavage at the appropriate dose.

Rats were euthanized with CO₂inhalation for 2 min, and cisterna

magna was quickly exposed by removing the skin and muscle above it.

CSF (50−100 μL) was collected with a 30 gauge needle through the

dura membrane covering the cisterna magna. Blood was withdrawn by

cardiac puncture and plasma obtained by centrifugation for drug

exposures. Brains were removed and, along with the CSF, immediately

frozen on dry ice and stored at −80 °C until use. The frozen brains

were subsequently homogenized in 10 volumes of (w/v) of 0.5%

Triton X-100 in TBS with protease inhibitors. The homogenates were

centrifuged at 100000 rpm for 30 min at 4 °C. The supernatants were

analyzed for Aβ40 levels by immunoassay as follows: Meso Scale 96-

well avidin plates were coated with Biotin-4G8 (Covance) and

detected with ruthenium-labeled Fab speciﬁc for Aβ40. Plates were

read in MSD Sector6000 imager according to manufacturer’s

recommended protocol (Meso Scale Discovery, Inc.). Aβ40

concentrations were plotted using Graphpad Prism and analyzed by

one-way ANOVA followed by Dunnett’s multiple comparison analysis

to compare drug-treated animals to vehicle-treated controls.

BACE1 Enzymatic Assay. BACE1 enzymatic activity was

determined by the enhancement of ﬂuorescence intensity upon

enzymatic cleavage of the ﬂuorescence resonance energy transfer

substrate. The BACE1 recognition and cleavage sequence of the

substrate is derived from the reported literature,¹⁸and the ﬂuorophore

and quencher dyes are attached to side chain of Lys residues at the

termini of the substrate peptide. The human recombinant BACE1¹⁹

assay was performed in 50 mM acetate, pH 4.5/8% DMSO/100 μM

Genepol/0.002% Brij-35. In dose−response IC₅₀assays, 10 point 1:3

serial dilutions of compound in DMSO were preincubated with the

enzyme for 60 min at room temperature. Subsequently, the substrate

was added to initiate the reaction. After 60 min at room temperature,

the reaction was stopped by addition of 0.1 M Tris base to raise the

pH above the enzyme active range, and the increase of ﬂuorescence

intensity was measured on Saﬁre II microplate reader (Tecan,

Mannedorf, Switzerland).

̈

Cell-Based Assay. Human embryonic kidney cells (HEK293)

stably expressing APP_SWwere plated at a density of 100K cells/well in

96-well plates (Costar). The cells were cultivated for 6 h at 37 °C and

5% CO₂in DMEM supplemented with 10% FBS. Cells were incubated

overnight with test compounds at concentrations ranging from 0.0005

to 10 μM. Following incubation with the test compounds, the

conditioned media was collected and the Aβ40 levels were determined

using a sandwich ELISA. The IC₅₀was calculated from the percent of

control Aβ40 as a function of the concentration of the test compound.

The sandwich ELISA to detect Aβ40 was performed in 96-well

microtiter plates, which were precoated with goat antirabbit IgG

(Pierce). The capture and detection antibody pair that was used to

detect Aβ40 from cell supernatants consists of aﬃnity puriﬁed pAβ40

(Invitrogen) and biotinylated 6E10 (Covance), respectively. Con-

ditioned media was incubated with capture antibody overnight at 4 °C,

followed by washing. The detecting antibody incubation was for 3 h at

4 °C, again followed by the wash steps as described previously. The

plate was developed using Delﬁa reagents (Streptavidin−Europium

and Enhancement solution (Perkin-Elmer)), and time-resolved

ﬂuorescence was measured on an EnVision multilabel plate reader

(Perkin-Elmer).

Permeability Assay. The wild-type cell line LLC-PK1 (porcine

renal epithelial cells, WT-LLC-PK1) was purchased from American

Type Culture Collection (ATCC, Manassass, VA). Transfections of

WT-LLC-PK1 cells with human MDR1 gene (hMDR1-LLC-PK1) and

rat mdr1a gene (rMdr1a-LLC-PK1) were generated. Cells were grown

in Medium 199 supplemented with 10% fetal bovine serum.²⁰Cells

were seeded onto matrigel-coated transwell ﬁlter membranes at a

density of 90000 cells/well. Media change was performed on day 3.

Compound incubations were performed 5−6 days post seeding. All

cultures were incubated at 37 °C in a humidiﬁed (95% relative

humidity) atmosphere of 5% CO₂/95% air.

CYP3A4 Inhibition Assay. Pooled human liver microsomes (0.1

mg/mL) were incubated at 37 °C in a phosphate buﬀer (pH 7.4) with

the selective CYP3A substrate midazolam at a concentration of 2.5 μM

in the presence and absence of test compound (at about a 3 μM

9034

dx.doi.org/10.1021/jm300118s | J. Med. Chem. 2012, 55, 9025−9044

Journal of Medicinal Chemistry

Article

concentration). The reaction was started with the addition of NADPH

(1 mM ﬁnal concentration). Incubations were stopped after 10 min by

the addition of organic solvent and 1-hydroxymidazolam. Metabolite

formation was measured by HPLC-MS. The ability of the test

compound to inhibit the activity of CYP3A4 was determined.

Estimated IC₅₀s were calculated based on the ratio of the

concentration of metabolite in the presence of test compound to

the concentration of metabolite in the absence of test compound.

Chemistry. Unless otherwise noted, all materials were obtained

from commercial suppliers and used without further puriﬁcation.

Sulﬁnyl imine 2 and azachroman 6 were prepared as described in ref 9.

