Highly potent, selective, and orally active phosphodiesterase 10A inhibitors

ARTICLE

pubs.acs.org/jmc

Highly Potent, Selective, and Orally Active Phosphodiesterase

10A Inhibitors^†

Michael S. Malamas,*^,‡Yike Ni,^‡James Erdei,^‡Hans Stange,^||Rudolf Schindler,^||Hans-Joachim Lankau,^||

Christian Grunwald,^||Kristi Yi Fan,^‡,#Kevin Parris,^‡Barbara Langen,^||Ute Egerland,^||Thorsten Hage,^||

Karen L. Marquis,^‡Steve Grauer,^‡Julie Brennan,^‡Rachel Navarra,^‡Radka Graf,^‡Boyd L. Harrison,^‡

Albert Robichaud,^{‡,^}Thomas Kronbach,^||Menelas N. Pangalos,^‡,∞Norbert Hoefgen,^||and

Nicholas J. Brandon^§

^‡Pﬁzer Neuroscience Princeton, 865 Ridge Road, Monmouth Junction, New Jersey 08852, United States

^§Pﬁzer Neuroscience, Eastern Point Road, Groton Connecticut 06340, United States

Biocrea Therapies GmbH, Meissner Strasse 191, 01445 Radebeul, Germany

S Supporting Information

b

ABSTRACT: The identiﬁcation of highly potent and orally

active phenylpyrazines for the inhibition of PDE10A is re-

ported. The new analogues exhibit subnanomolar potency for

PDE10A, demonstrate high selectivity against all other mem-

bers of the PDE family, and show desired druglike properties.

Employing structure-based drug design approaches, we meth-

odically explored two key regions of the binding pocket of the

PDE10A enzyme to alter the planarity of the parent compound 1 and optimize its aﬃnity for PDE10A. Bulky substituents at the C9

position led to elimination of the mutagenicity of 1, while a crucial hydrogen bond interaction with Glu716 markedly enhanced its

potency and selectivity. A systematic assessment of the ADME and PK properties of the new analogues led to druglike development

candidates. One of the more potent compounds, 96, displayed an IC₅₀for PDE10A of 0.7 nM and was active in predictive

antipsychotic animal models.

’ INTRODUCTION

is a dual substrate PDE.^6ꢀ9PDE10A hydrolyzes both cAMP and

cGMP, with a higher aﬃnity for cAMP (K_m= 0.05 μM) than for

cGMP (K_m=3 μM).¹⁰PDE10A is primarily a membrane bound

enzyme containing a catalytic domain in the C-terminal portion

of the protein. Key residues of the catalytic core form a well-

deﬁned hydrophobic clamp region that positions the planar rings

of the nucleotide for interaction with an absolutely conserved

glutamine residue (Gln716 in PDE10A). Free rotation of this

glutamine in PDE10A, and other dual substrate enzymes, is

believed to allow the binding of either cAMP or cGMP in the

substrate pocket.¹¹PDE10A mRNA is highly expressed only in

brain and testes.^7,8In the brain, both PDE10A mRNA and

protein are speciﬁcally enriched in the medium spiny neurons

(MSNs) of the striatum.¹²PDE10A has been suggested to play a

key role in regulating MSN activity and in turn striatal output, as

this region integrates dopaminergic and glutamatergic inputs

from midbrain and cortical regions, respectively, to action

motoric and cognitive function. Dysfunction in corticalꢀstriatal

neurotransmission has been implicated in the pathophysiology of

schizophrenia; thus, PDE10A inhibition has been suggested as a

therapeutic strategy for this disease.^13ꢀ15

Cyclic nucleotide phosphodiesterases (PDEs) are key regulators

of cellular signal transduction by hydrolyzing the 3⁰,5⁰-monopho-

sphate bond of the ubiquitous intracellular second messengers

adenosine 3⁰,5⁰-monophosphate (cAMP) and cyclic guanosine

3⁰,5⁰-monophosphate (cGMP). Both cAMP and cGMP are

important messengers in the signaling cascades of G-protein-

coupled receptors, and modulation of their intracellular levels

inﬂuences the magnitude and duration of cellular responses to

hormonal and neurotransmitter stimulation. There are 11 dis-

tinct phosphodiesterases, designated as PDE1ꢀPDE11. The

PDEs are encoded by 21 diﬀerent genes in humans, with over

50 isoforms expressed as a result of alternative splicing patterns.¹

PDEs are classiﬁed by their substrate speciﬁcity. Some selectively

hydrolyze cAMP or cGMP, while others will accept both cyclic

nucleotides as substrates.^2,3The PDE families diﬀer in the sub-

strate speciﬁcity for the cyclic nucleotides, the mechanism of

regulation, and the sensitivity to inhibitors. They are also local-

ized to speciﬁc subcellular sites, which allows for ﬁne spatial and

temporal control of the levels of the cyclic nucleotides.^4,5This

feature is thought to be an important contributing factor that

allows the enzyme to inﬂuence selective intracellular signaling

pathways in response to diﬀerent stimuli, in spite of the ubiqui-

tous intracellular distribution of the cyclic nucleotides. PDE10A

Received: July 15, 2011

Published: October 11, 2011

r

2011 American Chemical Society

7621

dx.doi.org/10.1021/jm2009138 J. Med. Chem. 2011, 54, 7621–7638

|

Journal of Medicinal Chemistry

ARTICLE

Figure 1

In a recent disclosure,¹⁶we have described the design of new

pyrido[2,3-e]pyrazines (1, Figure 1) as potent and selective

inhibitors of PDE10A. The reported research has resulted in

the discovery of novel templates for PDE10A inhibition and has

delineated key structural elements of the ligand and protein

interactions. These initial structureꢀactivity relationship (SAR)

