Optimization of brain penetrant 11β-hydroxysteroid dehydrogenase type i inhibitors and in vivo testing in diet-induced obese mice

Article

pubs.acs.org/jmc

Optimization of Brain Penetrant 11β-Hydroxysteroid Dehydrogenase

Type I Inhibitors and in Vivo Testing in Diet-Induced Obese Mice

Frederick W. Goldberg,* Alexander G. Dossetter,* James S. Scott, Graeme R. Robb, Scott Boyd,

Sam D. Groombridge, Paul D. Kemmitt, Tove Sjogren, Pablo Morentin Gutierrez,

̈

Joanne deSchoolmeester, John G. Swales, Andrew V. Turnbull, and Martin J. Wild

AstraZeneca, Mereside, Alderley Park, Macclesﬁeld SK10 4TG, United Kingdom

S

* Supporting Information

ABSTRACT: 11β-Hydroxysteroid dehydrogenase type 1 (11β-HSD1) has been widely considered by the pharmaceutical

industry as a target to treat metabolic syndrome in type II diabetics. We hypothesized that central nervous system (CNS)

penetration might be required to see eﬃcacy. Starting from a previously reported pyrimidine compound, we removed hydrogen-

bond donors to yield 3, which had modest CNS penetration. More signiﬁcant progress was achieved by changing the core to give

40, which combines good potency and CNS penetration. Compound 40 was dosed to diet-induced obese (DIO) mice and gave

excellent target engagement in the liver and high free exposures of drug, both peripherally and in the CNS. However, no body

weight reduction or eﬀects on glucose or insulin were observed in this model. Similar data were obtained with a structurally

diverse thiazole compound 51. This work casts doubt on the hypothesis that localized tissue modulation of 11β-HSD1 activity

alleviates metabolic syndrome.

published, with the primary focus on treatment of metabolic

syndrome.^10,11Incyte has taken an undisclosed compound as

far as phase II trials in patients with type II diabetes.¹²Doses of

200 mg resulted in small reductions in glycated hemoglobin

(hBA1c, −0.6%), fasting glucose (−24 mg/dL), and reduction

in homeostatic model assessment−insulin resistance (HOMA-

IR, −24%) compared with placebo. Merck has progressed two

compounds into development: 3-[1-(4-chlorophenyl)-3-ﬂuo-

rocyclobutyl]-4,5-dicyclopropyl-1,2,4-triazole (MK-0916),

which entered phase II in 2004, and 3-[1-(3-ethylsulfonyl-

propyl)-4-bicyclo[2.2.2]octanyl]-4-methyl-5-[2-

(triﬂuoromethyl)phenyl]-1,2,4-triazole (MK-0736), which was

originally developed as a treatment for hypertension and

entered phase II trials in 2005.¹³Initial clinical data have now

been reported showing no signiﬁcant improvement in fasting

plasma glucose at week 12 relative to placebo. Modest dose-

dependent decreases in blood pressure and body weight were

observed over the course of the study, together with a small but

signiﬁcant reduction of 0.3% in hBA1c at week 12.¹⁴

Collectively these data indicate that only high doses of 11β-

INTRODUCTION

■

11β-Hydroxysteroid dehydrogenase type 1 (11β-HSD1) is a

NADPH-dependent enzyme that is widely expressed, notably in

liver, adipose tissue, and brain.¹It catalyzes conversion of the

inactive glucocorticoid hormone cortisone to the active

glucocorticoid hormone cortisol and therefore plays a key

role in the regulation of intracellular cortisol concentrations.²

Although patients with metabolic syndrome do not exhibit

elevated plasma glucocorticoid levels,³it has been hypothesized

that elevated intracellular concentrations may play a crucial

role, and reduction of these has been proposed as an attractive

therapeutic paradigm.⁴⁻⁶Elevated cortisol is also associated

with cognitive dysfunction and neurotoxicity, and glucocorti-

coids increase amyloid Aβ in models of Alzheimer’s disease

(AD). 11β-HSD1 is highly expressed in brain regions important

for cognition and is increased in animal models of AD.

Transgenic overexpression of 11β-HSD1 in brain exacerbates

age-associated cognitive dysfunction, while 11β-HSD1 knock-

out mice are protected.^7,8Proof of concept has been obtained

in humans by use of the nonselective prototype 11β-HSD

inhibitor carbenoxolone.⁹

Over 250 11β-HSD1 patent applications from more than 25

pharmaceutical companies and academic groups have now been

Received: October 28, 2013

Published: January 14, 2014

970

dx.doi.org/10.1021/jm4016729 | J. Med. Chem. 2014, 57, 970−986

Journal of Medicinal Chemistry

Article

Figure 1. Structures of potential start points.

Figure 2. Optimization of pyrimidine-based 11β-HSD1 inhibitors for CNS properties. ^ah denotes human isoform; m denotes murine isoform. IC₅₀

data are the geometric mean of at least three independent measurements unless stated, with 68% conﬁdence limits in parentheses as calculated from

b

standard error of the mean (SEM). Dosed individually (HPMC/Tween suspension for po and solution for iv) to fed male Han Wistar rats.

HSD1 inhibitors improve glycemic control in humans and do

so only to a modest extent.

with good pharmacokinetics (PK) and are potent against both

human and murine isoforms.^15,21

Our own experience with 11β-HSD1 inhibitors in the

preclinical setting¹⁵mirrors the clinical work and indicates that

only very high doses/exposures of 11β-HSD1 inhibitors

provide eﬃcacy. We hypothesized that modest central nervous

system (CNS) penetration may account for the limited eﬃcacy

and high doses required. 11β-HSD1 is expressed in areas of the

brain relevant to metabolic control,¹⁶and glucocorticoids are

known to have eﬀects on neuropeptides in the hypothalamus,

including proopiomelanocortin (POMC) and neuropeptide Y

(NPY), which are crucially involved in food intake and have

also been shown to control glucose levels via the vagal

nerve.^17,18

Our initial strategy to improve CNS penetration was to

reduce the number of hydrogen-bond donors present in 1 with

a view to reducing the levels of eﬄux as measured by a Madin−

Darby canine kidney (MDCK) assay.²²Replacement of the

amino-tetrahydrofuran (THF) at C2 with a 3,5-cis-dimethyl-

substituted morpholine 2 (Figure 2) retained human potency

but was signiﬁcantly less potent in mouse, in accordance with

previous structure−activity relationships (SAR) on this series.¹⁵

The favorable DMPK properties in rat were retained despite a

signiﬁcant increase in log D_7.4(+1.2) associated with this

change. To lower the lipophilicity, we investigated ether

substitution into the 5-membered ring and found that the 2-

(R)-tetrahydrofuryl²³analogue 3 lowered log D_7.4(−1.3) while

maintaining human potency and improving mouse potency to a

level comparable to 1. This compound has two hydrogen-bond

donors with one potentially “masked” through formation of a 7-

membered internal hydrogen bond between the amide NH and

the ether, whereas 1 has three hydrogen-bond donors. It is

noteworthy that the ¹H NMR shift in CDCl₃of the amide NH

in 3 is shifted downﬁeld (δ_H= 7.90 ppm) relative to 2 (δ_H=

5.83 ppm), consistent with formation of an internal hydrogen

bond. MDCK permeability of 2 and 3 was notably improved

relative to 1 with no signiﬁcant eﬄux observed, presumably as a

consequence of the reduced hydrogen-bond donor count.²⁴

RESULTS AND DISCUSSION

■

To explore the hypothesis that 11β-HSD1 inhibition in the

brain (in addition to peripheral tissues) may cause eﬀects on

body weight or glycemic control, we required tool compounds

with good CNS exposure. Acidic compounds such as 2-[(3S)-1-

[5-(cyclohexylcarbamoyl)-6-propylsulfanyl-2-pyridyl]-3-

piperidyl]acetic acid (AZD4017, Figure 1)¹⁹and 4-[4-(2-

adamantylcarbamoyl)-5-tert-butylpyrazol-1-yl]benzoic acid

(AZD8329)²⁰were considered unlikely to deliver a brain

penetrant proﬁle, so we instead chose neutral pyrimidines such

as 1 as potential start points, as they are neutral compounds

971

dx.doi.org/10.1021/jm4016729 | J. Med. Chem. 2014, 57, 970−986

Journal of Medicinal Chemistry

Article

Figure 3. Changing the core to identify new CNS penetrant lead compounds. ^ah denotes human isoform; m denotes murine isoform. IC₅₀data are

the geometric mean of at least three independent measurements unless stated, with 68% conﬁdence limits in parentheses as calculated from the

b

c

d

SEM. n = 2. Dosed individually (HPMC/Tween suspension for po and solution for iv) to fed male Han Wistar rats. Ratio of the mean (n = 2)

drug concentrations in homogenized brain tissue and whole blood, 1 h after a 30 mg/kg dose to male Han Wistar rats.

Compound 3 achieved moderate CNS exposure (free ratio

0.24), although the molecular mass (456 Da) and polar surface

area (PSA; 94 Å², not accounting for intramolecular hydrogen

bond) were high compared to those of known neutral CNS-

penetrant drugs.²⁵Eﬀorts to reduce donor count further by

changing the 1-hydroxyladamantyl group were unsuccessful, as

they led to signiﬁcant reductions in potency and/or poor PK

properties. We therefore attempted to identify a new core that

might allow us to reduce molecular mass while maintaining

good potency against both human and murine isoforms. We

synthesized a range of core heterocycles, ﬁxing the 1-

hydroxyadamantyl from pyrimidine 3 and maintaining an

ortho group to the amide (Figure 3), as our experience had

shown that this was required for potency. A pragmatic high-

throughput approach was taken to assess CNS penetration,

where compounds with acceptable human and murine potency

in a homogeneous time-resolved ﬂuorescence (HTRF) assay

and rat PK in cassette were assessed for CNS penetration

individually, by orally dosing 30 mg/kg of compound to rats

and assessing the brain/blood ratio at a ﬁxed 1 h time point.

Rat was chosen as the species to assess CNS penetration as

blood contamination of brain samples can be an issue when

CNS penetration is assessed in mouse. Free [brain]/[blood]

ratios could then be determined by correcting the total ratio

with in vitro tests for plasma protein binding and brain slice

tissue binding. All rat PK measurements in this article are taken

from whole blood (not plasma), as compounds from this series

were observed to partition unequally between red blood cells

and plasma.

[brain]/[blood] ratio using the above method was poor

(0.1), albeit at very low log D_7.4(0.4). Other pyrazoles we

synthesized (with higher log D) gave poor oral exposure, so we

investigated alternative heterocycles. This work identiﬁed only

one compound, 5, that showed excellent CNS penetration

(∼1:1 free ratio at 1 h time point), with good potency against

both human and murine isoforms despite its low molecular

weight (322). Thiazole 5 was therefore selected as a suitable

lead compound for further optimization. Thiazole 5 had good

physicochemical properties (log D_7.4= 2.0, solubility >2 mM),

high free levels (Hu ppb = 55% free, brain tissue binding = 40%

free), and moderate oral bioavailability (23%) and clearance

(22 mL·min⁻¹·kg⁻¹) in rat. Our priority was therefore to

improve human/murine potency and oral exposure of 5 while

maintaining the excellent physicochemical properties (maintain

log D_7.4between 1 and 3).

Pyrimidines 1−3 were synthesized according to previously

published procedures and general methods.^15,21Pyrazole 4 was

synthesized by an amide coupling between (1S,4R)-4-amino-

adamantan-1-ol and 2-chloropyrazole 6, followed by methoxide

displacement of the chloride (Scheme 1). Thiazoles 5 and 15−

51 were typically synthesized by a late-stage coupling of a

thiazole acid with both R²and R⁴groups in place, formed from

saponiﬁcation of the corresponding ester (Scheme 2). The

route used to synthesize the ester precursor varied, depending

a

Scheme 1. Synthesis of Pyrazole 4

f

We started by synthesizing pyrazole 4 (Figure 3), given the

precedent of this ring system from a previous acidic series²⁰and

the knowledge from our previous work that alkoxy substituents

can achieve good human and murine potency.¹⁵Compound 4

had good human/murine potency and suitable PK for

assessment of CNS penetration, but the measured free

a

Reagents and conditions: (i) (1S,4R)-4-Aminoadamantan-1-ol hydro-

chloride, HATU, DMF, i-PrNEt₂, 68%. (ii) NaOMe, THF, 54%.

972

dx.doi.org/10.1021/jm4016729 | J. Med. Chem. 2014, 57, 970−986

Journal of Medicinal Chemistry

Article

a

Scheme 2. General Synthetic Routes to Vary R²and R⁴Groups on the Thiazole

a

Reagents and conditions: (i) Toluene or EtOH, 25−88%. (ii) NaOH, EtOH or KOSi(CH₃)₃, THF, 64−100%. (iii) (1S,4R)-4-Aminoadamantan-1-

ol hydrochloride, HATU, DMF or CH₂Cl₂, i-PrNEt₂, 31−91%. (iv) DIAD, THF, PPh₃, ROH, THF or RI, K₂CO₃, DMF, 26−99%. (v) t-BuONO,

THF, HCl, 1,4-dioxane, 68−78%. (vi) NaOR, ROH or THF, 48−85%

Table 1. Optimizing the R⁴Group

[brain]/ [bl-

a

b

IC₅₀, nM (68% CL)

ood]

d

h-11β-HSD1,

m-11β-HSD1,

h-11β-HSD1,

log

D_7.4

aq solubility,

PSA,

Å²

c

R²

R⁴

HTRF

HPLC

LLE

μM

total free

5

Me

OMe

61

203

28

5.2

2.0

1.7

>2000

2100

1.2

0.9

75

e

(52−72)

142

(163−252)

304

(26−30)

15

H

75

(126−160)

>20500

(266−347)

>24400

e

16

17

H

<3.9

3.8

0.8

1.3

>2900

>1500

66

Me

8600

6910

(7110−10400)

29

(4810−9920)

179

18

19

20

21

22

23

24

25

26

27

Me

OEt

4.9

4.5

4.9

4.8

5.5

4.8

4.2

5.3

4.9

5.8

2.6

3.1

3.0

3.2

1.5

3.1

1.5

1.9

1.8

170

210

75

84

66

74

(28−30)

23

(141−228)

637

OCH₂CF₃

O-i-Pr

8.3

(18−29)

13

(562−722)

364

(7.9−8.8)

65

(11−15)

10

(344−386)

125

O-c-Bu

64

(9−11)

104

(114−138)

4810

O-(3-oxetenyl)

S-n-Pr

1100

500

(95−113)

14

(3390−6820)

1040

(11−18)

2090

(908−1200)

9350

CF₃

610

(1730−2530)

69

(7870−11100)

f

c-Pr

395

920

1.6

1.5

1.2

2.9

(65−74)

216

(S)-2-THF

(R)-2-THF

11600

>2100

(193−242)

24

(8430−16000)

364

3.6

(19−29)

(189−700)

(2.6−4.9)

a

h denotes human isoform; m denotes murine isoform. IC₅₀data are the geometric mean of at least three independent measurements unless stated,

with 68% conﬁdence limits in parentheses as calculated from SEM. Ratio of mean (n = 2) drug concentrations in homogenized brain tissue and

whole blood, 1 h after a 30 mg/kg dose to male Han Wistar rats. Ligand lipophilicity eﬃciency, deﬁned as human HTRF pIC₅₀− log D_7.4.