Benzyl Grignard reagents were purchased from Rieke Metals, Inc. as

solutions in ether. Anhydrous solvents were obtained from Aldrich or

EM Science and used directly. All reactions involving air- or moisture-

sensitive reagents were performed under a nitrogen or argon

atmosphere. All microwave assisted reactions were conducted with a

Smith synthesizer from Personal Chemistry, Uppsala, Sweden. Silica

gel chromatography was performed using either glass columns packed

with silica gel (230−400 mesh, EMD Chemicals, Gibbstown, NJ) or

(S)-N-((2S,3S)-3,4-Bis(tert-butyldimethylsilyloxy)-1-(3-chloro-

5-ﬂuorophenyl)butan-2-yl)-2-methylpropane-2-sulﬁnamide

(4b). The title compound was prepared via a method analogous to the

preparation of compound 4a using (3-chloro-5-ﬂuorobenzyl)-

1

magnesium chloride, giving a pale-yellow oil (24.0 g, 71% yield). H

NMR (300 MHz, chloroform-d) δ ppm 6.96 (s, 1 H), 6.91 (dt, J =

8.40, 2.08 Hz, 1 H), 6.78 (dt, J = 9.21, 1.75 Hz, 1 H), 4.19 (ddd, J =

8.51, 5.88, 2.41 Hz, 1 H), 3.62−3.82 (m, 2 H), 3.43−3.62 (m, 2 H),

2.90 (dd, J = 14.10, 3.73 Hz, 1 H), 2.58 (dd, J = 14.10, 10.60 Hz, 1 H),

1.01 (s, 9 H), 0.94 (s, 9 H), 0.93 (s, 9 H), 0.19 (s, 3 H), 0.07−0.15 (m,

9 H).

(S)-N-((2S,3S)-3,4-Bis(tert-butyldimethylsilyloxy)-1-(3-chloro-

4-ﬂuorophenyl)butan-2-yl)-2-methylpropane-2-sulﬁnamide

(4c). The title compound was prepared via a method analogous to the

preparation of compound 4a using (3-chloro-4-ﬂuorobenzyl)-

1

magnesium bromide, giving a pale yellow oil (1.4 g, 80% yield). H

NMR (300 MHz, chloroform-d) δ ppm 7.14−7.22 (m, 1 H), 6.96−

7.08 (m, 2 H), 4.12−4.22 (m, 1 H), 3.51−3.79 (m, 4 H), 2.88 (dd, J =

14.25, 3.58 Hz, 1 H), 2.56 (dd, J = 14.10, 10.45 Hz, 1 H), 1.00 (s, 9

H), 0.94 (s, 9 H), 0.93 (s, 9 H), 0.18 (s, 3 H), 0.06−0.14 (m, 9 H).

(S)-N-((2S,3S)-3,4-Bis(tert-butyldimethylsilyloxy)-1-(5-chloro-

2-ﬂuorophenyl)butan-2-yl)-2-methylpropane-2-sulﬁnamide

(4d). The title compound was prepared via a method analogous to the

preparation of compound 4a using (5-chloro-2-ﬂuorobenzyl)-

1

prepacked silica gel cartridges (Biotage or ISCO). H NMR spectra

were recorded on a Bruker AV-400 (400 MHz) spectrometer at

ambient temperature or on a Varian 400 MHz spectrometer. Chemical

shifts are reported in parts per million (ppm, δ units) downﬁeld from

tetramethylsilane. Data are reported as follows: chemical shift,

multiplicity (s = singlet, d = doublet, t = triplet, q = quartet, br =

broad, m = multiplet), coupling constants, and number of protons.

Purity for ﬁnal compounds was greater than 95% unless otherwise

noted and was measured using Agilent 1100 series high performance

liquid chromatography (HPLC) systems with UV detection at 254 nm

(system A, Agilent Zorbax Eclipse XDB-C8 4.6 mm ×150 mm, 5 μm,

5−100% CH₃CN in H₂O with 0.1% TFA for 15 min at 1.5 mL/min;

system B, Waters Xterra 4.6 mm ×150 mm, 3.5 μm, 5−95% CH₃CN

in H₂O with 0.1% TFA for 15 min at 1.0 mL/min). Exact mass

conﬁrmation was performed on an Agilent 1100 series high

performance liquid chromatography (HPLC) system (Santa Clara,

CA, USA) by ﬂow injection analysis, eluting with a binary solvent

system A and B (A, water with 0.1% FA; B, ACN with 0.1% FA) under

isocratic conditions (50% A/50% B) at 0.2 mL/min with MS detection

by an Agilent G1969A time-of-ﬂight (TOF) mass spectrometer (Santa

Clara, CA, USA).

1

magnesium bromide, giving a pale-yellow oil (4.0 g, 63% yield). H

NMR (300 MHz, chloroform-d) δ ppm 7.05−7.19 (m, 2 H), 6.82−

6.98 (m, 1 H), 4.22 (ddd, J = 8.37, 5.88, 2.27 Hz, 1 H), 3.75 (ddd, J =

10.19, 4.71, 2.34 Hz, 1 H), 3.47−3.72 (m, 3 H), 2.65−2.88 (m, 2 H),

1.27 (s, 9 H), 0.97 (s, 9 H), 0.94 (s, 9 H), 0.20 (s, 3 H), 0.08−0.14 (m,

9 H).

(S)-N-((2S,3S)-3,4-Bis(tert-butyldimethylsilyloxy)-1-(4-

chlorophenyl)butan-2-yl)-2-methylpropane-2-sulﬁnamide

(4e). The title compound was prepared via a method analogous to the

preparation of compound 4a using 4-chlorobenzylmagnesium chloride,

giving a pale-yellow oil (86 g, 77% yield). ¹H NMR (400 MHz,

chloroform-d) δ ppm 7.23 (d, J = 8.53 Hz, 2 H), 7.03−7.13 (m, 2 H),

4.12−4.21 (m, 1 H), 3.70−3.82 (m, 1 H), 3.67 (dd, J = 10.04, 5.52 Hz,

1 H), 3.48−3.61 (m, 2 H), 2.89 (dd, J = 14.56, 4.02 Hz, 1 H), 2.60

(dd, J = 14.05, 10.54 Hz, 1 H), 0.98 (s, 9 H), 0.94 (s, 9 H), 0.93 (s, 9

H), 0.18 (s, 3 H), 0.07−0.13 (m, 9 H).

(S)-N-((2S,3S)-3,4-Bis(tert-butyldimethylsilyloxy)-1-(3-chloro-

2-ﬂuorophenyl)butan-2-yl)-2-methylpropane-2-sulﬁnamide

(4f). The title compound was prepared via a method analogous to the

preparation of compound 4a using (3-chloro-2-ﬂuorobenzyl)-

(S)-N-((2S, 3S)-1-(3-Bromophenyl)-3, 4-bis(tert-

butyldimethylsilyloxy)butan-2-yl)-2-methylpropane-2-sulﬁna-

mide (4a). A 2 L ﬂask was charged with (S,E)-N-((S)-2,3-bis(tert-

butyldimethylsilyloxy)propylidene)-2-methylpropane-2-sulﬁnamide

(2) (50.00 g, 119 mmol) and THF (300 mL), and the resulting

solution was cooled to −65 °C. (3-Bromobenzyl)magnesium chloride

(948 mL of a 0.25 M solution in diethyl ether, 237 mmol) was added

dropwise via cannula over 2.5 h such that the internal temperature did

not rise above −60 °C. The resulting yellow slurry was stirred for 2 h

at −65 °C, at which time TLC analysis indicated consumption of the

starting material. The reaction mixture was warmed −25 °C, at which

point a saturated aq ammonium chloride solution (200 mL) was

added. The resulting white solids were dissolved upon the addition of

water (150 mL). The mixture was warmed to room temperature, and

the layers were separated. The aqueous layer was extracted with ethyl

acetate (250 mL). The organic layers were combined, and the

combined solution was dried over sodium sulfate, ﬁltered, and

concentrated to aﬀord a yellow oil. The crude product was puriﬁed by

chromatography on silica gel (0−10% EtOAc/hexane) to aﬀord (S)-N-

((2S,3S)-1-(3-bromophenyl)-3,4-bis(tert-butyldimethylsilyloxy)butan-

2-yl)-2-methylpropane-2-sulﬁnamide (4a) as a pale-yellow oil that

1

magnesium bromide, giving a pale-yellow oil (24.5 g, 77% yield). H

NMR (400 MHz, acetonitrile-d₃) δ ppm 7.29 (ddd, J = 7.97, 7.09, 1.76

Hz, 1 H), 7.14−7.22 (m, 1 H), 7.03−7.10 (m, 1 H), 4.13 (ddd, J =

7.87, 5.72, 2.15 Hz, 1 H), 3.59−3.77 (m, 4 H), 2.84−2.91 (m, 2 H),

0.94 (s, 9 H), 0.94 (s, 9 H), 0.90 (s, 9 H), 0.17 (s, 3 H), 0.13 (s, 3 H),

0.12 (s, 6 H).

tert-Butyl (2S,3S)-1-(3-Bromophenyl)-3-(tert-butyldimethyl-

silyloxy)-4 Hydroxybutan-2-ylcarbamate (5a). A solution of

(S)-N-((2S,3S)-1-(3-bromophenyl)-3,4-bis(tert-butyldimethylsilyloxy)-

butan-2-yl)-2-methylpropane-2-sulﬁnamide (4a) (5.00 g, 8.43 mmol)

in ethanol (60 mL) at −40 °C was added hydrogen chloride (10.5 mL

of a 4N solution in 1,4-dioxane, 42.2 mmol) dropwise over 3 min such

that the maintained the temperature did not rise above −35 °C. The

reaction was brieﬂy warmed to −5 °C to dissolve the white precipitate

that had formed, and then the resulting homogeneous solution was

maintained at −20 °C for 16 h. After this time, N,N-diisopropylethyl-

amine (10.3 mL, 59.0 mmol) and di-tert-butyl dicarbonate (3.68 g,

16.9 mmol) were added in sequence, and the reaction was warmed to

23 °C. After stirring for 30 min, the reaction mixture was diluted with

dichloromethane (100 mL) and poured into a 50% aq ammonium

chloride solution (500 mL). The layers were separated, and the

aqueous layer was extracted with dichloromethane (3 × 100 mL). The

organic layers were combined, and the combined solution was washed

with water (100 mL), washed with brine (100 mL), dried over sodium

sulfate, ﬁltered, and concentrated. Puriﬁcation of the crude product by

chromatography on silica gel (0−25% EtOAc/Hexane) aﬀorded tert-

butyl (2S,3S)-1-(3-bromophenyl)-3-(tert-butyldimethylsilyloxy)-4 hy-

1

formed a waxy solid upon standing at room temperature. H NMR

(400 MHz, chloroform-d) δ ppm 7.29−7.36 (m, 2 H), 7.12 (ddd, J =

15.43, 7.65, 7.53 Hz, 2 H), 4.18 (ddd, J = 8.16, 5.90, 2.51 Hz, 1 H),

3.71−3.80 (m, 1 H), 3.67 (dd, J = 10.29, 5.77 Hz, 1 H), 3.49−3.60 (m,

2 H), 2.90 (dd, J = 14.05, 4.02 Hz, 1 H), 2.58 (dd, J = 13.80, 10.79 Hz,

1 H), 0.99 (s, 9 H), 0.94 (s, 9 H), 0.93 (s, 9 H), 0.18 (s, 3 H), 0.07−

0.14 (m, 6 H), 0.01 (s, 3 H). MS m/z = 592.2 [M + H]⁺. Calcd for

C₂₆H₅₁BrNO₃SSi₂: 592.2.

9035

dx.doi.org/10.1021/jm300118s | J. Med. Chem. 2012, 55, 9025−9044

Journal of Medicinal Chemistry

Article

droxybutan-2-ylcarbamate (5a) (2.75 g, 68.7% yield) as a colorless oil.

¹H NMR (400 MHz, chloroform-d) δ ppm 7.32−7.39 (m, 2 H), 7.09−

7.20 (m, 2 H), 4.67 (d, J = 8.51 Hz, 1 H), 4.02−4.24 (m, 1 H), 3.93

(br s, 1 H), 3.69−3.77 (m, 2 H), 3.59−3.68 (m, 1 H), 2.99 (dd, J =

14.28, 4.40 Hz, 1 H), 2.69 (dd, J = 13.55, 9.54 Hz, 1 H), 1.37 (s, 9 H),

0.94 (s, 9 H), 0.11 (d, J = 1.86 Hz, 6 H).

tert-Butyl (2S,3S)-3-(tert-Butyldimethylsilyloxy)-1-(3-chloro-

5-ﬂuorophenyl)-4 Hydroxybutan-2-ylcarbamate (5b). The title

compound was prepared from 4b via a method analogous to the

preparation of compound 5a, giving a colorless oil (36.0 g, 88% yield).

¹H NMR (300 MHz, chloroform-d) δ ppm 7.00 (s, 1 H), 6.94 (dt, J =

8.48, 2.05 Hz, 1 H), 6.84 (d, J = 9.35 Hz, 1 H), 4.68 (d, J = 8.92 Hz, 1

H), 3.99−4.24 (m, 1 H), 3.91 (tt, J = 9.59, 5.02 Hz, 1 H), 3.57−3.80

(m, 2 H), 2.99 (dd, J = 14.25, 4.17 Hz, 1 H), 2.60−2.75 (m, 1 H), 2.37

(br s, 1 H), 1.36 (s, 9 H), 1.19 (s, 9 H), 0.11 (s, 6 H).

tert-Butyl (2S,3S)-3-(tert-Butyldimethylsilyloxy)-1-(3-chloro-

4-ﬂuorophenyl)-4 Hydroxybutan-2-ylcarbamate (5c). The title

compound was prepared from 4c via a method analogous to the

preparation of compound 5a, giving a colorless oil (0.49 g, 84% yield).