Aqueous solubility measured at pH 7.4, performed under thermodynamic conditions from solid samples. n = 2. n = 1.

b

c

d

e

f

on the nature of the R²and R⁴groups. Where R²and R⁴were

both alkyl groups, classic Hantzsch ring synthesis between a 2-

chloro ketoester 7 and a thioamide was used to give ester 8,

which could then be saponiﬁed and coupled with (1S,4R)-4-

973

dx.doi.org/10.1021/jm4016729 | J. Med. Chem. 2014, 57, 970−986

Journal of Medicinal Chemistry

Article

Table 2. Optimizing the R²Position, with R⁴Fixed as either (R)-2-THF or Cyclopropyl

[brain]/

[blood]

a

b

IC₅₀, nM (68% CL)

d

h-11β-HSD1,

HTRF

m-11β-HSD1,

h-11β-HSD1,

HPLC

log

D_7.4

aq solubility,

rat F (%)/rat CL

PSA,

c

e

R²

HTRF

LLE

μM

(mL·min⁻¹·kg⁻¹

)

total/free

Å²

R⁴= THF

27 Me

24

364

3.6

5.8

5.0

6.4

1.8

2.4

1.6

>2100

1600

11/17

1.5/2.9

74

(19−29)

43

(189−700)

668

(2.6−4.9)

28 Et

(36−51)

10

(604−738)

360

29

H

>2600

0/52

(8−13)

48

(329−393)

3260

30 CH₂OMe

31 CH₂CH₂OMe

32 NH₂

(34−66)

50

(2820−3760)

553

5.4

1.9

>7500

83

(40−62)

47

(493−621)

964

5.5

5.6

1.8

1.7

83

710

560

102

(36−61)

45

(878−1060)

1210

33 NHMe

5.3

2.1

89

(36−56)

230

(1060−1370)

3980

34 NMe₂

(206−255)

486

(2690−5880)

2480

3.9

5.9

2.7

0.4

140

75

35 CH₂SO₂Me

36 OMe

1300

65/17

40/21

24/8

<0.05/<0.05

1.1/0.7

111

(358−658)

7.2

(1920−3210)

34

0.40

5.8

5.1

2.3

2.1

2.8

490

980

84

92

f

(6.5−8.0)

12

(30−37)

9.9

(0.30−0.53)

0.49

37 OCH₂CH₂OMe

0.8/1.0

f

(10−15)

13

(9.7−10.1)

8.9

(0.40−0.60)

0.45

38 (S)-OCH(Me)

16/13

0.9/0.5

CH₂OMe

f

(11−15)

7.2

(6.5−12.2)

11.8

(0.40−0.50)

39 O-(3-CN)-c-butyl

40 O-4-THP

5.7

5.2

2.4

2.6

120

300

102

92

(6.7−7.6)

17

(8.9−15.6)

14

1.7

83/34

1.1/0.9

f

(13−24)

(10−19)

(0.7−4.0)

R⁴= c-Pr

25 Me

g

69

390

5.3

5.2

4.6

6.4

5.6

4.9

1.9

2.6

2.9

3.4

1.6

2.2

2.6

920

570

140

33

12/14

45/15

1.6/1.2

1.8/0.7

66

75

97

84

(65−74)

14

41 OMe

53

4.4

f

(12−16)

9.1

(49−57)

34

(4.398−4.44)

1.7

42 OEt

f

(8.6−9.6)

10

(30−39)

12

(1.5−2.0)

1.5

43 O-i-Pr

(9−12)

9.9

(11−13)

13

(1.1−2.0)

44 OCH₂CH₂OH

45 OCH₂CH₂OMe

200

750

2700

25/5

0.2/0.1

1.2/0.6

(8.9−11.0)

17

(11−16)

23

2.8

23/27

f

(16−19)

30

(21−26)

14

(2.7−3.0)

9.0

46 (R)-OCH(Me)

CH₂OCH₃

(27−33)

(11−18)

(5.5−14.8)

47 (S)-OCH(Me)

12

8.8

1.3

5.1

2.8

960

84

CH₂OMe

f

(11−14)

12

(7.0−11.0)

19

(0.9−2.0)

1.4

48 O-(R)-2-THF

49 O-4-THP

5.5

5.0

2.4

2.7

260

180

84

f

(11−14)

18

(16−23)

4.8

(1.0−2.0)

6.9

53/21

1.2/0.6

974

dx.doi.org/10.1021/jm4016729 | J. Med. Chem. 2014, 57, 970−986

Journal of Medicinal Chemistry

Article

Table 2. continued

[brain]/

[blood]

a

b

IC₅₀, nM (68% CL)

d

h-11β-HSD1,

HTRF

m-11β-HSD1,

h-11β-HSD1,

HPLC

log

D_7.4

aq solubility,

rat F (%)/rat CL

PSA,

c

e

R²

HTRF

LLE

μM

(mL·min⁻¹·kg⁻¹

)

total/free

Å²

R⁴= c-Pr

f

(15−22)

9.1

(4.1−5.7)

5.4

(5.1−9.5)

0.43

50 OCH₂-c-Pr

4.4

5.1

3.6

2.6

10

12

75

93

(8.3−9.9)

22

(4.7−6.1)

13

(0.40−0.46)

8.1

51 OCH₂-(1-CN)-c-Pr

79/16

0.4/0.4

f

(19−25)

(8−21)

(4.0−16.5)

a

h denotes human isoform; m denotes murine isoform. IC₅₀data are the geometric mean of at least three independent measurements unless stated,

b

with 68% conﬁdence limits in parentheses as calculated from SEM. Ratio of mean (n = 2) drug concentrations in homogenized brain tissue and

c

whole blood, 1 h after a 30 mg/kg dose to male Han Wistar rats. Ligand lipophilicity eﬃciency, deﬁned as human HTRF pIC₅₀− log D_7.4.

d

e

Aqueous solubility measured at pH 7.4, performed under thermodynamic conditions from solid samples. CL = clearance; dosed individually

f

g

(HPMC/Tween suspension for po and solution for iv) to fed male Han Wistar rats. n = 2. n = 1.

aminoadamantan-1-ol to give the target compounds. Where R⁴

= alkoxy, thioacetamide was condensed with diethyl 2-

bromopropanedioate 9 to give 10, which was then converted

to alkoxy thiazole 11 via alkylation or Mitsunobu chemistry,

and then converted to the ﬁnal compounds as before. An

exception to this route was 15, where the R⁴= OMe group was

installed by displacement of a chloro with methoxide (see

Supporting Information). The corresponding thioethers at R⁴

were generated from alkylation of a thiol as detailed in

Supporting Information for 23. To synthesize examples with R²

= alkoxy, we ﬁrst condensed thiourea with bromo ketoester 12

to generate 2-aminothiazole 13, followed by conversion to the

corresponding 2-chlorothiazole 14. Compound 14 could be

converted through to the ﬁnal compounds by displacing the

chloro with an alkoxide, followed by saponiﬁcation of the ester

and amide coupling. Alternatively the ester of 14 could be

saponiﬁed ﬁrst, followed by amide coupling and displacement

of the chloro as the last step. 2-Aminothiazole 12 could also be

functionalized directly when alkylamine R²groups were desired

in the ﬁnal compound, as detailed in Supporting Information

for 33 and 34.

and 27 as start points for optimization, we checked whether the

favorable [brain]/[blood] ratio had been maintained, and we

also assessed potency of 5 and 27 in a low-throughput HPLC

11β-HSD1 assay. The HPLC assay was better at accurately

assessing the potency of very potent compounds that approach

the tight binding limit of the HTRF assay, as the greater

sensitivity of the HPLC assay meant that we could lower the

enzyme concentration 10-fold, so very potent compounds

appeared more potent in the HPLC assay. The THF

compound 27 was more potent in the HPLC assay than 5,

and both 25 and 27 maintained high [brain]/[blood] ratios,

consistent with the low PSA (all compounds in Table 1 have

calculated PSA < 80 Å²). Note the unusually high ratio of 2.9

for 27 is due primarily to the measured diﬀerences in brain

tissue versus plasma protein binding for that compound; the

total [brain]/[blood] ratio was 1.5. We felt that alkyl R⁴groups

such as THF and cyclopropyl oﬀered broader scope for further

optimization than alkoxy groups, as they allow us to vary R²

more widely, as for example, if R²and R⁴are both alkoxy, the

resulting thiazoles were found to be unstable.

We then assessed the SAR of the R²position (Table 2),

focusing primarily on compounds with R⁴= THF, although we

also synthesized some examples with R⁴= c-Pr. Removing the

methyl to give 29 was tolerated, and in fact the resulting

compound has excellent LLE (6.4) as a consequence. However,

as 29 gave modest murine potency and poor rat PK, we

subsequently focused on compounds with R²substituents.

Ethyl 28 was of similar potency to methyl 27 against both

isoforms, with worse LLE (5.0). However, 28 did allow us to

generate our ﬁrst high-resolution crystal structure from this

series (a structure had been obtained previously with 5 but at

low resolution). The crystal structure for 28 (4c7k, resolution

1.9 Å, Figure 4) shows the expected pair of hydrogen bonds

from the ligand carbonyl to Ser170 and Tyr183, and the

adamantyl hydroxy group takes part in an extended hydrogen-

bonding network with several crystallographic waters and the

NADP(H) cofactor. When comparing this structure to the

previously obtained structure of pyrimidine 1 (4bb5), we

observed the loss of a weak hydrogen bond to Gly216 due to

the absence of the corresponding hydrogen-bond acceptor in

the thiazole ring. However, this is compensated by an additional

interaction between Leu215 and the sulfur atom of the thiazole,

the result of electron density transfer from the electron-rich

carbonyl to the electron-deﬁcient sulfur. The strength of this

interaction is expected to be similar in magnitude to a typical

hydrogen bond.²⁶The dihedral angle between the aromatic ring

These synthetic routes allowed us to vary the R²and R⁴

groups on the thiazole and assess the SAR. Simple thiazole

amides 15 and 16 (Table 1) suggested that removal of the

methyl at R²was tolerated but that a group at R⁴was required

for potency. Compound 17 demonstrated that a methyl group

at R⁴was much less potent than methoxy. Larger lipophilic

alkoxide groups at R⁴(18−21) increased potency, although

LLE (ligand lipophilicity eﬃciency, pIC₅₀− log D) was not

improved. A more polar 3-oxetanyl group 22 was well tolerated

for human potency, considering the low log D and favorable

physical properties, but gave poor mouse potency. Sulfur-linked

alkyl chains were also explored to give 23, given the precedent

from a previous program.¹⁹This gave excellent potency against

the human isoform but poor potency against the murine

isoform and very poor stability in rat hepatocytes, so this

approach was not pursued further. The murine potency result

for 23 was in line with our expectation from the published

crystal structures that the murine isoform does not tolerate very

large groups at R⁴. We also explored alkyl R⁴groups, where

larger groups such as cyclopropyl 25 and (S)-2-THF 27 gave

comparable HTRF potency, metabolic stability, and solubility

to the methoxy start point 5. The S-THF enantiomer 26 was 10

times less potent than the R enantiomer, as expected from our

experience with pyrimidines, where the R stereochemistry in 3

was also the more potent enantiomer. To further proﬁle 5, 25,

975

dx.doi.org/10.1021/jm4016729 | J. Med. Chem. 2014, 57, 970−986

Journal of Medicinal Chemistry

Article

Figure 4. Selected ligands cocrystallized with human isoform of 11β-HSD1: (a) ligand binding pocket of 28 (4c7k); (b) interaction map showing

directional interactions; (c) ligand binding pocket of 1 (4bb5); (d) overlay of 28 (green carbons) and 1 (cyan carbons).

and the amide is approximately the same between pyrimidine 1

and thiazole 28 (56° for pyrimidine, 41° for thiazole; note the

presence of an intramolecular hydrogen bond for the thiazole

that is not present in the pyrimidine). This positions the R⁴

group of both ligands into the same lipophilic pocket, with the

potential for larger groups to be tolerated. The orientation of

the R²group of the thiazole diﬀers somewhat from the

equivalent group in the pyrimidine ligand and so we would not

expect substituent SAR to transfer directly from one series to

the other.

was less active. A side chain bearing a sulfone group 35,

designed with the intention of making a polar interaction, was

signiﬁcantly less active, although LLE is maintained due to a

low log D (0.4). As well as giving a lower log D than we

typically explored, 35 gives a high calculated PSA (111 Å²), and

interestingly we observed poor CNS exposure for 35 (total

ratio 0.03), despite having high bioavailability in rat (65%).

Switching to O-linked R²groups gave us a breakthrough:

methoxy 36 had excellent potency against both isoforms, close

to the tight binding limit of the human assay, so potency was

assessed in the HPLC assay as being 0.4 nM. Further alkoxy

examples 37−40 all proved to be very potent against both

isoforms, and 37, 38, and 40 were bioavailable in rat and had

excellent CNS penetration (free ratio ∼1:1 at 1 h time point).

The alkoxy groups in 37−40 were selected for synthesis on the

basis of the more extensive set made with R⁴= c-Pr (vide infra).

These compounds were further proﬁled in mouse, and 40 was

taken forward as a key probe compound to explore preclinical

eﬃcacy.

By maintaining R⁴= THF, we were able to explore a wide

diversity of R²groups; in particular, we aimed to place polar

groups in the R²region, as we expected from the crystal

structure that these may be tolerated. Changing methyl to

methoxymethyl (30, Table 2) maintained human potency but

reduced murine potency. However, spacing the methoxy out

with an additional methylene to give 31 restored the murine

potency, which demonstrated that correct placement of polarity

could maintain potency against both isoforms. SAR remained

ﬂat with NH₂and NHMe groups (32 and 33), but NMe₂34

976

dx.doi.org/10.1021/jm4016729 | J. Med. Chem. 2014, 57, 970−986

Journal of Medicinal Chemistry

Article

With the data in hand on the signiﬁcant beneﬁt of R²=

alkoxy, we also explored compounds with other alkyl groups in

the R⁴position, in particular R⁴= c-Pr. As shown in Table 2, the

SAR with R⁴= c-Pr is broadly similar to that obtained with R⁴=

THF. Methoxy 41 was again found to be more potent than

methyl 25, particularly against the murine isoform. This result

led to the synthesis of a range of alkoxy groups 42−51, which

generated several very potent compounds with suitable PK

properties to be considered as probe compounds. The crystal

structure of 28 was used to inform the design of these

compounds, notably 44 (pendant alcohol), which was designed

to displace an observed crystallographic water and thus gain

additional potency and improved LLE. Figure 5 shows the

donor count in this case, appears to reduce the free [brain]/

[blood] ratio.