¹H NMR (300 MHz, chloroform-d) δ ppm 7.24 (d, J = 7.45 Hz, 1 H),

6.99−7.12 (m, 2 H), 4.67 (d, J = 8.77 Hz, 1 H), 3.83−3.97 (m, J =

9.37, 9.37, 4.71, 4.46 Hz, 1 H), 3.58−3.79 (m, 3 H), 2.98 (dd, J =

14.18, 4.24 Hz, 1 H), 2.65 (dd, J = 14.18, 10.08 Hz, 1 H), 2.40 (t, J =

6.07 Hz, 1 H), 1.36 (s, 9 H), 0.90−0.95 (m, 9 H), 0.07−0.15 (m, 6 H).

MS m/z = 470.1 [M + Na]⁺. Calcd for C₂₁H₃₅ClFNNaO₄Si: 470.1.

tert-Butyl (2S,3S)-3-(tert-Butyldimethylsilyloxy)-1-(5-chloro-

2-ﬂuorophenyl)-4 hydroxybutan-2-ylcarbamate (5d). The title

compound was prepared from 4d via a method analogous to the

preparation of compound 5a, giving a colorless oil (2.9 g, 92% yield).

¹H NMR (300 MHz, chloroform-d) δ ppm 7.06−7.25 (m, 2 H), 6.97

(t, J = 8.99 Hz, 1 H), 4.77 (d, J = 9.06 Hz, 1 H), 4.00−4.31 (m, 1 H),

3.84−4.00 (m, 1 H), 3.56−3.85 (m, 2 H), 3.01 (dd, J = 14.47, 3.65 Hz,

1 H), 2.52−2.82 (m, 1 H), 2.32 (t, J = 6.14 Hz, 1 H), 1.36 (s, 9 H),

0.94 (s, 9 H), 0.12 (s, 6 H).

tert-Butyl (2S,3S)-3-(tert-Butyldimethylsilyloxy)-1-(4-chloro-

phenyl)-4 hydroxybutan-2-ylcarbamate (5e). The title com-

pound was prepared from 4e via a method analogous to the

preparation of compound 5a, giving a colorless oil (17.8 g, 84% yield).

¹H NMR (400 MHz, chloroform-d) δ ppm 7.22−7.33 (m, 2 H), 7.14

(d, J = 8.41 Hz, 2 H), 4.66 (d, J = 8.41 Hz, 1 H), 3.87−4.01 (m, 1 H),

3.51−3.78 (m, 2 H), 2.98 (dd, J = 14.18, 4.40 Hz, 1 H), 2.68 (dd, J =

13.40, 10.07 Hz, 1 H), 2.39 (br s, 1 H), 1.31−1.42 (m, 9 H), 0.93 (s, 9

H), 0.11 (s, 6 H).

tert-Butyl (2S,3S)-3-(tert-Butyldimethylsilyloxy)-1-(3-chloro-

2-ﬂuorophenyl)-4 hydroxybutan-2-ylcarbamate (5f). The title

compound was prepared from 4f via a method analogous to the

preparation of compound 5a, giving a colorless oil (17.1 g, 88% yield).

¹H NMR (400 MHz, acetonitrile-d₃) δ ppm 7.31 (t, J = 7.29 Hz, 1 H),

7.15−7.25 (m, 1 H), 7.01−7.12 (m, 1 H), 5.16 (d, J = 9.29 Hz, 1 H),

3.98−4.16 (m, 1 H), 3.84−3.97 (m, 1 H), 3.75 (q, J = 4.86 Hz, 1 H),

3.52−3.63 (m, 1 H), 3.03 (dd, J = 13.30, 2.93 Hz, 1 H), 2.87 (t, J =

5.82 Hz, 1 H), 2.66 (t, J = 12.52 Hz, 1 H), 1.27 (br s, 9 H), 0.93 (s, 9

H), 0.11 (s, 6 H).

tert-Butyl (2S,3R)-1-(3-Bromophenyl)-3-(tert-butyldimethyl-

silyloxy)-4-((S)-6′-neopentyl-3′,4′-dihydrospiro[cyclobutane-

1,2′-pyrano[2,3-b]pyridine]-4′-ylamino)butan-2-ylcarbamate

(7a). A solution of tert-butyl (2S,3S)-1-(3-bromophenyl)-3-(tert-

butyldimethylsilyloxy)-4-oxobutan-2-ylcarbamate (5a) (0.68 g, 1.4

mmol) and (S)-6′-neopentyl-3′,4′-dihydrospiro[cyclobutane-1,2′-

pyrano[2,3-b]pyridin]-4′-amine (6) (0.37 g, 1.4 mmol) in dichloro-

methane (10 mL) was treated with trimethyl orthoformate (1.3 mL,

12 mmol). The solution was stirred for 1 h, then sodium

triacetoxyborohydride (1.2 g, 5.8 mmol) was added, and the resulting

mixture was stirred for an additional hour. The reaction mixture was

diluted with dichloromethane (25 mL) and saturated aq Rochelle’s salt

solution (50 mL). The biphasic mixture was stirred vigorously for 1 h,

and then the layers were separated. The aq layer was extracted with

dichloromethane (2 × 25 mL). The organic extracts were combined,

and the combined solution was washed with water, washed with brine,

dried over sodium sulfate, ﬁltered, and concentrated. The crude

product was puriﬁed by chromatography on silica gel (0−30% EtOAc/

Hexane) to aﬀord tert-butyl (2S,3R)-1-(3-bromophenyl)-3-(tert-

butyldimethylsilyloxy)-4-((S)-6′-neopentyl-3′,4′-dihydrospiro-

[cyclobutane-1,2′-pyrano[2,3-b]pyridine]-4′-ylamino)butan-2-ylcarba-

1

mate (7a) (0.845 g, 82% yield) as a white solid. H NMR (400 MHz,

chloroform-d) δ ppm 7.89 (br s, 1 H), 7.57 (br s, 1 H), 7.37 (s, 2 H),

7.10−7.21 (m, 2 H), 3.75−4.10 (m, 3 H), 2.87−3.00 (m, 1 H), 2.81

(d, J = 7.82 Hz, 2 H), 2.50−2.66 (m, 1 H), 2.29−2.47 (m, 2 H), 2.17−

2.29 (m, 1 H), 2.04−2.15 (m, 1 H), 1.92−2.03 (m, 1 H), 1.65−1.82

(m, 1 H), 1.39 (br s, 9 H), 0.90 (s, 18 H), 0.83−0.89 (m, 6 H), 0.07

(br s, 3 H), 0.05 (br s, 3 H). MS m/z = 716.2 [M + H]⁺. Calcd for

C₃₇H₅₉BrN₃O₄Si: 716.4.