In general, it can be seen that the compounds with R⁴= THF

are more potent than the corresponding matched pairs with R⁴

= c-Pr, and 40 in particular was considered a suitable probe

compound for further studies. However, as we were keen to

take two structurally diverse compounds into eﬃcacy studies,

we proﬁled both 40 and 51 in mouse PK, where they both

showed good oral exposure (Table 3). Both compounds were

tested for 11β-HSD2 inhibition and were found to be inactive,

which is important as 11β-HSD2 is believed to be responsible

for the reverse reaction in vivo (11β-HSD1 converts cortisone

to cortisol, whereas 11β-HSD2 converts cortisol to cortisone),

and 11β-HSD2 activity has been associated with increased risk

of hypertension.²⁷

Before proceeding to eﬃcacy studies in mice, we examined

whether these compounds could deliver target engagement in

the brain and established a pharmacokinetic/pharmacodynamic

(PK/PD) relationship, which due to technical issues was done

in the rat. Compound 51 was selected for this study as it had

good potency against the rat isoform with HTRF IC₅₀= 55 nM

(68% conﬁdence limit from SEM 42−70 nM), whereas 51 had

modest rat potency of 345 nM (265−449 nM). A dose

response study was designed with rats receiving a single oral

dose of 51. One hour postdose (to coincide with C_maxin

plasma) the animals were terminated, compound concen-

trations in plasma and brain were quantiﬁed, and target

engagement in CNS was assessed by comparing the ratio of

corticosterone/dehydrocorticosterone (DHC) between the

groups. Figure 6 shows the PK/PD relationship between

corticosterone/DHC ratio and drug concentrations, in both

plasma and CNS. The resulting PK/PD ﬁt suggests that the in

vivo IC₅₀(49 nM vs plasma PK, 20 nM vs brain PK)

corresponds closely to the in vitro IC₅₀(55 nM by HTRF)

within error. It should be noted that there is considerable

variability in the PD data as evident from the plots in Figure 6,

giving a 2-fold variability in the estimated IC₅₀. This study

gave us conﬁdence that where the free compound concen-

tration exceeded the in vitro IC₅₀, we should obtain target

engagement in vivo, both peripherally and in the brain. More

speciﬁcally, where free compound concentration >10× IC₅₀,

then total target engagement should be obtained. A reduction

of the activity of 11β-HSD1 in the CNS was also observed in

Figure 5. Overlay of 28 (R⁴= ethyl, ligand carbons and solvent atoms

in green, 4c7k) and 44 (R⁴= OCH₂CH₂OH, ligand carbons and

solvent atoms in magenta, 4c7j). The hydroxyl group of 44 displaces a

crystallographic water present in the crystal structure of 28.

overlay of the crystal structures of 28 and 44 (4c7j, resolution

2.2 Å), with the side chain of 44 displacing the water molecule

as intended. This is further reﬂected in the measured potency,

with potency maintained with respect to ethoxy 42 at reduced

log D (Δlog D = −1.3), and consequently 44 has excellent LLE

(6.4). However, although CNS penetration for the R⁴= c-Pr

examples was generally good, 44 is an exception. It is

interesting to contrast 44 (hydroxy, free ratio 0.1) with the

close analogue 45 (methoxy, free ratio 0.6), where once again it

appears that high PSA, or more speciﬁcally hydrogen-bond

3

this study by means of an ex vivo assay (conversion of H-

3

cortisone to H-cortisol).

Table 3. Pharmacokinetic Data for 40 and 51

a

b

species

PPB % free

hep CL_int, μL·min⁻¹·10⁻⁶

iv/po dose, μmol·kg⁻¹

iv CL, mL·min⁻¹·kg⁻¹

iv V_ss, L·kg⁻¹

po AUC, μM·h

F, %

Compound 40

c

rat

20

22

23

12

11

<3

5.2/7.4

34

26

1.0

1.7

3.9

83

d

mouse

3.5/17.6

20.6

139

human

Compound 51

c

rat

20

22

16

22

25

<3

3.1/12.3

4.0/14.3

16

18

0.6

1.0

7.9

79

73

d

mouse

12.0

human

a

b

In vitro median intrinsic clearance observed with hepatocytes from the appropriate species. All measurements were taken from whole blood (not

plasma), as compounds from this series were observed to partition unequally. iv, intravenous; po, per os; CL, clearance; V_ss, steady-state volume of

c

distribution; AUC, area under curve. Dosed individually (HPMC/Tween suspension for po and solution for iv) to fed male Han Wistar rats.

d

Dosed individually (HPMC/Tween suspension for po and solution for iv) to fed male C57B6 mice.

977

dx.doi.org/10.1021/jm4016729 | J. Med. Chem. 2014, 57, 970−986

Journal of Medicinal Chemistry

Article

Figure 6. PK/PD relationship of 51 in the rat, comparing PD activity in the brain versus PK in plasma (left, IC₅₀= 49 nM) or PK in brain (right,

IC₅₀= 20 nM). Points shown represent individual animals. More detailed parameters for model ﬁtting can be found in Supporting Information

(Table S2).

Compounds 40 and 51 were then evaluated for eﬃcacy in a

21-day compound-in-diet (CID) study in C57B6 mice.

Following similar protocols in the literature,²⁸the mice in

this study were fed a daily diet of 60% high-fat chow. The

compound was ﬁnely ground and mixed with food and the

mice, preconditioned on 60% high-fat diet, were allowed to eat

2.5 g/day. The CID protocol was devised to eliminate stress to

the mice and avoid the possibility of adverse readings driven by

elevated cortisol. In a scoping study for 5 days, all of the

animals ate the required amount of food each day, which

eliminated the possibility of any taste aversion stopping the

animal receiving the correct calorie intake. Both compounds

provided suﬃcient drug concentrations in the mouse CNS to

test the 11β-HSD1 CNS hypothesis. Compound 51 gave a

modest CNS ratio (free [brain]/[blood] ratio = 0.2) but still

provided a free drug concentration in the brain of 20× IC₅₀

Figure 7. Ex vivo target engagement of 40 and 51 in liver in a mouse

(HPLC) at 20 mg/kg. Compound 40 gave better CNS

diet-induced obesity (DIO) eﬃcacy study. Compound treated arms

exposure (free [brain]/[blood] = 0.5−0.7) and gave free drug

(for 40 and 51) are in combination with 60% high-fat diet (HFD).

concentrations of 20× IC₅₀(HPLC) at 4 mg/kg and 200× IC₅₀

(HPLC) at 40 mg/kg. Target engagement in the liver was

conﬁrmed by use of an ex vivo protocol, with both compounds

showing high inhibition of 11β-HSD1 in the liver as shown in

Figure 7. We could not directly assess CNS target engagement

in mouse. However, as we achieved 20−200× IC₅₀free drug

concentration in the CNS and have both evidence of target

engagement in the liver and conﬁdence in our predictions from

the PK/PD relationships in both plasma and brain in the rat, we

were conﬁdent that we had achieved CNS target engagement in

the eﬃcacy study. Target engagement in adipose tissue could

also not be directly shown in mouse. However, we have data

from rats demonstrating that complete target engagement is

observed when the free exposure exceeds the HPLC IC₅₀by

>10-fold. As target engagement has been shown in the liver in

the eﬃcacy study in all study groups, and high peripheral IC₅₀

cover has been achieved in the eﬃcacy study (>10× HPLC

IC₅₀in all study groups; for example, 280× HPLC IC₅₀for 40

at 40 mg/kg), we conclude that target engagement in adipose

tissue was also achieved in this study.

Error bars denote 68% conﬁdence limits as calculated from SEM (n =

6 for lean control, n = 12 for HFD control, n = 10 for each treated

group).

syndrome. It should be noted that these data are consistent

with the ﬁndings from recently published knockout data.²⁹We

noted that 11β-HSD1 programs in the literature had

experienced activation of the pregnane X receptor

(PXR)³⁰⁻³²and inhibition of soluble epoxide hydrolase

(sEH)³³⁻³⁵and also general concerns that the positive

preclinical eﬃcacy of other 11β-HSD1 inhibitors may be due

to oﬀ-target activities;¹³therefore, we assessed our compounds

for PXR activation and sEH inhibition. All 11 compounds

tested (including 51, but 40 was not assessed) were found to

give <20% activation of PXR at 10 μM. When assessed for sEH

inhibition, 40 was found to be inactive (IC₅₀> 33 μM),

although 51 showed modest inhibition with IC₅₀of 12 μM

(68% conﬁdence limit 7−20 μM). Interestingly, pyrimidine 1 is

a potent inhibitor of sEH with an IC₅₀of 22 nM (18−27 nM)

and has previously demonstrated glucose and body weight

eﬃcacy at high doses.¹⁵Compound 1 did not provide glucose

eﬃcacy at lower doses, despite providing target engagement of

11β-HSD1, and this study casts doubt on the hypothesis that

this was due to it having poor CNS penetration. We conclude

that it is possible that the positive eﬃcacy observed at higher

doses with 1 may have been due to sEH inhibition and/or

No decrease in body weight or changes in insulin or glucose

levels (and therefore HOMA-IR) in any of the animal groups

were observed in the eﬃcacy study, compared to the high-fat

diet control (Figure 8). These data cast doubt on the

hypothesis that localized tissue modulation of 11β-HSD1

activity, either peripheral or central, would alleviate metabolic

978

dx.doi.org/10.1021/jm4016729 | J. Med. Chem. 2014, 57, 970−986

Journal of Medicinal Chemistry

Article

atherosclerosis), cognitive (e.g., Alzheimer’s disease), or

inﬂammatory nature remains to be determined.

EXPERIMENTAL SECTION

■

General. All solvents and chemical used were reagent-grade. Purity

and characterization of compounds were established by a combination

of LCMS and NMR. LCMS spectra were obtained by use of a Waters

liquid chromatography mass spectrometry system, where purity was

determined by UV absorption at a wavelength of 254 nm and the mass

ion was determined by electrospray ionization (Micromass instru-

ment). All test compounds were >95% purity as assessed by LCMS

and ¹H NMR. ¹H NMR spectra were recorded on a Varian AV400 FT

spectrometer or via ﬂow NMR process on an Avance 500 FT

spectrometer, in deuterated dimethyl sulfoxide (DMSO-d₆) or CDCl₃,

with the data expressed as chemical shifts in parts per million (ppm)

from internal standard tetramethylsilane (TMS). Preparative HPLC

was performed on a Waters or Phenomenex column with decreasingly

polar mixtures of water (containing 1% formic acid or 1% aqueous

NH₄OH) and MeCN. Puriﬁcation by ﬂash column chromatography

(FCC) was typically performed on silica gel (Merck 7734 grade), and

solvent mixtures and gradients are recorded herein. All reactions were

performed under nitrogen unless otherwise stated. All IC₅₀data are

quoted as geometric means, and the 68% conﬁdence intervals shown

in parentheses are calculated from the SEM.

Synthesis of Representative Key Examples (3, 5, 27, 36, 40,

and 51). (2R)-Tetrahydrofuran-2-carbonyl Chloride. N,N-Dimethyl-

formamide (DMF; 0.015 mL, 0.20 mmol) was added to (2R)-

tetrahydrofuran-2-carboxylic acid (2.3 g, 19.8 mmol) and oxalyl

chloride (3.64 mL, 41.6 mmol) in dichloromethane (DCM; 25 mL) at

room temperature (rt). The reaction mixture was stirred at rt for 20 h

and then evaporated to aﬀord the title compound (2.60 g, 98%), which

1

was used without further puriﬁcation. H NMR (400 MHz, CDCl₃)

1.93−2.04 (2H, m), 2.19−2.28 (1H, m), 2.33−2.42 (1H, m), 3.99−

4.07 (2H, m), 4.71−4.74 (1H, m), CO₂H not observed. LCMS not

observed.

Figure 8. Eﬀect on insulin resistance and fasting glucose of 40 and 51

in a mouse diet-induced obesity (DIO) 21-day study, showing no

signiﬁcant eﬃcacy. HOMA-IR is deﬁned as [fasting glucose

(millimolar) × fasting insulin (milliunits per liter)]/22.5. Error bars

denote 68% conﬁdence limits as calculated from SEM (n = 6 for lean

control, n = 12 for HFD control, n = 10 for treated groups).

2,2-Dimethyl-5-[(2R)-tetrahydrofuran-2-carbonyl]-1,3-dioxane-

4,6-dione. (2R)-Tetrahydrofuran-2-carbonyl chloride (2.6 g, 19.3

mmol) in DCM (5 mL) was added dropwise to 2,2-dimethyl-1,3-

dioxane-4,6-dione (3.06 g, 21.3 mmol) and pyridine (3.13 mL, 38.6

mmol) in DCM (25 mL) at 0 °C. The reaction mixture was stirred at

0 °C for 1 h and at rt for 2 h and then was evaporated, diluted with

DCM, and washed sequentially with 1 M aqueous citric acid, water,

and saturated brine. The organic layer was dried over MgSO₄, ﬁltered,

and evaporated to aﬀord the title compound (4.28 g, 91%) as a brown

another unknown oﬀ-target mechanism, rather than due to

11β-HSD1 inhibition.

1

oil that was used in the next stage without further puriﬁcation. H

CONCLUSION

■

NMR (400 MHz, CDCl₃) 1.74 (6H, s), 1.86−1.94 (1H, m), 1.98−

2.05 (2H, m), 2.57−2.68 (1H, m), 4.01−4.15 (2H, m), 5.65−5.70

(1H, m), 15.42 (1H, s). m/z [M − H]⁻= 241. HPLC t_R= 0.43 min.