tert-Butyl (2S,3R)-3-(tert-Butyldimethylsilyloxy)-1-(3-chloro-

5-ﬂuorophenyl)-4-((S)-6′-neopentyl-3′,4′-dihydrospiro-

[cyclobutane-1,2′-pyrano[2,3-b]pyridine]-4′-ylamino)butan-2-

ylcarbamate (7b). The title compound was prepared from 5b and 6

via a method analogous to the preparation of 7a, giving a white solid

(2.3 g, 99% yield). MS m/z = 690.2 [M + H]⁺. Calcd for

C₃₇H₅₈ClFN₃O₄Si: 690.4.

tert-Butyl (2S,3R)-3-(tert-Butyldimethylsilyloxy)-1-(3-chloro-

4-ﬂuorophenyl)-4-((S)-6′-neopentyl-3′,4′-dihydrospiro-

[cyclobutane-1,2′-pyrano[2,3-b]pyridine]-4′-ylamino)butan-2-

ylcarbamate (7c). The title compound was prepared from 5c and 6

via a method analogous to the preparation of 7a, giving a white solid

(3.8 g, 83% yield). MS m/z = 690.3 [M + H]⁺. Calcd for

C₃₇H₅₈ClFN₃O₄Si: 690.4.

tert-Butyl (2S,3R)-3-(tert-Butyldimethylsilyloxy)-1-(5-chloro-

2-ﬂuorophenyl)-4-((S)-6′-neopentyl-3′,4′-dihydrospiro-

[cyclobutane-1,2′-pyrano[2,3-b]pyridine]-4′-ylamino)butan-2-

ylcarbamate (7d). The title compound was prepared from 5d and 6

via a method analogous to the preparation of 7a, giving a white solid

(0.13 g, 78% yield). MS m/z = 690.3 [M + H]⁺. Calcd for

C₃₇H₅₈ClFN₃O₄Si: 690.4.

tert-Butyl (2S,3R)-3-(tert-Butyldimethylsilyloxy)-1-(4-chloro-

phenyl)-4-((S)-6′-neopentyl-3′,4′-dihydrospiro[cyclobutane-

1,2′-pyrano[2,3-b]pyridine]-4′-ylamino)butan-2-ylcarbamate

(7e). The title compound was prepared from 5e and 6 via a method

analogous to the preparation of 7a, giving a white solid (0.50 g, 54%

yield). MS m/z = 671.8 [M + H]⁺. Calcd for C₃₇H₅₉ClN₃O₄Si: 672.4.

tert-Butyl (2S,3R)-3-(tert-Butyldimethylsilyloxy)-1-(3-chloro-

2-ﬂuorophenyl)-4-((S)-6′-neopentyl-3′,4′-dihydrospiro-

[cyclobutane-1,2′-pyrano[2,3-b]pyridine]-4′-ylamino)butan-2-

ylcarbamate (7f). The title compound was prepared from 5f and 6

via a method analogous to the preparation of 7a, giving a white solid

(7.9 g, 72% yield). MS m/z = 690.1 [M + H]⁺. Calcd for

C₃₇H₅₈ClFN₃O₄Si: 690.4.

N-((2S,3R)-1-(3-Bromophenyl)-3-hydroxy-4-((S)-6′-neopen-

tyl-3′,4′-dihydrospiro[cyclobutane-1,2′-pyrano[2,3-b]-

pyridine]-4′-ylamino)butan-2-yl)-2-methoxyacetamide (9a). A

solution of (2S,3R)-1-(3-bromophenyl)-3-(tert-butyldimethylsilyloxy)-

4-((S)-6′-neopentyl-3′,4′-dihydrospiro[cyclobutane-1,2′-pyrano[2,3-b]-

pyridine]-4′-ylamino)butan-2-ylcarbamate (7a) (1.45 g, 2.00 mmol) in

methanol (5 mL) was treated with hydrogen chloride (10 mL of a 4N

solution in 1,4-dioxane, 40 mmol), and the resulting solution was

stirred for 16 h. The reaction mixture was concentrated in vacuo, and

the residue was taken up in dichloromethane (25 mL) and N,N-

diisopropylethylamine (7.8 mL, 44 mmol) to give a clear solution. A

solution of 1-(1H-imidazol-1-yl)-2-methoxyethanone (1.0 mL, 2.0

mmol) in dichloromethane (10 mL) was added, and the reaction

mixture was to stirred at room temperature for 15 h. The reaction

mixture was washed with a saturated aq sodium bicarbonate solution

(50 mL), the layers were separated, and the aqueous layer was

extracted with dichloromethane (2 × 25 mL). The organic layers were

combined, and the combined solution was washed with water (50

mL), washed with brine (50 mL), dried over sodium sulfate, ﬁltered,

and concentrated. The crude product was puriﬁed by chromatography

on silica gel (0−5% MeOH/DCM) to aﬀord N-((2S,3R)-1-(3-

bromophenyl)-3-hydroxy-4-((S)-6′-neopentyl-3′,4′-dihydrospiro-

[cyclobutane-1,2′-pyrano[2,3-b]pyridine]-4′-ylamino)butan-2-yl)-2-

methoxyacetamide (9a) (1.090 g, 94% yield) as a white solid. ¹H

NMR (400 MHz, acetonitrile-d₃) δ ppm 7.79 (d, J = 1.76 Hz, 1 H),

7.70 (dd, J = 2.40, 1.03 Hz, 1 H), 7.40−7.45 (m, 1 H), 7.32−7.39 (m,

9036

dx.doi.org/10.1021/jm300118s | J. Med. Chem. 2012, 55, 9025−9044

Journal of Medicinal Chemistry

Article

1 H), 7.17−7.24 (m, 2 H), 6.89 (d, J = 10.17 Hz, 1 H), 4.02−4.13 (m,

1 H), 3.90 (dd, J = 10.47, 5.48 Hz, 1 H), 3.53−3.79 (m, 4 H), 3.36−

3.46 (m, 1 H), 3.25 (s, 2 H), 3.09 (dd, J = 14.04, 3.86 Hz, 1 H), 2.65−

2.82 (m, 3 H), 2.43 (s, 4 H), 2.32−2.42 (m, 1 H), 2.17−2.24 (m, 1 H),

2.08−2.13 (m, 1 H), 1.87 (br s, 1 H), 1.71−1.79 (m, 1 H), 1.65 (ddd,

J = 13.25, 10.51, 1.17 Hz, 1 H), 0.88 (s, 9 H). MS m/z = 574.2 [M +

H]⁺. Calcd for C₂₉H₄₁BrN₃O₄: 574.2.

N-((2S,3R)-1-(3-Chloro-5-ﬂuorophenyl)-3-hydroxy-4-((S)-6′-

neopentyl-3′,4′-dihydrospiro[cyclobutane-1,2′-pyrano[2,3-b]-

pyridine]-4′-ylamino)butan-2-yl)-2-methoxyacetamide (9b).

The title compound was prepared from 7b via a method analogous

to the preparation of 9a, giving a white solid (0.64 g, 85% yield).

N-((2S,3R)-1-(3-Chloro-4-ﬂuorophenyl)-3-hydroxy-4-((S)-6′-

neopentyl-3′,4′-dihydrospiro[cyclobutane-1,2′-pyrano[2,3-b]-

pyridine]-4′-ylamino)butan-2-yl)-2-methoxyacetamide (9c).

The title compound was prepared from 7c via a method analogous

to the preparation of 9a, giving a white solid (0.87 g, 64% yield). MS

m/z = 548.2 [M + H]⁺. Calcd for C₂₉H₄₀ClFN₃O₄: 548.3.

silica gel with the aid of ethyl acetate. The ﬁltrate was concentrated,

and the residue was puriﬁed by chromatography on silica gel (10−60%

EtOAc/hexane) to give tert-butyl (2S,3R)-3-(tert-butyldimethylsily-

loxy)-1-(2-ﬂuoro-3-((trimethylsilyl)ethynyl)phenyl)-4-((S)-6′-neopen-

tyl-3′,4′-dihydrospiro[cyclobutane-1,2′-pyrano[2,3-b]pyridine]-4′-

1

ylamino)butan-2-ylcarbamate as a white solid (3.67 g, 85% yield). H

NMR (400 MHz, acetonitrile-d₃) δ ppm 7.81 (d, J = 2.15 Hz, 1 H),

7.65 (s, 1 H), 7.19−7.33 (m, 2 H), 7.06 (t, J = 7.63 Hz, 1 H), 5.42 (d, J

= 9.39 Hz, 1 H), 3.96−4.09 (m, 1 H), 3.78−3.96 (m, 2 H), 2.86−3.06

(m, 2 H), 2.68−2.83 (m, 1 H), 2.63 (t, J = 13.01 Hz, 1 H), 2.32−2.42

(m, 2 H), 2.20−2.32 (m, 1 H), 2.04−2.20 (m, 2 H), 1.84−1.93 (m, 1

H), 1.69−1.82 (m, 2 H), 1.23−1.32 (m, 9 H), 0.90−0.98 (m, 9 H),

0.83−0.90 (m, 9 H), 0.20−0.30 (m, 9 H), 0.10 (s, 3 H), 0.09 (s, 3 H).

tert-Butyl (2S,3R)-3-(tert-Butyldimethylsilyloxy)-4-((S)-6′-

neopentyl-3′,4′-dihydrospiro[cyclobutane-1,2′-pyrano[2,3-b]-

pyridine]-4′-ylamino)-1-(4-((trimethylsilyl)ethynyl)phenyl)-

butan-2-ylcarbamate (10i). The title compound was prepared from

7e via a method analogous to the preparation of 10h, giving a white

solid (0.30 g, 93% yield). ¹H NMR (400 MHz, acetonitrile-d₃) δ ppm

7.81 (br s, 1 H), 7.54−7.68 (m, 1 H), 7.33 (d, J = 8.02 Hz, 2 H), 7.21

(d, J = 8.12 Hz, 2 H), 5.57 (d, J = 10.27 Hz, 1 H), 3.92−4.03 (m, 1 H),

3.79−3.92 (m, 2 H), 2.84−2.98 (m, 2 H), 2.67−2.75 (m, 1 H), 2.57−

2.67 (m, 1 H), 2.31−2.40 (m, 3 H), 2.22−2.31 (m, 1 H), 2.14−2.23

(m, 3 H), 1.83−1.92 (m, 1 H), 1.77 (d, J = 2.54 Hz, 2 H), 1.30 (s, 9

H), 0.90 (s, 9 H), 0.88 (s, 9 H), 0.22 (s, 9 H), 0.04−0.09 (m, 6 H). m/

z = 734.4 [M + H]⁺. Calcd for C₄₂H₆₈N₃O₄Si₂: 734.5.

N-((2S,3R)-3-Hydroxy-4-((S)-6′-neopentyl-3′,4′-dihydrospiro-

[cyclobutane-1,2′-pyrano[2,3-b]pyridine]-4′-ylamino)-1-(3-

(pyridin-4-yl)phenyl)butan-2-yl)-2-methoxyacetamide (11). A

solution of N-((2S,3R)-1-(3-bromophenyl)-3-hydroxy-4-((S)-6′-neo-

pentyl-3′,4′-dihydrospiro[cyclobutane-1,2′-pyrano[2,3-b]pyridine]-4′-

ylamino)butan-2-yl)-2-methoxyacetamide (9a) (0.30 g, 0.52 mmol), 4-

(tributylstannyl)pyridine (0.21 g, 0.57 mmol), and tetrakis-

(tripheynlphosphine)palladium (0.12 g, 0.10 mmol) in 1,4-dioxane

(5 mL) was heated to 140 °C for 30 min in a microwave reactor. The

reaction mixture was cooled, stirred with 10% aq potassium ﬂuoride

solution (10 mL), and extracted with EtOAc (3 × 10 mL). The

organic layers were combined, and the combined solution was washed

with brine, dried over magnesium sulfate, ﬁltered, and concentrated.

The residue was puriﬁed by reverse-phase chromatography (10−90%

CH₃CN/H₂O with 0.l% TFA). Pure fractions containing product were

combined in saturated aq sodium bicarbonate solution and extracted

with DCM. The organic extract was dried over sodium sulfate, ﬁltered,

and evaporated to give N-((2S,3R)-3-hydroxy-4-((S)-6′-neopentyl-

3′,4′-dihydrospiro[cyclobutane-1,2′-pyrano[2,3-b]pyridine]-4′-ylami-

no)-1-(3-(pyridin-4-yl)phenyl)butan-2-yl)-2-methoxyacetamide (11)

(0.105 g, 35% yield) as a white solid. ¹H NMR (400 MHz,

acetonitrile-d₃) δ ppm 8.61 (d, J = 6.16 Hz, 2 H), 7.79 (d, J = 2.25

Hz, 1 H), 7.70 (dd, J = 2.40, 1.03 Hz, 1 H), 7.54−7.61 (m, 4 H), 7.41

(dt, J = 1.00, 7.62 Hz, 1 H), 7.33 (dt, J = 7.53, 1.00 Hz, 1 H), 6.93 (d, J

= 9.39 Hz, 1 H), 4.11−4.25 (m, 1 H), 3.91 (dd, J = 10.51, 5.53 Hz, 1

H), 3.74 (dd, J = 15.16 Hz, 1 H), 3.60−3.65 (m, 1 H), 3.58 (dd, J =

15.20 Hz, 1 H), 3.14−3.23 (m, 3 H), 2.76−2.89 (m, 2 H), 2.67−2.75

(m, 1 H), 2.32−2.47 (m, 4 H), 2.06−2.26 (m, 6 H), 1.81−1.91 (m, 1

H), 1.61−1.78 (m, 2 H), 0.86 (s, 9 H). HRMS (ESI+) m/z =

573.3427 [M + H]⁺. Calcd for C₃₄H₄₄N₄O₄: 573.3425.