Methyl 3-Oxo-3-[(2R)-tetrahydrofuran-2-yl]propanoate. Metha-

nol (25 mL) was added to 2,2-dimethyl-5-[(2R)-tetrahydrofuran-2-

carbonyl]-1,3-dioxane-4,6-dione (4.28 g, 17.7 mmol) in toluene (50

mL) at rt. The reaction mixture was heated at reﬂux for 4 h and then

allowed to cool to rt, evaporated, and puriﬁed by FCC, elution

gradient 5−20% EtOAc in isohexane, to aﬀord the title compound

Starting from 3, with low CNS penetration, we synthesized

compounds with diverse heterocyclic cores and identiﬁed

thiazole 5, which had good potency, CNS penetration, and high

LLE. We successfully optimized R²and R⁴positions on the

thiazole and found that an oxygen atom at either position (but

especially at R²) increased potency. An oxygen linking atom at

R²and a lipophilic group at R⁴were key for maximum

inhibition. Compounds 40 and 51 were selected for eﬃcacy

studies on the basis of good murine potency, oral exposure, and

CNS penetration. Both compounds were dosed to DIO mice

by a CID protocol over 21 days. Excellent target engagement

was achieved in liver, as assessed by an ex vivo assay, and by

comparing the PK/PD relationship in rat to the high PK

exposures in mouse in the eﬃcacy study, we conclude that

target engagement was also achieved in brain and adipose

tissue. However, no body weight reduction or eﬀects on glucose

or insulin (and so HOMA-IR) were observed. This study does

not support the hypothesis that 11β-HSD1 inhibition, either

peripheral or central, markedly impacts body weight or glucose

homeostasis. Whether 11β-HSD1 inhibitors have potential

utility in other indications of either metabolic (e.g.,

1

(1.80 g, 59%) as a colorless oil. H NMR (400 MHz, CDCl₃) 1.85−

1.95 (2H, m), 2.00−2.09 (1H, m), 2.15−2.24 (1H, m), 3.58 (2H, s),

3.74 (3H, s), 3.88−3.93 (2H, m), 4.36−4.39 (1H, m). m/z MH⁺=

173. HPLC t_R= 0.80 min.

Methyl (Z)-3-(Dimethylamino)-2-[(2R)-tetrahydrofuran-2-

carbonyl]prop-2-enoate. DMF dimethyl acetal (1.67 mL, 12.6

mmol) was added in one portion to methyl 3-oxo-3-[(2R)-

tetrahydrofuran-2-yl]propanoate (1.8 g, 10.5 mmol) in 1,4-dioxane

(25 mL) at rt. The resulting solution was stirred at 100 °C for 2 h. The

reaction mixture was evaporated, and the resulting crude product was

puriﬁed by FCC, elution gradient 50−100% EtOAc in isohexane, to

aﬀord the title compound (1.80 g, 76%) as a yellow oil. ¹H NMR (400

MHz, CDCl₃) 1.83−1.92 (2H, m), 2.00−2.08 (1H, m), 2.12−2.21

(1H, m), 3.04 (6H, s), 3.73 (3H, s), 3.83−3.96 (2H, m), 4.97 (1H, t),

7.67 (1H, s). m/z [M − H]⁻= 226. HPLC t_R= 1.25 min.

979

dx.doi.org/10.1021/jm4016729 | J. Med. Chem. 2014, 57, 970−986

Journal of Medicinal Chemistry

Article

Methyl 2-Methylsulfanyl-4-[(2R)-tetrahydrofuran-2-yl]-

pyrimidine-5-carboxylate. 2-Methyl-2-thiopseudourea sulfate (1.54

g, 11.1 mmol) was added to methyl (Z)-3-(dimethylamino)-2-[(2R)-

tetrahydrofuran-2-carbonyl]prop-2-enoate (1.8 g, 7.92 mmol) and

sodium acetate (2.73 g, 33.3 mmol) in DMF (30 mL) at rt. The

resulting solution was stirred at 80 °C for 3 h. Water was added to the

cooled solution, the mixture was extracted with EtOAc, and the

organic layer was washed with water, dried over MgSO₄, ﬁltered, and

evaporated. The resulting crude product was puriﬁed by FCC, elution

gradient 5−30% EtOAc in isohexane, to aﬀord the title compound

m), 4.63 (2H, d), 5.08 (1H, t), 7.90 (1H, d), 8.75 (1H, s). m/z MH⁺=

457. HPLC t_R= 1.96 min.

Ethyl 4-Hydroxy-2-methylthiazole-5-carboxylate (10). Thioaceta-

mide (10.1 g, 134 mmol) was added to diethyl bromomalonate 9 (22.9

mL, 134 mmol) in toluene (77 mL). The reaction mixture was stirred

at 110 °C for 90 min, allowed to cool to rt, ﬁltered through Celite, and

washed with toluene. The ﬁltrate was evaporated and the residue was

slurried in water for 5 min. The precipitate was collected by ﬁltration

and then washed with water and isohexane and dried in vacuo to aﬀord

1

the title compound 10 (6.18 g, 25%) as a pale yellow solid. H NMR

1

(1.46 g, 73%) as a colorless oil. H NMR (400 MHz, CDCl₃) 1.96−

(400 MHz, DMSO) 1.22 (3H, t), 2.56 (3H, s), 4.17 (2H, q), 11.72

(1H, s). m/z MH⁺= HATU188. HPLC t_R= 1.38 min.

2.10 (3H, m), 2.38−2.49 (1H, m), 2.60 (3H, s), 3.91 (3H, s), 4.00−

4.05 (1H, m), 4.13−4.19 (1H, m), 5.69−5.74 (1H, m), 8.88 (1H, s).

m/z MH⁺= 255. HPLC t_R= 1.88 min.

Ethyl 4-Methoxy-2-methyl-thiazole-5-carboxylate. Diisopropyl

azodicarboxylate (DIAD; 4.67 mL, 24.0 mmol) in THF (16.7 mL)

was added to ethyl 4-hydroxy-2-methylthiazole-5-carboxylate 10 (3 g,

16.0 mmol), PPh₃(4.62 g, 17.6 mmol), and MeOH (1.95 mL, 48.1

mmol) in THF (85 mL) at 10 °C. The reaction mixture was stirred at

rt overnight and then heated at 50 °C for 2 h. The reaction mixture

was concentrated in vacuo, and then 3:1 isohexane/EtOAc (50 mL)

was added and the mixture was ﬁltered. The ﬁltrate was concentrated

in vacuo and then partially puriﬁed by FCC, elution gradient 10−60%

EtOAc in isohexane. The resulting impure product was dissolved in

EtOAc and washed with 2 M aqueous NaOH, and the precipitate was

discarded. The resulting organic layer was dried over MgSO₄, ﬁltered,

and evaporated to aﬀord the title compound (1.56 g, 49%) as a

colorless oil. ¹H NMR (400 MHz, DMSO) 1.22 (3H, t), 2.62 (3H, s),

4.01 (3H, d), 4.17 (2H, q). m/z MH⁺= 202. HPLC t_R= 1.74 min.

4-Methoxy-2-methylthiazole-5-carboxylic Acid. Aqueous NaOH

(2 M, 8.05 mL, 16.1 mmol) was added to ethyl 4-methoxy-2-

methylthiazole-5-carboxylate (1.08 g, 5.37 mmol) in ethanol (20 mL),

and the reaction mixture was stirred at rt for 1 h and then evaporated.

The resulting residue was diluted with water and acidiﬁed with 2 M

aqueous HCl (8.05 mL), and the resulting precipitate was isolated by

ﬁltration and dried in vacuo to aﬀord the title compound (0.718 mg,

77%) as a white solid. ¹H NMR (400 MHz, DMSO) 2.67 (3H, s), 4.04

(3H, s), 12.64 (1H, s). m/z MH⁺= 174. HPLC t_R= 1.01 min.

N-(trans-5-Hydroxy-2-adamantyl)-4-methoxy-2-methylthiazole-

5-carboxamide (5). HATU (1.14 g, 2.99 mmol) was added to a stirred

solution of 4-methoxy-2-methylthiazole-5-carboxylic acid (431 mg,

2.49 mmol), (1S,4R)-4-aminoadamantan-1-ol hydrochloride (507 mg,

2.49 mmol), and i-Pr₂NEt (1.30 mL, 7.47 mmol) in DMF (5 mL), and

the reaction mixture was stirred at rt for 17 h. The reaction mixture

was poured onto water and extracted with EtOAc, and the combined

organic layers were washed with saturated brine, dried over MgSO₄,

ﬁltered, and evaporated. The resulting crude product was puriﬁed by

preparative HPLC to aﬀord the title compound 5 (0.49 g, 61%) as a

2-Methylsulfanyl-4-[(2R)-tetrahydrofuran-2-yl]pyrimidine-5-car-

boxylic acid. LiOH·H₂O (0.482 g, 11.48 mmol) in water (10 mL) was

added to a stirred solution of methyl 2-methylsulfanyl-4-[(2R)-

tetrahydrofuran-2-yl]pyrimidine-5-carboxylate (1.46 g, 5.74 mmol) in

THF (20 mL) at rt. The resulting mixture was stirred at rt for 3 days

and then partially evaporated. The resulting aqueous layer was washed

with EtOAc and then was acidiﬁed with 1 M aqueous citric acid and

extracted with EtOAc. The organic layer was washed with saturated

brine, dried over MgSO₄, ﬁltered, and evaporated to aﬀord the title

compound (1.23 g, 89%) as a white solid. ¹H NMR (400 MHz,

CDCl₃) 1.99−2.20 (3H, m), 2.39−2.49 (1H, m), 2.62 (3H, s), 4.03−

4.11 (1H, m), 4.16−4.22 (1H, m), 5.66−5.70 (1H, m), 9.02 (1H, s),

CO₂H not observed. m/z MH⁺= 241. HPLC t_R= 0.59 min.

N-(5-Hydroxy-2-adamantyl)-2-methylsulfanyl-4-[(2R)-oxolan-2-

yl]pyrimidine-5-carboxamide. i-Pr₂NEt (3.57 mL, 20.48 mmol) was

added to 2-methylsulfanyl-4-[(2R)-tetrahydrofuran-2-yl]pyrimidine-5-

carboxylic acid (1.23 g, 5.12 mmol), (1S,4R)-4-aminoadamantan-1-ol

hydrochloride (1.043 g, 5.12 mmol), and 2-(3H-[1,2,3]triazolo[4,5-

b]pyridin-3-yl)-1,1,3,3-tetramethylisouronium hexaﬂuorophosphate-

(V) (HATU; 2.34 g, 6.14 mmol) in DMF (15 mL) at rt. The

resulting solution was stirred at rt for 16 h and then was evaporated

and taken up in EtOAc. The organic layer was washed with water and

saturated brine, dried over MgSO₄, ﬁltered, and evaporated, and the

resulting crude product was puriﬁed by FCC, elution gradient 1−6%

DCM in MeOH, to aﬀord the title compound (1.18 g, 59%) as an oﬀ-

1

white solid. H NMR (400 MHz, CDCl₃) 1.50−1.59 (3H, m), 1.75−

1.83 (6H, m), 1.90−1.97 (2H, m), 2.03−2.27 (6H, m), 2.59 (3H, s),

2.80−2.91 (1H, m), 3.89−3.93 (1H, m), 3.97−4.02 (1H, m), 4.20−

4.26 (1H, m), 5.14 (1H, t), 7.91 (1H, d), 8.86 (1H, s). m/z MH⁺=

390. HPLC t_R= 1.69 min.

N-(5-Hydroxy-2-adamantyl)-2-methylsulfonyl-4-[(2R)-oxolan-2-

yl]pyrimidine-5-carboxamide. 3-Chloroperoxybenzoic acid (70%)

(1.392 g, 5.65 mmol) was added in one portion to N-(5-hydroxy-2-

adamantyl)-2-methylsulfanyl-4-[(2R)-oxolan-2-yl]pyrimidine-5-car-

boxamide (1.1 g, 2.82 mmol) in DCM (45 mL) at 0 °C. The resulting

solution was stirred at rt for 24 h. The reaction mixture was taken up

in DCM and washed with saturated aqueous NaHCO₃and saturated

brine. The organic layer was dried over MgSO₄, ﬁltered, and

evaporated to aﬀord the title compound (1.18 g, 99%) as a white

1

white solid. H NMR (400 MHz, CDCl₃) 1.51−1.58 (1H, m), 1.69−

1.83 (7H, m), 1.85−1.97 (2H, m), 2.12−2.22 (3H, m), 2.62 (3H, s),

4.16 (4H, d), 7.32 (1H, d), OH signal not observed. m/z MH⁺= 323.

HPLC t_R= 1.59 min.

Ethyl 2-Chloro-3-oxo-3-[(2R)-tetrahydrofuran-2-yl]propanoate.

Sulfuryl chloride (12.0 mL, 149 mmol) was added dropwise to ethyl

3-oxo-3-[(2R)-tetrahydrofuran-2-yl]propanoate (27.7 g, 150 mmol) in

DCM (500 mL) at 0 °C. The resulting solution was allowed to warm

to rt and was stirred for 2 h and then quenched by pouring cautiously

onto stirred water (250 mL). The organic layer was isolated, and the

aqueous layer was extracted with DCM. The combined organic layers

were washed with saturated brine, dried over MgSO₄, and evaporated

to aﬀord crude title compound (29.8 g, 91%) that was used without

1

solid. H NMR (400 MHz, CDCl₃) 1.50−1.60 (3H, m), 1.74−1.85

(6H, m), 1.90−1.98 (2H, m), 2.08−2.31 (6H, m), 2.79−2.90 (1H, m),

3.36 (3H, s), 3.90−4.04 (2H, m), 4.23−4.30 (1H, m), 5.24 (1H, t),

7.88 (1H, d), 9.17 (1H, s). m/z MH⁺= 422. HPLC t_R= 1.31 min.

2-[(2S,6R)-2,6-Dimethylmorpholin-4-yl]-N-(5-hydroxy-2-ada-

mantyl)-4-[(2R)-oxolan-2-yl]pyrimidine-5-carboxamide (3). N-(5-

Hydroxy-2-adamantyl)-2-methylsulfonyl-4-[(2R)-oxolan-2-yl]-

pyrimidine-5-carboxamide (12.3 g, 29.2 mmol) and (2R,6S)-2,6-

dimethylmorpholine (15 mL, 121.1 mmol) were dissolved in THF

(150 mL), and the resulting solution was stirred at rt for 24 h. The

reaction mixture was evaporated and the crude product was puriﬁed by

FCC, elution gradient 1−5% MeOH in EtOAc, and then triturated

with diethyl ether to aﬀord the title compound 3 (7.80 g, 59%) as a

1

further puriﬁcation. H NMR (400 MHz, CDCl₃) 1.19−1.3 (3H, m),

1.78−1.93 (2H, m), 1.99−2.26 (2H, m), 3.76−3.9 (2H, m), 4.15−4.34

(2H, m), 4.45−4.59 (1H, m), 5.10 (1H, d). Mass ion not observed by

LCMS.

Ethyl 2-Methyl-4-[(2R)-tetrahydrofuran-2-yl]thiazole-5-carboxy-

late. Ethanethioamide (232 mg, 3.08 mmol) was added to a stirred

solution of crude ethyl 2-chloro-3-oxo-3-[(2R)-tetrahydrofuran-2-

yl]propanoate (682 mg, 3.08 mmol if pure) in EtOH (6 mL), and

the resulting solution was stirred at 80 °C for 90 min. The reaction

mixture was allowed to cool to rt, evaporated, and puriﬁed by FCC,

elution gradient 0−100% EtOAc in isohexane, to aﬀord the title

1

white solid. H NMR (400 MHz, CDCl₃) 1.26 (6H, d), 1.41 (1H, s),

1.48−1.58 (2H, m), 1.75−1.85 (6H, m), 1.89−1.96 (2H, m), 2.01−

2.09 (2H, m), 2.14−2.23 (4H, m), 2.62 (2H, t), 2.74−2.82 (1H, m),

3.57−3.66 (2H, m), 3.91 (1H, q), 3.98−4.03 (1H, m), 4.19−4.26 (1H,

980

dx.doi.org/10.1021/jm4016729 | J. Med. Chem. 2014, 57, 970−986

Journal of Medicinal Chemistry

Article

compound (250 mg, 26%) as a yellow oil. ¹H NMR (400 MHz,

DMSO) 1.26 (3H, t), 1.85−2.10 (3H, m), 2.13−2.26 (1H, m), 2.65

(3H, s), 3.78 (1H, td), 3.89−4.01 (1H, m), 4.26 (2H, q), 5.47−5.83

(1H, m). m/z MH⁺= 242. HPLC t_R= 1.84 min.

with 1 M aqueous citric acid, and extracted with EtOAc. The organic

layer was washed with water and saturated brine, dried over MgSO₄,

and evaporated. The resulting crude product was puriﬁed by FCC,

elution gradient 20−100% EtOAc in heptanes, to aﬀord the title

1

2-Methyl-4-[(2R)-tetrahydrofuran-2-yl]thiazole-5-carboxylic Acid.

Aqueous NaOH (2 M, 1.55 mL, 3.11 mmol) was added to ethyl 2-

methyl-4-[(2R)-tetrahydrofuran-2-yl]thiazole-5-carboxylate (250 mg,

1.04 mmol) in MeOH (4 mL). The reaction mixture was stirred at rt

for 3 h and then acidiﬁed with 1 M aqueous citric acid and diluted with

EtOAc, and the organic layer was isolated. The aqueous layer was

extracted with EtOAc, and the combined organic layers were washed

with water and saturated brine, dried over MgSO₄, ﬁltered, and

evaporated to aﬀord the title compound (215 mg, 97%) as a cream-

colored solid. ¹H NMR (400 MHz, DMSO) 1.84−2.22 (4H, m), 2.64

(3H, s), 3.72−3.82 (1H, m), 3.89−3.98 (1H, m), 5.63 (1H, t), 13.39

(1H, br s). m/z MH⁺= 214. HPLC t_R= 0.66 min.

N-(trans-5-Hydroxy-2-adamantyl)-2-methyl-4-[(2R)-tetrahydro-

furan-2-yl]thiazole-5-carboxamide (27). HATU (460 mg, 1.21

mmol) was added to a stirred solution of 2-methyl-4-[(2R)-

tetrahydrofuran-2-yl]thiazole-5-carboxylic acid (215 mg, 1.01 mmol),

(1S,4R)-4-aminoadamantan-1-ol hydrochloride (226 mg, 1.11 mmol),

and i-Pr₂NEt (0.53 mL, 3.02 mmol) in DMF (5 mL), and the reaction

mixture was stirred at rt for 17 h. The reaction mixture was poured

onto water and extracted with EtOAc, and the combined organic layers

were washed with saturated brine, dried over MgSO₄, ﬁltered, and

evaporated. The resulting crude product was puriﬁed by preparative

HPLC to aﬀord the title compound 27 (150 mg, 41%) as a white solid.

¹H NMR (400 MHz, DMSO) 1.58 (2H, d), 1.72−2.04 (8H, m), 2.16

compound (2.20 g, 85%) as a pale yellow foam. H NMR (400 MHz,

DMSO) 1.84−1.98 (2H, m), 2.03−2.20 (2H, m), 3.76−3.82 (1H, m),

3.79 (3H, s), 3.91−3.97 (1H, m), 4.06 (3H, s), 5.54−5.59 (1H, m).

m/z MH⁺= 244. LCMS t_R= 1.88 min.

2-Methoxy-4-[(2R)-tetrahydrofuran-2-yl]thiazole-5-carboxylic

acid. KOSi(CH₃)₃(2.22 g, 17.3 mmol) was added to methyl 2-

methoxy-4-[(2R)-tetrahydrofuran-2-yl]thiazole-5-carboxylate (2.10 g,

8.63 mmol) in THF (100 mL), and the reaction mixture was stirred at

rt overnight. The reaction mixture was quenched with 1 M aqueous

citric acid and extracted with EtOAc (100 mL). The organic layer was

washed with water and saturated brine, dried over MgSO₄, ﬁltered, and

evaporated to aﬀord the title compound (1.90 g, 96%) as a yellow

1

solid. H NMR (400 MHz, DMSO) 1.84−1.98 (2H, m), 2.02−2.18

(2H, m), 3.71−3.78 (1H, m), 3.89−3.96 (1H, m), 4.03 (3H, s), 5.56−

5.61 (1H, m), 13.25 (1H, br s). m/z MH⁺= 230. HPLC t_R= 1.39 min.

N-(trans-5-Hydroxy-2-adamantyl)-2-methoxy-4-[(2R)-tetrahy-

drofuran-2-yl]thiazole-5-carboxamide (36). HATU (3.78 g, 9.95

mmol) was added to a solution of 2-methoxy-4-[(2R)-tetrahydrofuran-

2-yl]thiazole-5-carboxylic acid (1.90 g, 8.29 mmol), (1S,4R)-4-

aminoadamantan-1-ol hydrochloride (2.53 g, 12.4 mmol), and i-

Pr₂NEt (4.33 mL, 24.9 mmol) in DMF (25 mL). The reaction mixture

was stirred at rt for 17 h and then was evaporated, diluted with water

(50 mL), stirred for 10 min, and ﬁltered. The resulting residue was

washed with water, dried in vacuo, and then dissolved in EtOAc, dried

over MgSO₄, ﬁltered, and evaporated. The resulting crude product was

puriﬁed by crystallization from EtOAc to aﬀord the title compound

(1.70 g, 54%) as a cream-colored solid. ¹H NMR (400 MHz, DMSO)

1.38−1.47 (2H, m), 1.60−1.67 (4H, m), 1.69−1.80 (4H, m), 1.90−

2.06 (5H, m), 2.16−2.27 (1H, m), 2.29−2.39 (1H, m), 3.80−3.85

(1H, m), 3.86−3.91 (1H, m), 3.90−3.98 (1H, m), 4.02 (3H, s), 4.41

(1H, s), 5.04 (1H, t), 8.50−8.57 (1H, m). m/z MH⁺= 379. HPLC t_R

= 1.96 min. HRMS (ES+) for C₁₉H₂₆N₂O₄S (MH⁺): calcd 379.1686,

found 379.1684.

(5H, dd), 2.31−2.46 (1H, m), 2.51−2.71 (1H, m), 2.77 (3H, s), 3.97

(1H, td), 4.08 (2H, dd), 4.58 (1H, s), 5.26 (1H, t), 8.68 (1H, d). m/z

MH⁺= 363. HPLC t_R= 1.70 min. HRMS (ES+) for C₁₉H₂₆N₂O₃S

(MH⁺): calcd 363.1737, found 363.1735.

Ethyl 2-Amino-4-[(2R)-tetrahydrofuran-2-yl]thiazole-5-carboxy-

late. Thiourea (10.3 g, 135 mmol) was added to a solution of ethyl

2-chloro-3-oxo-3-[(2R)-tetrahydrofuran-2-yl]propanoate (29.8 g, 135

mmol) in ethanol (300 mL), and the resulting solution was stirred at

80 °C for 2 h. The reaction mixture was treated with triethylamine

(18.9 mL, 135 mmol), evaporated, and puriﬁed by FCC, elution

gradient 0.3−15% MeOH in DCM, to aﬀord the title compound (26.5

2-Chloro-4-[(2R)-tetrahydrofuran-2-yl]thiazole-5-carboxylic acid.

KOSi(CH₃)₃(7.35 g, 57.3 mmol) was added to a solution of ethyl 2-

chloro-4-[(2R)-tetrahydrofuran-2-yl]thiazole-5-carboxylate (5.00 g,

19.1 mmol) in THF (50 mL), and the reaction mixture was stirred

at rt for 30 min. The reaction mixture was acidiﬁed with 4 M HCl in

1,4-dioxane (20 mL) and evaporated. The resulting residue was

partitioned between DCM and 1 M aqueous citric acid. The organic

layer was isolated and passed through a phase-separation cartridge and

evaporated to aﬀord the title compound (4.43 g, 99%) as a white solid.

¹H NMR (400 MHz, DMSO) 1.92−2.06 (2H, m), 2.14 (1H, ddd),

1

g, 81%) as a pale yellow oil that crystallized upon standing. H NMR

(400 MHz, CDCl₃) 1.17−1.31 (3H, m), 1.75−2.05 (3H, m), 2.21−

2.35 (1H, m), 3.78−3.87 (1H, m), 3.99−4.07 (1H, m), 4.1−4.26 (2H,

m), 5.49 (1H, t), 5.87 (2H, s). m/z MH⁺= 243. HPLC t_R= 1.47 min.

Ethyl 2-Chloro-4-[(2R)-tetrahydrofuran-2-yl]thiazole-5-carboxy-

late. To a stirred suspension of ethyl 2-amino-4-[(2R)-tetrahydrofur-

an-2-yl]thiazole-5-carboxylate (26.5 g, 109 mmol) in 1,4-dioxane (100

mL) was added 4 M HCl in 1,4-dioxane (32.8 mL, 131 mmol)

dropwise. The reaction mixture was stirred at rt for 30 min and

evaporated. THF (300 mL) was added to the residue and the mixture

was stirred at 65 °C before the dropwise addition of tert-butyl nitrite

(17.1 mL, 142 mmol) as a solution in THF (50 mL) over 1 h. The

reaction mixture was stirred for a further 1 h at 65 °C, cooled to rt,

evaporated, and puriﬁed by FCC, eluting with 5−60% EtOAc in

heptanes, to aﬀord the title compound (22.4 g, 78%) as a pale yellow

2.27 (1H, ddt), 3.82−3.9 (1H, m), 4.00 (1H, dd), 5.70 (1H, dd), 14.05

(1H, s). m/z MH⁺= 232, 234. HPLC t_R= 1.39 min.

2-Chloro-N-(5-hydroxy-2-adamantyl)-4-[(2R)-tetrahydrofuran-2-

yl]thiazole-5-carboxamide. 1,1′-Carbonyldiimidazole (CDI; 4.02 g,

24.8 mmol) was added to 2-chloro-4-[(2R)-tetrahydrofuran-2-yl]-

thiazole-5-carboxylic acid (3.86 g, 16.5 mmol) in DMF (50 mL) at rt,

and the resulting solution was stirred at rt for 90 min. (1S,4R)-4-

Aminoadamantan-1-ol hydrochloride (3.37 g, 16.5 mmol) and 1,8-

diazabicyclo[5.4.0]undec-7-ene (DBU; 4.94 mL, 33 mmol) were

added, and the resulting solution was stirred at rt for 2 days. The

reaction mixture was taken up in water and EtOAc and extracted with

EtOAc. The combined organic layers were washed with brine, dried

over MgSO₄, evaporated, and puriﬁed by FCC, elution gradient 0−

60% EtOAc in heptanes, to aﬀord the title compound (4.16 g, 66%) as

1

oil. H NMR (400 MHz, DMSO) 1.28 (3H, t), 1.87−2 (2H, m),

2.01−2.12 (1H, m), 2.18−2.28 (1H, m), 3.77−3.84 (1H, m), 3.91−

3.98 (1H, m), 4.26−4.33 (2H, m), 5.58−5.63 (1H, m). m/z MH⁺=

262. HPLC t_R= 2.26 min. A small amount of dechlorinated material

ethyl 4-[(2R)-tetrahydrofuran-2-yl]thiazole-5-carboxylate (1.24 g, 5%)

1

was obtained as a byproduct, and this was used to synthesize 29. H

NMR (400 MHz, DMSO) 1.29 (3H, t), 1.87−2.03 (2H, m), 2.05−

2.15 (1H, m), 2.15−2.28 (1H, m), 3.76−3.84 (1H, m), 3.91−4.01

(1H, m), 4.24−4.35 (2H, m), 5.69 (1H, dd), 9.22 (1H, s). m/z MH⁺=

228. HPLC t_R= 1.62 min.

Methyl 2-Methoxy-4-[(2R)-tetrahydrofuran-2-yl]thiazole-5-car-

boxylate. A solution of 25% (w/w) NaOCH₃in MeOH (2.23 g,

10.3 mmol) was added to a suspension of ethyl 2-chloro-4-[(2R)-

tetrahydrofuran-2-yl]thiazole-5-carboxylate (2.70 g, 10.3 mmol) in

MeOH (30 mL). The reaction mixture was heated in a microwave

reactor at 100 °C for 6 h and then was allowed to cool to rt, diluted

1

a yellow gum. H NMR (400 MHz, DMSO) 1.43 (2H, d), 1.65 (4H,

d), 1.69−1.83 (4H, m), 1.93−2.01 (2H, m), 2.05 (3H, d), 2.27 (2H,

dt), 3.84 (1H, ddd), 3.94 (2H, dd), 4.43 (1H, s), 5.11−5.18 (1H, m),

8.72 (1H, d). m/z MH⁺= 383. HPLC t_R= 2.05 min.

N-(trans-5-Hydroxy-2-adamantyl)-4-[(2R)-tetrahydrofuran-2-yl]-

2-tetrahydropyran-4-yloxythiazole-5-carboxamide (40). Sodium

hydride (60% w/w in oil, 5.08 g, 127 mmol) was added to

tetrahydro-2H-pyran-4-ol (81.0 mL, 846 mmol) in THF (40 mL) at

5 °C. The reaction mixture was stirred at 5 °C for 30 min, and then a

981

dx.doi.org/10.1021/jm4016729 | J. Med. Chem. 2014, 57, 970−986

Journal of Medicinal Chemistry

Article

solution of 2-chloro-N-(5-hydroxy-2-adamantyl)-4-[(2R)-tetrahydro-

furan-2-yl]thiazole-5-carboxamide (12.0 g, 28.2 mmol) in THF (40

mL) was added. The reaction was stirred at rt for 16 h and then was

poured into 1 M aqueous citric acid and extracted with EtOAc. The

combined organic layers were washed with brine, evaporated, puriﬁed

by FCC, elution gradient 60−100% EtOAc in heptane, and then

triturated with diethyl ether to aﬀord the title compound (6.02 g, 48%)

2-[(1-Cyanocyclopropyl)methoxy]-4-cyclopropyl-N-(trans-5-hy-

droxy-2-adamantyl)thiazole-5-carboxamide (51). Sodium hydride

(60% w/w in oil, 0.51 g, 12.8 mmol) was added to 1-(hydroxymethyl)-

cyclopropanecarbonitrile (1.24 g, 12.8 mmol) in THF (15 mL) at 0

°C. The reaction mixture was allowed to warm to rt and stirred for 30

min, and then 2-chloro-4-cyclopropyl-N-(5-hydroxy-2-adamantyl)-

thiazole-5-carboxamide (1.5 g, 4.25 mmol) was added. The reaction

mixture was stirred at rt for 16 h and then evaporated, and the residue

was taken up in DCM and 1 M aqueous citric acid. The organic layer

was isolated, evaporated, and puriﬁed by FCC, elution gradient 20−

100% EtOAc in heptane, to aﬀord the title compound 51 (1.41 g,

1

as a colorless solid. H NMR (400 MHz, DMSO) 1.41 (2H, d), 1.63

(4H, d), 1.67−1.8 (6H, m), 1.93−2.09 (7H, m), 2.14−2.24 (1H, m),

2.26−2.36 (1H, m), 3.45−3.52 (2H, m), 3.78−3.85 (3H, m), 3.87−

3.96 (2H, m), 4.42 (1H, s), 5.01−5.05 (2H, m), 8.51 (1H, d). m/z

MH⁺= 449. HPLC t_R= 1.99 min. HRMS (ES+) for C₂₃H₃₂N₂O₅S

(MH⁺): calcd 449.2105, found 449.2100.