N-((2S,3R)-1-(5-Chloro-2-ﬂuorophenyl)-3-hydroxy-4-((S)-6′-

neopentyl-3′,4′-dihydrospiro[cyclobutane-1,2′-pyrano[2,3-b]-

pyridine]-4′-ylamino)butan-2-yl)-2-methoxyacetamide (9d).

The title compound was prepared from 7d via a method analogous

to the preparation of 9a, giving a yellow glass (1.2 g, 61% yield). MS

m/z = 548.1 [M + H]⁺. Calcd for C₂₉H₄₀ClFN₃O₄: 548.3.

N-((2S,3R)-1-(4-Chlorophenyl)-3-hydroxy-4-((S)-6′-neopen-

tyl-3′,4′-dihydrospiro[cyclobutane-1,2′-pyrano[2,3-b]-

pyridine]-4′-ylamino)butan-2-yl)-2-methoxyacetamide (9e).

The title compound was prepared from 7e via a method analogous

to the preparation of 9a, giving a light-yellow solid. (0.81 g, 70% yield).

MS m/z = 530.2 [M + H]⁺. Calcd for C₂₉H₄₁ClN₃O₄: 530.3.

tert-Butyl (2S,3R)-3-(tert-Butyldimethylsilyloxy)-1-(2-ﬂuoro-

3-(thiazol-2-yl)phenyl)-4-((S)-6′-neopentyl-3′,4′-dihydrospiro-

[cyclobutane-1,2′-pyrano[2,3-b]pyridine]-4′-ylamino)butan-2-

ylcarbamate (10g). A solution of tert-butyl (2S,3R)-3-(tert-

butyldimethylsilyloxy)-1-(3-chloro-2-ﬂuorophenyl)-4-((S)-6′-neopen-

tyl-3′,4′-dihydrospiro[cyclobutane-1,2′-pyrano[2,3-b]pyridine]-4′-

ylamino)butan-2-ylcarbamate (7f) (3.97 g, 5.75 mmol), dicyclohexyl-

(2′,4′,6′-triisopropylbiphenyl-2-yl)phosphine (1.64 g, 3.45 mmol), and

Pd₂(dba)₃(0.79 g, 0.86 mmol) in THF (15 mL) was degassed by

sparging with N₂(g) for 15 min. 2-(Tributylstannyl)thiazole (3.23 g,

8.63 mmol) was added, and the resulting mixture was heated to 135

°C for 40 min in a microwave reactor. The reaction mixture was

cooled and ﬁltered through a plug of silica gel with the aid of ethyl

acetate. The ﬁltrate was concentrated, and the residue was puriﬁed by

chromatography on silica gel (10−60% EtOAc/hexane) to give tert-

butyl (2S,3R)-3-(tert-butyldimethylsilyloxy)-1-(2-ﬂuoro-3-(thiazol-2-

yl)phenyl)-4-((S)-6′-neopentyl-3′,4′-dihydrospiro[cyclobutane-1,2′-

pyrano[2,3-b]pyridine]-4′-ylamino)butan-2-ylcarbamate (10g) as a

1

white solid (3.90 g, 92% yield). H NMR (400 MHz, acetonitrile-

d₃) δ ppm 8.04−8.17 (m, 1 H), 7.87−7.98 (m, 1 H), 7.80 (s, 1 H),

7.66 (s, 1 H), 7.63 (d, J = 3.33 Hz, 1 H), 7.36 (t, J = 8.17 Hz, 1 H),

7.22 (t, J = 7.68 Hz, 1 H), 5.51 (d, J = 8.61 Hz, 1 H), 4.01−4.25 (m, 1

H), 3.84−3.97 (m, 2 H), 3.07 (dd, J = 14.33, 4.65 Hz, 1 H), 2.94 (dd, J

= 12.62, 3.91 Hz, 1 H), 2.69−2.82 (m, 2 H), 2.33−2.41 (m, 2 H),

2.21−2.33 (m, 1 H), 2.07−2.21 (m, 2 H), 1.84−1.93 (m, 1 H), 1.69−

1.82 (m, 1 H), 1.60 (br s, 1 H), 1.22 (s, 9 H), 0.92 (s, 9 H), 0.85 (s, 9

H), 0.10 (s, 3 H), 0.09 (s, 3 H). MS m/z = 739.2 [M + H]⁺. Calcd for

C₄₀H₆₀FN₄O₄SSi: 739.4.

tert-Butyl (2S,3R)-3-(tert-Butyldimethylsilyloxy)-1-(2-ﬂuoro-

3-((trimethylsilyl)ethynyl)phenyl)-4-((S)-6′-neopentyl-3′,4′-

dihydrospiro[cyclobutane-1,2′-pyrano[2,3-b]pyridine]-4′-

ylamino)butan-2-ylcarbamate (10h). A solution of tert-butyl

(2S,3R)-3-(tert-butyldimethylsilyloxy)-1-(3-chloro-2-ﬂuorophenyl)-4-

((S)-6′-neopentyl-3′,4′-dihydrospiro[cyclobutane-1,2′-pyrano[2,3-b]-

pyridine]-4′-ylamino)butan-2-ylcarbamate (7f) (3.97 g, 5.75 mmol),

dicyclohexyl(2′,4′,6′-triisopropylbiphenyl-2-yl)phosphine (1.64 g, 3.45

mmol), and Pd₂(dba)₃(0.79 g, 0.86 mmol) in THF (15 mL) was

degassed by sparging with N₂(g) for 15 min. Trimethyl(2-

(tributylstannyl)ethynyl)silane (3.34 g, 8.63 mmol) was added, and

the resulting mixture was heated to 135 °C for 40 min in a microwave

reactor. The reaction mixture was cooled and ﬁltered through a plug of

N-((2S,3R)-3-Hydroxy-4-((S)-6′-neopentyl-3′,4′-dihydrospiro-

[cyclobutane-1,2′-pyrano[2,3-b]pyridine]-4′-ylamino)-1-(3-

(pyridin-2-yl)phenyl)butan-2-yl)-2-methoxyacetamide (12).

The title compound was prepared from 9a and 2-(tributylstannyl)-

pyridine via a method analogous to the preparation of 12, giving a

white solid (0.088 g, 32% yield). ¹H NMR (300 MHz, chloroform-d) δ

ppm 8.69 (dt, J = 4.75, 1.35 Hz, 1 H), 7.82−7.99 (m, 3 H), 7.64−7.80

(m, 2 H), 7.56 (d, J = 1.61 Hz, 1 H), 7.42 (t, J = 7.60 Hz, 1 H), 7.18−

7.35 (m, 2 H), 6.62 (d, J = 9.06 Hz, 1 H), 4.09−4.28 (m, J = 8.51, 8.20,

8.20, 4.46 Hz, 1 H), 3.89−3.99 (m, 1 H), 3.66−3.89 (m, 2 H), 3.47−

3.62 (m, 1 H), 3.16−3.29 (m, 4 H), 2.94−3.07 (m, 1 H), 2.83 (dd, J =

12.35, 4.75 Hz, 1 H), 2.65 (dd, J = 12.42, 3.95 Hz, 1 H), 2.54 (t, J =

10.23 Hz, 1 H), 2.41 (s, 2 H), 2.33 (dd, J = 13.15, 5.55 Hz, 1 H),

2.12−2.28 (m, 1 H), 2.06 (ddd, J = 7.71, 4.24, 3.98 Hz, 1 H), 1.87−

9037

dx.doi.org/10.1021/jm300118s | J. Med. Chem. 2012, 55, 9025−9044

Journal of Medicinal Chemistry

Article

1.99 (m, 1 H), 1.60−1.86 (m, 3 H), 0.89 (s, 9 H). HRMS (ESI+) m/z

= 573.3439 [M + H]⁺. Calcd for C₃₄H₄₅N₄O₄: 573.3425.

N-((2S,3R)-3-Hydroxy-4-((S)-6′-neopentyl-3′,4′-dihydrospiro-

[cyclobutane-1,2′-pyrano[2,3-b]pyridine]-4′-ylamino)-1-(3-

(thiazol-5-yl)phenyl)butan-2-yl)-2-methoxyacetamide (16).

The title compound was prepared from 9a and 5-thiazolyltributyl-

stannane via a method analogous to the preparation of 14, giving a

white solid (0.16 g, 55% yield). ¹H NMR (400 MHz, acetonitrile-d₃) δ

ppm 7.86 (t, J = 1.57 Hz, 1 H), 7.85 (d, J = 3.23 Hz, 1 H), 7.79−7.83

(m, 2 H), 7.73−7.75 (m, 1 H), 7.50 (d, J = 3.33 Hz, 1 H), 7.38 (t, J =

7.58 Hz, 1 H), 7.30−7.34 (m, 1 H), 6.96 (d, J = 9.68 Hz, 1 H), 4.10−

4.20 (m, 1 H), 4.00−4.09 (m, 1 H), 3.74 (d, J = 15.26 Hz, 1 H), 3.64−

3.71 (m, 1 H), 3.59 (d, J = 15.26 Hz, 1 H), 3.16−3.23 (m, 4 H), 2.78−

2.88 (m, 2 H), 2.36−2.49 (m, 3 H), 2.07−2.26 (m, 5 H), 1.84 (s, 1 H),

1.67−1.79 (m, 2 H), 1.18 (s, 2 H), 0.87 (s, 9 H). HRMS (ESI+) m/z =

579.3001 [M + H]⁺. Calcd for C₃₂H₄₃N₄O₄S: 579.2989.

N-((2S,3R)-3-Hydroxy-4-((S)-6′-neopentyl-3′,4′-dihydrospiro-

[cyclobutane-1,2′-pyrano[2,3-b]pyridine]-4′-ylamino)-1-(3-

(thiazol-4-yl)phenyl)butan-2-yl)-2-methoxyacetamide (17).

The title compound was prepared from 9a and 4-(tributylstannyl)-

thiazole via a method analogous to the preparation of 14, giving a

white solid (0.078 g, 44% yield). ¹H NMR (400 MHz, acetonitrile-d₃)

δ ppm 7.87 (t, J = 1.52 Hz, 1 H), 7.77−7.82 (m, 3 H), 7.75 (d, J = 1.96

Hz, 1 H), 7.71 (dd, J = 2.35, 0.98 Hz, 1 H), 7.35 (t, J = 7.68 Hz, 1 H),

7.22 (ddd, J = 7.80, 1.34, 1.12 Hz, 1 H), 6.93 (d, J = 9.49 Hz, 1 H),

4.11−4.20 (m, 1 H), 3.93 (dd, J = 10.32, 5.43 Hz, 1 H), 3.72 (d, J =

15.16 Hz, 1 H), 3.61−3.65 (m, 1 H), 3.52 (d, J = 15.20 Hz, 1 H),

3.14−3.21 (m, 4 H), 2.78−2.85 (m, 2 H), 2.70−2.76 (m, 1 H), 2.35−

2.46 (m, 4 H), 2.07−2.25 (m, 5 H), 1.80−1.91 (m, 1 H), 1.62−1.78

(m, 2 H), 0.87 (s, 9 H). HRMS (ESI+) m/z = 579.2998 [M + H]⁺.

Calcd for C₃₂H₄₃N₄O₄S: 579.2989.

N-((2S,3R)-3-Hydroxy-4-((S)-6′-neopentyl-3′,4′-dihydrospiro-

[cyclobutane-1,2′-pyrano[2,3-b]pyridine]-4′-ylamino)-1-(3-(ox-

azol-2-yl)phenyl)butan-2-yl)-2-methoxyacetamide (13). A sol-

ution of N-((2S,3R)-1-(3-bromophenyl)-3-hydroxy-4-((S)-6′-neopen-

tyl-3′,4′-dihydrospiro[cyclobutane-1,2′-pyrano[2,3-b]pyridine]-4′-

ylamino)butan-2-yl)-2-methoxyacetamide (9a) (0.28 g, 0.50 mmol), 2-

(tributylstannyl)oxazole (0.20 g, 0.55 mmol), and tetrakis-