Ethyl 2-Chloro-3-cyclopropyl-3-oxopropanoate. Sulfuryl chloride

(8.97 mL, 110 mmol) was added dropwise to a solution of ethyl 3-

cyclopropyl-3-oxopropanoate (17.2 g, 110 mmol) in DCM (150 mL)

at 0 °C. The reaction mixture was allowed to warm to rt and was

stirred at rt for 90 min and then quenched by pouring slowly into

water. The organic layer was isolated, the aqueous layer was extracted

with DCM, and the combined organic layers were washed with water

and saturated brine, dried over Na₂SO₄, and evaporated to aﬀord

impure ethyl 2-chloro-3-cyclopropyl-3-oxopropanoate (20.5 g, 98% if

pure) that was used without puriﬁcation or characterization.

1

80%) as a white solid. H NMR (400 MHz, DMSO) 0.86−0.97 (4H,

m), 1.2−1.25 (2H, m), 1.31−1.39 (4H, m), 1.57−1.64 (4H, m), 1.65−

1.72 (2H, m), 1.8−1.89 (2H, m), 1.95−2.01 (1H, m), 2.06−2.12 (2H,

m), 2.43−2.52 (1H, m), 3.82−3.88 (1H, m), 4.38 (2H, s), 4.39 (1H,

s), 7.57 (1H, d). m/z MH⁺= 414. HPLC t_R= 1.98 min. HRMS (ES+)

for C₂₂H₂₇N₃O₃S (MH⁺): calcd 414.1846, found 414.1842.

Human and Murine in Vitro 11β-HSD1 Inhibition: Homoge-

neous Time-Resolved Fluorescence Assay. The conversion of

cortisone to the active steroid cortisol by 11β-HSD1 oxoreductase

activity, can be measured by a competitive HTRF assay (CisBio

International, R&D, Administration and Europe Oﬃce, In Vitro

̀

Technologies - HTRF/Bioassays BP 84175, 30204 Bagnols/Ceze

Cedex, France; cortisol bulk HTRF kit, catalog no. 62CO2PEC). The

evaluation of compounds described herein was carried out with a

baculovirus-expressed N-terminal 6-His-tagged full-length human or

murine 11β-HSD1 enzyme. The enzyme was puriﬁed from a

detergent-solubilized cell lysate, on a copper chelate column.

Compounds to be tested were dissolved in DMSO to 10 mM and

diluted further in assay buﬀer containing 1% DMSO to 10-fold the

ﬁnal assay concentration. Diluted compounds were then plated into

black 384-well plates (Matrix, Hudson, NH). The assay was carried

out in a total volume of 20 μL consisting of cortisone (Sigma, Poole,

Dorset, U.K., 160 nM), glucose 6-phosphate (Roche Diagnostics, 1

mM), NADPH (Sigma, Poole, Dorset, U.K., 100 μM), glucose-6-

phosphate dehydrogenase (Roche Diagnostics, 12.5 μg/mL), ethyl-

enediaminetetraacetic acid (EDTA; Sigma, Poole, Dorset, U.K., 1

mM), assay buﬀer (K₂HPO₄/KH₂PO₄, 100 mM) pH 7.5, recombinant

human or murine 11β-HSD1 [at an appropriate dilution to give a

viable assay window; an example of a suitable dilution may be 1:1000

dilution of stock enzyme] plus test compound. The assay plates were

incubated for 25 min at 37 °C, after which time the reaction was

stopped by the addition of 10 μL of 0.5 mM glycerrhetinic acid plus

conjugated cortisol (D2). Anticortisol Cryptate (10 μL) was then

added, and the plates were sealed and incubated for 6 h at rt.

Fluorescence at 665 and 620 nm was measured and the 665/620 nm

ratio was calculated by use of an Envision plate reader. These data

were then used to calculate IC₅₀values for each compound (Origin

7.5, Microcal software, Northampton, MA).

Human in Vitro 11β-HSD1 Inhibition: High-Performance

Liquid Chromatography Assay. The assay was carried out in a total

volume of 1 mL consisting of cortisone (Sigma, Poole, Dorset, U.K.,

160 nM), ³H-cortisone (PerkinElmer, 20 nmol/L, 1 μCi/mL), glucose

6-phosphate (Roche Diagnostics, 1 mM), NADPH (Sigma, Poole,

Dorset, U.K., 100 μM), glucose-6-phosphate dehydrogenase (Roche

Diagnostics, 12.5 μg/mL), EDTA (Sigma, Poole, Dorset, U.K., 1 mM),

assay buﬀer (K₂HPO₄/KH₂PO₄, 100 mM) pH 7.5, recombinant 11β-

HSD1, diluted 10-fold over HTRF enzyme dilution (E/10), plus test

compound. The assay plates were incubated for 5 h at 37 °C, after

which time the reaction was stopped by the addition of 100 μL of 0.5

mM glycerrhetinic acid. Radiolabeled steroids were extracted with

EtOAc, and samples were evaporated to dryness under nitrogen and

resuspended in mobile phase for HPLC analysis (methanol/H₂O

50:50). Radiolabeled steroids were separated by reverse-phase HPLC

(Agilent 1200) on a Kromasil C18 5 μm column, 4.6 mm × 250 mm

(Crawford Scientiﬁc, Lanarkshire, U.K.) with methanol/H₂O (50:50)

at a ﬂow rate of 1.5 mL/min. Radioactivity was measured by use of a

ﬂow scintillation analyzer (Radiomatic series 500TR, Perkin-Elmer

Analytical Instruments) with FLO-ONE software.

Ethyl 2-Amino-4-cyclopropylthiazole-5-carboxylate. Thiourea

(4.06 g, 53.4 mmol) was added to a stirred solution of crude ethyl

2-chloro-3-cyclopropyl-3-oxopropanoate (10.2 g, 53.4 mmol if pure)

in EtOH (70 mL), and the resulting solution was stirred at 80 °C for

90 min. The resulting precipitate was isolated by ﬁltration, washed

with EtOH, and dried in vacuo to aﬀord the title compound (9.93 g,

88%) as a white crystalline solid, which was used without further

1

puriﬁcation. H NMR (400 MHz, DMSO) 0.89−0.98 (4H, m), 1.23

(3H, t), 2.80−2.88 (1H, m), 4.17 (2H, q), NH₂not assigned. m/z

MH⁺= 213. HPLC t_R= 1.69 min.

Ethyl 2-Chloro-4-cyclopropylthiazole-5-carboxylate. HCl in 1,4-

dioxane (4 M, 30 mL, 120 mmol) was added to a suspension of ethyl

2-amino-4-cyclopropylthiazole-5-carboxylate (9.70 g, 45.7 mmol) in

1,4-dioxane (50 mL), and the reaction mixture was stirred for 30 min

and then evaporated. THF (150 mL) was added to the resulting white

solid, and tert-butyl nitrite (8.22 mL, 68.6 mmol) was added dropwise

at 65 °C over a period of 1 h. The reaction mixture was stirred at 65

°C for 1 h and then poured into water and extracted with EtOAc. The

organic layer was isolated, dried over Na₂SO₄, evaporated, and puriﬁed

by FCC, elution gradient 0−20% EtOAc in heptanes, to aﬀord the title

1

compound (7.40 g, 70%) as a colorless solid. H NMR (400 MHz,

DMSO) 1.00−1.07 (2H, m), 1.14−1.20 (2H, m), 1.35 (3H, t), 2.93−

3.01 (1H, m), 4.36 (2H, q). m/z MH⁺= 232. HPLC t_R= 2.92 min.

2-Chloro-4-cyclopropylthiazole-5-carboxylic Acid. KOSi(CH₃)₃

(8.31 g, 64.7 mmol) was added to a solution of ethyl 2-chloro-4-

cyclopropylthiazole-5-carboxylate (5.00 g, 21.6 mmol) in THF (100

mL), and the reaction mixture was stirred at rt for 1 h. The reaction

mixture was quenched with 4 M HCl in 1,4-dioxane and evaporated,

and the resulting residue was taken up in DCM and water. The organic

layer was isolated and evaporated to aﬀord the title compound (4.39 g,

100%) as a pale yellow solid. ¹H NMR (400 MHz, DMSO) 0.91−0.97

(2H, m), 1.04−1.11 (2H, m), 2.94 (1H, m), CO₂H not observed. m/z

MH⁺= 204. HPLC t_R= 1.99 min.

2-Chloro-4-cyclopropyl-N-(5-hydroxy-2-adamantyl)thiazole-5-

carboxamide. HATU (11.6 g, 30.6 mmol), (1S,4R)-4-amino-

adamantan-1-ol hydrochloride (4.98 g, 24.5 mmol), and i-Pr₂NEt

(10.6 mL, 61.1 mmol) were added to 2-chloro-4-cyclopropylthiazole-

5-carboxylic acid (4.15 g, 20.4 mmol) in DCM (90 mL). The reaction

mixture was stirred at rt for 2 h and then was taken up in DCM and

water. The organic layer was isolated, dried over Na₂SO₄, evaporated,

and puriﬁed by FCC, elution gradient 0−100% EtOAc in heptanes, to

1

aﬀord the title compound (6.56 g, 91%) as a solid. H NMR (400

MHz, DMSO) 0.92−0.85 (m, 2H), 1.04−0.95 (m, 2H), 1.34 (d, 2H),

1.65−1.58 (m, 4H), 1.73−1.65 (m, 2H), 1.86 (d, 2H), 2.03−1.98 (m,

1H), 2.08 (s, 2H), 2.52−2.41 (m, 1H), 3.90−3.84 (m, 1H), 4.40 (s,

1H), 8.03 (d, 1H). m/z MH⁺= 353. HPLC t_R= 2.04 min.

982

dx.doi.org/10.1021/jm4016729 | J. Med. Chem. 2014, 57, 970−986

Journal of Medicinal Chemistry

Article

Procedures for Crystallization, Data Collection, and Struc-

ture Solution for 28 (4c7k) and 44 (4c7j) in Complex with

Human 11β-HSD1. Protein expression and puriﬁcation was done as

described previously.¹⁵The protein was concentrated to 8 mg/mL in a

buﬀer consisting of 20 mM N-(2-hydroxyethyl)piperazine-N′-

ethanesulfonic acid (HEPES) pH 8.0, 100 mM NaCl, 2 mM TCEP,

and 10 μM NADP. Prior to crystallization compounds were added to

the protein sample to yield a ﬁnal compound concentration of 2 μM.

The ﬁnal DMSO concentration was 1%. Crystals were grown by the

hanging drop vapor diﬀusion method. Protein (0.5 μL) was mixed

with 0.5 μL of well solution containing 20% PEG-3350 and 200 mM

ammonium nitrate and allowed to equilibrate over a reservoir

containing well solution. For data collection, crystals were brieﬂy

transferred to a well solution containing 20% glycerol and ﬂash-frozen

in liquid nitrogen. Diﬀraction data of 28 and 44 in complex with 11β-

HSD1 were collected at the European Synchrotron Radiation Facility

(ESRF) or on a rotating-anode X-ray source to resolutions of 1.9 and

2.2 Å, respectively. Data were processed with MOSFLM³⁶and scaled

and further reduced by use of the CCP4 suite of programs.³⁷Initial

phasing was done by molecular replacement using 4bb5¹⁵as a starting

model. The initial models were further reﬁned by alternative cycles of

model rebuilding in Coot³⁸and reﬁnement in Refmac5³⁹and/or

AutoBuster.⁴⁰In both data sets the F_o− F_cdiﬀerence maps showed

positive density in the h-11β-HSD1 active site, which allowed

unambiguous modeling of the bound ligands in three out of the

four molecules of the asymmetric unit, with mole bound inhibitors.

Data collection and reﬁnement statistics can be found in the

Supporting Information (Table S1).

Han Wistar Rat Central Nervous System Penetration Model.

Male Han Wistar rats were sourced from the AstraZeneca breeding

colony at Alderley Park, U.K. Compounds were dosed orally in a

HPMC/Tween (0.5% Methocel/0.1% polysorbate) vehicle at 30 mg/

kg (n = 2 rats/compound). Animals were euthanized at 1 h postdose,

and terminal blood (100 μL of blood +100 μL of water) and brain

samples were collected from each animal. Samples were frozen at −20

°C for storage prior to analysis. Blood samples (50 μL) were protein-

precipitated with ice-cold CH₃CN containing a generic internal

standard [(N-(1-adamantyl)-2-(1,1-dioxothiolan-3-yl)sulfanylpyridine-

3-carboxamide]. The samples were then mixed and centrifuged (4500g

for 10 min). A portion (50 μL) of the resulting supernatant was

removed and diluted with water (300 μL) for injection (50 μL) onto

the LC-MS/MS system. Brain samples were homogenized after

addition of water (4× dilution by weight) by use of a FisherBrand 500

homogenizer (ThermoFisher Scientiﬁc, Loughborough, U.K.), and in a

similar manner to above, a portion (50 μL) of the resulting

homogenate was then protein-precipitated with ice-cold CH₃CN

containing internal standard, centrifuged, and diluted with water prior

to injection onto the LC-MS/MS system. Blood and brain samples

were quantiﬁed against freshly prepared calibration curves with

standards at 1, 5, 10, 50, 100, 500, 1000, 2000, 5000, and 10 000

ng/mL, prepared in the same manner as above. Quality controls were

run before and after the unknown samples at concentrations of 10,

100, and 1000 ng/mL. The LC-MS/MS system is as described, with a

Surveyor MS Pump Plus HPLC pump (Thermo Fisher Scientiﬁc,

Hemel Hempstead, U.K.) and a CTC Analytics HTS PAL autosampler

(Presearch Ltd., Basingstoke, U.K.) used to introduce the samples to

the mass spectrometer. Chromatography was performed on a Max-RP

(50 mm × 2.1 mm i.d., 5 μM) HPLC column (Phenomenex,

Macclesﬁeld, U.K.) with mobile phase consisting of varying ratios of

water containing 0.1% formic acid and methanol containing 0.1%

formic acid, ﬂow rate 750 μL/min. Samples were analyzed on a TSQ

Quantum Vantage mass spectrometer (ThermoFisher Scientiﬁc,

Hemel Hempstead, U.K.). Parent and product masses were autotuned

into the mass spectrometer via QuickQuan (version 2.4) and Xcalibur

(version 2.1) software (ThermoFisher Scientiﬁc, San Jose, CA). The

capillary temperature was 270 °C, the vaporizer temperature was 300

°C, and the sheath gas and auxiliary gas ﬂows were set at 60 and 20

arbitrary units, respectively. Subsequent to analysis, the data were

quantiﬁed by use of QuickCalc software (Gubb Inc., Alpharetta, GA)

by back-calculation against the relevant calibration curve. A 1/x

weighting was applied to the calibration curves, which were

constructed via a quadratic ﬁt. Relevant dilution factors were applied

to the unknown samples (2-fold for blood samples, 4-fold for brain

samples) and a brain/blood ratio was calculated for each animal; this

ratio was used to benchmark the compounds. For the brain protein

binding protocol, rat protein binding and drug−brain tissue binding

were determined by equilibrium dialysis combined with liquid

chromatography following protocols in the literature;^41,42thus the

free brain/blood ratio could be determined.

Target Engagement in Rat. Determination of Corticosteroids in

Tissue Samples. Rat corticosteroid concentrations were measured by

published extraction methods.⁴³

Analysis of Plasma Samples for Drug Concentration. Rat blood

(50 μL) was protein-precipitated with 200 μL of ice-cold MeCN

containing internal standard. The samples were then centrifuged at

4500g for 10 min prior to dilution of a 50 μL aliquot of the

supernatant with 300 μL of deionized water. The resulting extract (50

μL) was injected onto the LC-MS/MS system for subsequent

quantiﬁcation. Unknown samples were measured against calibration

standards prepared from control rat blood and subsequently spiked

with analyte at known concentrations. Calibration standards were

subjected to the same extraction procedure as the unknowns.

Analysis of Brain Tissue Samples for Drug Concentration. Whole

rat brains were weighed and diluted 4× (w/v) with deionized water.

The samples were then homogenized by use of a homogenization

probe (Fisher Scientiﬁc, Loughborough, U.K.). The resulting

homogenate (50 μL) was protein-precipitated with 200 μL of ice-

cold MeCN containing an internal standard (a compound structurally

similar to the analytes from the AstraZeneca compound library). The

samples were then centrifuged at 4500g for 10 min prior to dilution of

a 50 μL aliquot of the supernatant with 300 μL of deionized water.

The resulting extract (50 μL) was injected onto the LC-MS/MS

system for subsequent quantiﬁcation. Unknown samples were

measured against calibration standards prepared from control brain

and subsequently spiked with analyte at known concentrations.

Calibration standards were subjected to the same extraction procedure

as the unknowns.

LC-MS/MS Conditions. The chromatography system consisted of a

HP1100 HPLC pump (Agilent, Cheadle, U.K.) and a HTS PAL

autosampler (Presearch Ltd., Basingstoke, U.K.). The HPLC column

used was a Synergy Max-RP 50 × 2.1 mm (Phenomenex, Macclesﬁeld,

U.K.) maintained at a constant temperature of 50 °C. The mobile

phase consisted of variable ratios of 0.1% formic acid in water and

0.1% formic acid in methanol, ﬂow rate 0.75 mL/min. A TSQ

Quantum Vantage MS ﬁtted with a heated ES probe (ThermoFisher

Scientiﬁc, Hemel Hempstead, U.K.). Compounds were auto-optimized

for tube lens voltage, parent ion m/z, collision energy, and product ion

m/z by use of QuickQuan version 2.1 software (ThermoFisher

Scientiﬁc, San Jose, CA). The MS settings were generic for all

compounds tested with spray voltage at 3500 V, vaporizer temperature

at 350 °C, sheath and auxiliary gas pressure at 60 and 20 arbitrary

units, respectively, capillary temperature at 270 °C, and collision cell

pressure at 1.5 mTorr.

Treatment of Results. Corticosterone (Cort) and 11-dehydrocorti-

costerone (11-DHC) concentration levels were used to generate a

Cort/11-DHC ratio for use in target engagement assessment.

Compound levels in plasma and brain were combined with

experimental protein binding values in the relevant tissues (see

previous section) to generate free drug concentrations for the

construction of Cort/11-DHC versus [free plasma] or [free brain]

plots for the compound of interest.

Diet-Induced Obese Mouse Compound-in-Diet Experiment.

Mice (C57B6, Alderley Park, U.K.) were housed in groups of three

upon arrival and allowed ad-lib access to water and control or 60 kcal

% fat rodent diet in meal form (D12492Mi Research Diet USA) for 30

weeks. Animals were acclimatized in scantainer and individually

housed prior to study phase. Body weight was monitored weekly

during high-fat feeding. At the start of study, animals were randomized

on the basis of body weight (n = 10 each group). Animals were

maintained on control or compound-containing diets for 25 days. Lean

983

dx.doi.org/10.1021/jm4016729 | J. Med. Chem. 2014, 57, 970−986

Journal of Medicinal Chemistry

Article

controls were maintained on 10 kcal % fat control diet (D12450 Bi,

Research Diets USA). Compounds were formulated 1 day prior to

beginning of live phase. Compounds 51 (240 mg) and 40 (48, 144, or

480 mg) were weighed out into a marble mortar and crushed with

pestle before addition to preweighed powdered diet (30 g). After

mixing in a blender for 10 min, the compound/diet mixture was

compressed into balls of approximately 28−30 g each and stored at 4

°C until use. Individual animals were fed 2.5 g of CID mix a day, and

the chow mix was checked to ensure consumption. During the

treatment phase, body weight and food intake were monitored 3 times

a week at 8 a.m. (1 h before lights oﬀ). On day 21, following a 13 h

fast, blood samples were taken by tail-prick for glucose measurements

(Accu-Chek strip and read by Aviva Accu-Chek monitor). Each

sampling was done within 2 min to minimize disturbance to the

animals. Animals were re-fed following the procedures. Following a

short recovery, on day 24 plasma exposure of individual animal was

analyzed at 11 a.m. and 3 p.m.. Animals were terminated on day 25 at

9 a.m. without fast. Animals were euthanized by rising CO₂, and a

blood sample was taken via cardiac bleed immediately. Plasma was

frozen down for compound exposure, and liver samples were dissected

for ex vivo enzyme activity measurement.

adjusted to a concentration of 3.5 × 10⁵cells/mL (350 000), and then

cell suspension was added to each well (75 μL). Test compounds (25

μL) were transferred to give ﬁnal concentrations of 10 and 1 μM, and

the plate was incubated at 37 °C, 5% CO₂for 72 h. After 72 h, a

steadylite HTS assay kit was prepared (PerkinElmer 6016987, 500 mL

kit) as per instructions to yield reconstituted substrate. The medium

was aspirated oﬀ the plate and the cells were washed twice at rt with

Dulbecco’s PBS buﬀer (200 μL/well). Reconstituted substrate was

added (100 μL/well) to the culture plate, and the plate was agitated

gently under subdued light conditions. The plate was sealed and

incubated for 5 min in the dark at rt and then transferred to a plate

reader, and luminescence was measured. The result was expressed as a

percentage of control at the two concentrations.

ASSOCIATED CONTENT

■

S

* Supporting Information

Additional text describing preparation and characterization for

additional ﬁnal compounds, and two tables listing data

collection and reﬁnement statistics for X-ray crystallography

and parameters for PK/PD modeling. This material is available

free of charge via the Internet at http://pubs.acs.org.

Ex Vivo Enzyme Activity. Liver samples were stored at −80 °C

prior to assay. Liver samples were weighed (150−200 mg) and

transferred to a 24-well plate on ice. Warmed medium [1 mL of

Dulbecco’s modiﬁed Eagle medium (DMEM) Ham F12 medium

supplemented with 1% penicillin/streptomycin/ampicillin and 10%

fetal calf serum (FCS)] containing 20 nM [³H]cortisone (1 μCi/mL,

∼20 nM ﬁnal assay concentration, NET1178001MC, lot no. 3632365,

Perkin-Elmer) was added to the well, and tissue was cut into 2−3 mm³

pieces with scissors. Liver samples were incubated for 10 min at 37 °C

in 95% O₂/5% CO₂. Following incubation, medium from liver assays

was transferred to an Eppendorf tube and centrifuged (2 min, 3600

rpm), and the resulting supernatant was transferred to a 5-in. × ⁵/₈-in.

glass test tube prior to extraction. Extraction of radiolabeled steroids

was carried out by the addition of 2.5 mL of EtOAc to each sample.

Following a 10 s vortex mixing, the top solvent layer of each sample,

Accession Codes

The crystal structures of human 11β-HSD1 in complex with 28

and with 44 have been deposited in the Protein Data Bank

(PDB accession codes 4c7k and 4c7j).

AUTHOR INFORMATION

■

Corresponding Authors

*Phone +44(0)1625519064; e-mail frederick.goldberg@

astrazeneca.com.

*Phone +44(0)7720685539; e-mail al.dossetter@medchemica.

com.

4

containing the tritiated steroids, was transferred to a 4-in. × /₈-in.

Notes

glass tube. EtOAc was evaporated oﬀ at 50 °C, under nitrogen.

Samples were resuspended in 140 μL of 1:1 MeOH/water (mobile

phase). Samples were run on a HPLC-linked scintillation counter

[Agilent 1200 system and Perkin-Elmer Flo-Scintillation analyzer

(500TR)] with a Kromasil C18 5 μm column (4.6 mm × 250 mm).

Each sample was run for 12 min at 50 °C with 1:1 MeOH/water at 1.5

mL/min.

The authors declare no competing ﬁnancial interest.

ACKNOWLEDGMENTS

■

We thank Brendan Leighton for his leadership of the bioscience

strategy and for providing references, Peter Ceuppens for

statistical support, Alice Yu for in vivo bioscience support,

Elizabeth Kelly for running the DIO study, Pam Tyers for

Pharmacokinetics. On day 24 at 2 and 6 h from lights oﬀ, 20-μL

blood samples were taken via conscious tail-prick and diluted

immediately in 80 μL of phosphate-buﬀered saline (PBS) and

centrifuged at 6000g for 10 min at 4 °C. Terminal blood sample

was taken on day 25 at the time of lights oﬀ. Neat and diluted plasma

samples were transferred to a 96-well plate and analyzed for

compound exposure. HOMA-IR was deﬁned from [fasting glucose

(millimolar) × fasting insulin (milliunits per liter)]/22.5.

Pregnane X Receptor Inhibition Assay. A medium for cells was

prepared: add supplements to DMEM-phenol red free (Mediatech)

and warm at 37 °C in water bath [add 5 mL of nonessential amino

acids (Mediatech) + 5 mL of 2 mM ^L-glutamine (InVitrogen) + 50 mL

of 10% charcoal-stripped FBS (Hyclone)]. Test compounds were

prepared from 10 mM stock solution DMSO by dilution to ﬁnal

concentrations of 10 and 1 μM. Initially compounds were diluted on

plate to 4× the ﬁnal assay concentration (40 μM) in HepG2 medium

containing DMSO, and so 10 μL of stock was taken and diluted in

2490 μL of medium for a 250× dilution. The compound (1000 μL of

40 μM solution) was placed in a deep well place in one well, with 900

μL of blank medium in subsequent wells. Stock concentration of

rifampicin (10 mM) and dexamethasone (30 mM) were made up.

HepG2 cells, CryoTES (Applied Cell Sciences), were thawed by

placing vials in a 37 °C water bath and HepG2 medium was added

slowly to bring up to 40 mL volume, and then the samples were

centrifuged and resuspended in 5 mL of HepG2 medium. Cell density

and viability was determined by the trypan blue method and thus

́

generating high-quality HPLC data, and Gisela Branden and

̈

Cristian Bodin for their contributions to the X-ray crystallog-

raphy work.

ABBREVIATIONS USED

■

CID, compound in diet; CDI, 1,1′-carbonyldiimidazole; DIAD,

diisopropyl azodicarboxylate; DIO, diet-induced obese mice;

FCC, ﬂash column chromatography; HATU, 2-(3H-[1,2,3]-

triazolo[4,5-b]pyridin-3-yl)-1,1,3,3-tetramethylisouronium

hexaﬂuorophosphate(V); hBA1c, glycated hemoglobin; HFD,

60% high fat diet of chow; HOMA-IR, homeostasis model

assessment−insulin resistance; HPMC, hydroxypropyl methyl-

cellulose; HRMS, high-resolution mass spectrometry; 11β-

HSD1, 11β-hydroxysteroid dehydrogenase type 1; h-11β-

HSD1, human isoform of 11β-hydroxysteroid dehydrogenase

type 1; m-11β-HSD1, murine isoform of 11β-hydroxysteroid

dehydrogenase type 1; HTRF, homogeneous time-resolved

ﬂuorescence; LLE, ligand lipophilicity eﬃciency; MDCK,

Madin-Darby canine kidney; NADP(H), nicotinamide adenine

dinucelotide phosphate (reduced); PEG, poly(ethylene glycol);

POMC, proopiomelanocortin

984

dx.doi.org/10.1021/jm4016729 | J. Med. Chem. 2014, 57, 970−986

Journal of Medicinal Chemistry

Article

(17) Savontaus, E.; Conwell, I. M.; Wardlaw, S. L. Effects of

adrenalectomy on AGRP, POMC, NPY and CART gene expression in

the basal hypothalamus of fed and fasted rats. Brain Res. 2002, 958,

130−138.

REFERENCES

■

(1) Edwards, C. R. W.; Benediktsson, R.; Lindsay, R. S.; Seckl, J. R.

11β-Hydroxysteroid dehydrogenases: key enzymes in determining

tissue-specific glucocorticoid effects. Steroids 1996, 61, 263−269.

(2) Walker, E. A.; Stewart, P. M. 11β-Hydroxysteroid dehydrogenase:

unexpected connections. Trends Endocrinol. Metab. 2003, 14, 334−

339.

́

(18) Beaulieu, S.; Gagne, B.; Barden, N. Glucocorticoid regulation of

proopiomelanocortin messenger ribonucleic acid content of rat

hypothalamus. Mol. Endocrinol. 1988, 2, 727−731.

(19) Scott, J. S.; Bowker, S. S.; deSchoolmeester, J.; Gerhardt, S.;

Hargreaves, D.; Kilgour, E.; Lloyd, A.; Mayers, R. M.; McCoull, W.;

Newcombe, N. J.; Ogg, D.; Packer, M. J.; Rees, A.; Revill, J.; Schofield,

P.; Selmi, N.; Swales, J. G.; Whittamore, P. R. O. Discovery of a

potent, selective, and orally bioavailable acidic 11β-hydroxysteroid

dehydrogenase type 1 (11β-HSD1) inhibitor: Discovery of 2-[(3S)-1-

[5-(cyclohexylcarbamoyl)-6-propylsulfanylpyridin-2-yl]-3-piperidyl]-

acetic acid (AZD4017). J. Med. Chem. 2012, 55, 5951−5964.

(20) Scott, J. S.; deSchoolmeester, J.; Kilgour, E.; Mayers, R. M.;

Packer, M. P.; Hargreaves, D.; Gerhardy, S.; Ogg, D. J.; Rees, A.; Selmi,

N.; Stocker, A.; Swales, J. G.; Whittamore, P. R. O. Novel acidic 11β-

hydroxysteroid dehydrogenase type 1 (11β-HSD1) inhibitor with

reduced acyl glucuronide liability: The discovery of 4-[4-(2-

adamantylcarbamoyl)-5-tert-butyl-pyrazol-1-yl]benzoic acid

(AZD8329). J. Med. Chem. 2012, 55, 10136−10147.

(3) Strain, G. W.; Zumoff, B.; Strain, J. J.; Levin, J.; Fukushima, D. K.

Cortisol production in obesity. Metab., Clin. Exp. 1980, 29, 980−985.

(4) Thieringer, R.; Hermanowski-Vosatka, A. Inhibition of 11β-

HSD1 as a novel treatment for the metabolic syndrome: do

glucocorticoids play a role? Expert Rev. Cardiovasc. Ther. 2005, 3,

911−924.

(5) Walker, B. R. Glucocorticoids and cardiovascular disease. Eur. J.

Endocrinol. 2007, 157, 545−559.

(6) Tomlinson, J. W.; Walker, E. A.; Bujalska, I. J.; Draper, N.;

Lavery, G. G.; Cooper, M. S.; Hewison, M.; Stewart, P. M. 11β-

Hydroxysteroid dehydrogenase type 1: a tissue-specific regulator of

glucocorticoid response. Endocr. Rev. 2004, 25, 831−866.

(7) Yau, J. L.; Noble, J.; Kenyon, C. J.; Hibberd, C.; Kotelevtsev, Y.;

Mullins, J. J.; Seckl, J. R. Lack of tissue glucocorticoid reactivation in

11β-hydroxysteroid dehydrogenase type 1 knockout mice ameliorates

age-related learning impairments. Proc. Natl. Acad. Sci. U.S.A. 2001, 98,

4716−4721.

(8) Holmes, M. C.; Carter, R. N.; Noble, J.; Chitnis, S.; Dutia, A.;

Paterson, J. M.; Mullins, J. J.; Seckl, J. R.; Yau, J. L. 11β-Hydroxysteroid

dehydrogenase type 1 expression is increased in the aged mouse

hippocampus and parietal cortex and causes memory impairments. J.

Neurosci. 2010, 30, 6916−6920.

(9) Sandeep, T. C.; Yau, J. L.; MacLullich, A. M.; Noble, J.; Deary, I.

J.; Walker, B. R.; Seckl, J. R. 11β-Hydroxysteroid dehydrogenase

inhibition improves cognitive function in healthy elderly men and type

2 diabetics. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 6734−6739.

(10) Scott, J. S.; Goldberg, F. W.; Turnbull, A. V. Medicinal

chemistry of 11β-hydroxysteroid dehydrogenase type 1 (11β-HSD1).

J. Med. Chem. [Online early access]. DOI: 10.1021/jm4014746.

Published online: December 2, 2013.

(11) Anagnostis, P.; Katsiki, N.; Adamidou, F.; Athyros, V. G.;

Karagiannis, A.; Kita, M.; Mikhailidis, D. P. 11β-Hydroxysteroid

dehydrogenase type 1 inhibitors: novel agents for the treatement of

metabolic syndrome and obesity-related disorders? Metabolism 2013,

52, 21−33.

(12) Rosenstock, J.; Banarer, S.; Fonseca, V. A.; Inzucchi, S. E.; Sun,

W.; Yao, W.; Hollis, G.; Flores, R.; Levy, R.; Williams, W. V.; Seckl, J.

R.; Huber, R. The 11-β-hydroxysteroid dehydrogenase type 1 inhibitor

INCB13739 improves hyperglycemia in patients with type 2 diabetes

inadequately controlled by metformin monotherapy. Diabetes Care

2010, 33, 1516−1522.

(13) Bauman, D. R.; Whitehead, A.; Contino, L. C.; Cui, J.; Garcia-

Calvo, M.; Gu, X.; Kevin, N.; Ma, X.; Pai, L.-Y.; Shah, K.; Shen, X.;

Stribling, S.; Zokian, H. J.; Metzger, J.; Shevell, D. E.; Waddell, S. T..

Evaluation of selective inhibitors of 11β-HSD1 for the treatment of

hypertension. Bioorg. Med. Chem. Lett. 2013, 23, 3650−3653.

(14) Shah, S.; Hermanowski-Vosatka, A.; Gibson, K.; Ruck, R. A.; Jia,

G.; Zhang, J.; Hwang, P. M. T.; Ryan, N. W.; Langdon, R. B.; Feig, P.

U. Efficacy and safety of the selective 11β-HSD1 inhibitors MK-0736

and MK-0916 in overweight and obese patients with hypertension. J.

Am. Soc. Hypertens. 2011, 5, 166−176.

(15) Goldberg, F. W.; Leach, A. G.; Scott, J. S.; Snelson, W. L.;

Groombridge, S. D.; Donald, C. S.; Bennett, S. N. L.; Bodin, C.;

Gutierrez, P. M.; Gyte, A. C. Free-Wilson and structural approaches to

co-optimizing human and rodent isoform potency for 11β-

hydroxysteroid dehydrogenase type 1 (11β-HSD1) inhibitors. J. Med.

Chem. 2012, 55, 10652−10661.

(16) Bisschop, P. H.; Dekker, M. J.; Osterthun, W.; Kwakkel, J.;

Anink, J. J.; Boelen, A.; Unmehopa, U. A.; Koper, J. W.; Lamberts, S.

W.; Stewart, P. M.; Swaab, D. F.; Fliers, E. Expression of 11β-

hydroxysteroid dehydrogenase type 1 in the human hypothalamus. J.

Neuroendocrinol. 2013, 25, 425−32.

(21) Scott, J. S.; Gill, A. L.; Godfrey, L.; Groombridge, S. D.; Rees,

A.; Revill, J.; Schofield, P.; Sorme, P.; Stocker, A.; Swales, J. G.;

̈

Whittamore, P. R. O. Optimisation of pharmacokinetic properties in a

neutral series of 11β-HSD1 inhibitors. Bioorg. Med. Chem. Lett. 2012,

22, 6756−6761.

(22) Di, L.; Rong, H.; Feng, B. Demystifying brain penetration in

central nervous system drug discovery. J. Med. Chem. 2013, 56, 2−12.

(23) The absolute conﬁguration of 3 was established from synthesis

from (2R)-tetrahydrofuran-2-carboxylic acid. The corresponding S

analogue was much less potent against both the human (IC₅₀1.2 μM)

and mouse (IC₅₀2.1 μM) forms of 11β-HSD1.

(24) Desai, P. V.; Raub, T. J.; Blanco, M.-J. How hydrogen bonds

impact P-glycoprotein transport and permeability. Bioorg. Med. Chem.

Lett. 2012, 22, 6540−6548.

(25) Heffron, T. P.; Salphati, L.; Alicke, B.; Cheong, J.; Dotson, J.;

Edgar, K.; Goldsmith, R.; Gould, S. E.; Lee, L. B.; Lesnick, J. D.; Lewis,

C.; Ndubaku, C.; Nonomiya, J.; Olivero, A. G.; Pang, J.; Plise, E. G.;

Sideris, S.; Trapp, S.; Wallin, J.; Wang, L.; Zhang, X. The design and

identification of brain penetrant inhibitors of phosphoinositide 3-

kinase α. J. Med. Chem. 2012, 55, 8007−8020.

(26) Bissantz, C.; Kuhn, B.; Stahl, M. A medicinal chemist’s guide to

molecular interactions. J. Med. Chem. 2010, 53, 5061−5084.

(27) Kotelevtsev, Y.; Brown, R. W.; Fleming, S.; Kenyon, C.;

Edwards, C. R. W.; Seckl, J. R.; Mullins, J. J. Hypertension in mice

lacking 11β-hydroxysteroid dehydrogenase type 2. J. Clin. Invest. 1999,

103, 683−689.

(28) Wan, Z.; Chenail, E.; Xiang, J.; Li, H.; Ipek, M.; Bard, J.;

Svenson, K.; Mansour, T. S.; Xu, X.; Tian, X.; Suri, V.; Hahm, S.; Xin,

Y.; Johnson, C. E.; Li, X.; Qadri, A.; Panza, D.; Perreault, M.; Tobin, J.

F.; Saiah, E. Efficacious 11β-hydroxysteroid dehydrogenase type I

inhibitors in the diet-induced obesity mouse model. J. Med. Chem.

2009, 52, 5449−5461.

(29) Harno, E.; Cottrell, E. C.; Yu, A.; deSchoolmeester, J.; Gutierrez,

P. M.; Denn, M.; Swales, J. G.; Goldberg, F. W.; Bohlooly-Y, M.;

Andersen, H.; Wild, M. J.; Turnbull, A. V.; Leighton, B.; White, A.

́

11β-HSD1 inhibitors still improve metabolic phenotype in male 11β-

HSD1 knock-out mice suggesting off-target mechanisms. Endocrinology

2013, 154 (10), 4580−4593.

(30) Fotsch, C.; Bartberger, M. D.; Bercot, E. A.; Chen, M.; Cupples,

R.; Emery, M.; Fretland, J.; Guram, A.; Hale, C.; Han, N.; Hickman,

D.; Hungate, R. W.; Hayashi, M.; Komorowski, R.; Liu, Q.;

Matsumoto, G.; St. Jean, D. J., Jr.; Ursu, S.; Veniant, M.; Xu, G.; Ye,

Q.; Yuan, C.; Zhang, J.; Zhang, X.; Tu, H.; Wang, M. Further studies

with the 2-amino-1,3-thiazol-4(5H)-one class of 11beta-hydroxysteroid

dehydrogenase type 1 inhibitors: reducing pregnane X receptor activity

and exploring activity in a monkey pharmacodynamic model. J. Med.

Chem. 2008, 51, 7953−7967.

985

dx.doi.org/10.1021/jm4016729 | J. Med. Chem. 2014, 57, 970−986

Journal of Medicinal Chemistry

Article

(31) Zhu, Y.; Olson, S. H.; Hermanowski-Vosatka, A.; Mundt, S.;

Shah, K.; Springer, M.; Thieringer, R.; Wright, S.; Xiao, J.; Zokian, H.;

Balkovec, J. M. 4-Methyl-5-phenyl triazoles as selective inhibitors of

11β-hydroxysteroid dehydrogenase type I. Bioorg. Med. Chem. Lett.

2008, 18, 3405−3411.

(32) Gao, Y. D.; Olson, S. H.; Balkovec, J. M.; Zhu, Y.; Royo, I.;

Yabut, J.; Evers, R.; Tan, E. Y.; Tang, W.; Hartley, D. P.; Mosley, R. T.

Attenuating pregnane X receptor (PXR) activation: A molecular

modelling approach. Xenobiotica 2007, 37, 124−138.

(33) Marino, J. P., Jr.; Mcatee, J. J.; Washburn, D. G. Preparation of

nitrogen heterocyclic compounds, particularly substituted piperidines

and pyrrolidines, as soluble epoxide hydrolase (sEH) and 11β-

hydroxysteroid dehydrogenase type 1 (11β-HSD1) inhibitors and their

use. World Patent WO2009070497, 2009.

(34) Iyer, A.; Kauter, K.; Alam, M. A.; Hwang, S. H.; Morisseau, C.;

Hammock, B. D.; Brown, L. Pharmacological inhibition of soluble

epoxide hydrolase ameliorates diet-induced metabolic syndrome in

rats. Exp. Diabetes Res. 2012, 758614.

(35) Luria, A.; Bettaieb, A.; Xi, Y.; Shieh, G. J.; Liu, H. C.; Inoue, H.;

Tsai, H. J.; Imig, J. D.; Haj, F. G.; Hammock, B. D. Soluble epoxide

hydrolase deficiency alters pancreatic islet size and improves glucose

homeostasis in a model of insulin resistance. Proc. Natl. Acad. Sci.

U.S.A. 2011, 108, 9038−9043.

(36) Leslie, A. G. Integration of macromolecular diffraction data. Acta

Crystallogr. 2006, D62, 48−57.

(37) Collaborative Computational Project 4.. The CCP4 suite:

Programs for protein crystallography. Acta Crystallogr. 1994, D50,

750−763.

(38) Emsley, P.; Cowtan, K. Coot: model-building tools for

molecular graphics. Acta Crystallogr. 2004, D60, 2126−2132.

(39) Murshudov, G. N.; Vagin, A. A.; Lebedev, A.; Wilson, K. S.;

Dodson, E. J. Efficient anisotropic refinement of macromolecular

structures using FFT. Acta Crystallogr. 1999, D55, 247−255.

(40) Blanc, E.; Roversi, P.; Vonrhein, C.; Flensburg, C.; Lea, S. M.;

Bricogne, G. Refinement of severely incomplete structures with

maximum likelihood in BUSTER-TNT. Acta Crystallogr. 2004, D60,

2210−2221.

(41) Wan, H.; Rehngren, M. High-throughput screening of protein

binding by equilibrium dialysis combined with liquid chromatography

and mass spectrometry. J. Chromatogr. A 2006, 1102, 125−134.

(42) Wan, H.; Rehngren, M.; Giordanetto, F.; Bergstrom, F.; Tunek,

A. High-throughput screening of drug-brain tissue binding and in silico

prediction for assessment of central nervous system drug delivery. J.

Med. Chem. 2007, 50, 4606−4615.

(43) Ronquist-Nii, Y.; Edlund, P. O. Determination of corticosteroids

̈

in tissue samples by liquid chromatography−tandem mass spectrom-

etry. J. Pharm. Biomed. Anal. 2005, 37, 341−350.

986

dx.doi.org/10.1021/jm4016729 | J. Med. Chem. 2014, 57, 970−986

Article Doi

Optimization of brain penetrant 11β-hydroxysteroid dehydrogenase type i inhibitors and in vivo testing in diet-induced obese mice

DOI: 10.1021/jm4016729

Source and publish data:

Authors:

Article abstract of DOI:10.1021/jm4016729

Full text of DOI:10.1021/jm4016729

Products guided by the article

R&D Labs maybe for 1161825-34-8

Relevant to this article

Hot Product