studies have paved the path to a better understanding of the

requirements needed to enhance the ligand’s potency and

selectivity. X-ray crystallography and molecular modeling calcu-

lations were employed to assist with the design of new PDE10A

inhibitors. The initial research eﬀort has also encountered some

shortcomings related to the pharmacokinetic properties and

safety concerns because of the planar structure of the chemical

scaﬀold that may hamper further development of these PDE10A

inhibitors. In addition, we were aiming to further optimize the

overall biology proﬁle of the initial series, like target aﬃnity and

selectivity and in vitro potency. Armed with this information, we

have employed a chemistry strategy to address these issues and

further improve the pharmacokinetic and safety properties of our

PDE10A inhibitors. Our key focus centered on the potential risk

of mutagenicity of the early compounds. Activation of strain

TA1537 (AMES test) should be indicative for ligand intercala-

tion to DNA. Intercalation occurs when ligands of an appropriate

size and chemical nature align between the base pairs of DNA.

These ligands are mostly polycyclic, aromatic, and planar and

therefore able to insert into the hydrophobic environment found

between the base pairs. These structural modiﬁcations can lead to

functional changes, often to the inhibition of transcription and

replication and DNA repair processes, which make intercalators

potent mutagens. The pyrido[2,3-e]pyrazine scaﬀold of our early

inhibitors represents a fused polycyclic planar structure, which

potentially intercalates with DNA. Initially, we have designed a

new template to alter the three-dimensional structure of the

parent compound 1. Examination of the X-ray crystal structure

of 1:PDE10A¹⁶(Figure 2) revealed the presence of a large

unoccupied region of the ligand binding pocket. We intended

to take advantage of this observation and introduce bulky groups

at the C9 position of the fused tricyclic ring system of the parent

compound (Figure 1) in an attempt to alter the overall three-

dimensional planarity of the parent compound. Such structural

arrangement will break the three-dimensional planarity of the

parent structure and prevent its intercalation with DNA. Further-

more, our chemistry strategy will focus on improving the

metabolic stability and pharmacokinetic properties of the parent

compound 1. Isolation and characterization of the metabolites of

1 have indicated the areas of the molecule (2-OMe, 6- and 7-Me,

9-propyl), which were prone to metabolic oxidation (Figure 3).

Masking of the metabolically labile alkyl functionalities of the

parent compound 1 with more stable ﬂuorine-containing groups

will result in metabolic stability enhancement. In order to achieve

these objectives in a timely manner, we have designed a uniﬁed

chemistry methodology (Schemes 1ꢀ4). This new synthetic

Figure 2. Crystal structure of PDE10A (rat) complexed with 1,¹⁶

highlighting the key hydrogen-bonding interactions between ligand

and protein at the invariant Glu716, and the water-mediated contacts

with Asp664 and Tyr514. The fused ring system of the ligand is

sandwiched between residues Phe719, Ile682, and Phe686, making

π-stacking interactions with Phe719, π-edge stacking with Phe686,

and hydrophobic contact with Ile682.

Figure 3. Metabolic disposition of 1 (shown are identiﬁed metabolites).

strategy enables us to introduce multiple group modiﬁcations on

common synthetic blocks using parallel synthesis techniques.

The scope of this paper will cover the advancements in safety,

physicochemical properties, and pharmacokinetic properties of

the parent compound 1, as well as further improvements in

ligand potency, selectivity, and in vivo eﬃcacy.

’ CHEMISTRY

The compounds needed to delineate the SAR for this study

were prepared according to synthetic Schemes 1ꢀ4. To expedite

SAR development, we have designed a synthetic strategy to allow

for parallel synthesis, enabling us to rapidly explore the C9

position of the parent template (compound 1), as well as new

scaﬀolds. An important process throughout the SAR studies was

the preparation of bromoimidazole building blocks (i.e., 6),

which upon coupling with aryl groups, by employing the

palladium-catalyzed cross-coupling Suzuki protocol, produced

the target compounds. In Scheme 1, 2-chloropyridine 2 was

treated with 4-methylimidazole in the presence of potassium

hydroxide or potassium carbonate to aﬀord the coupled product

3 in 4:1 isomeric ratio. Reduction of the nitro group of 3a with

Pd/C and ammonium formate produced aniline 3b, which upon

treatment with acetic anhydride in toluene furnished amide 4.

Cyclization of 4 to the tricyclic intermediate 5 was accomplished

with phosphorus oxychloride and phosphorus pentoxide. Bro-

mination of imidazole 5 with N-bromosuccinimide in acetonitrile

produced bromoimidazole 6, which was used for the preparation

7622

dx.doi.org/10.1021/jm2009138 |J. Med. Chem. 2011, 54, 7621–7638

Journal of Medicinal Chemistry

ARTICLE

Scheme 1^a

^aReagents: (a) KOH or K₂CO₃, DMF; (b) Pd/C, HCO₂NH₄, MeOH; (c) AC₂O, toluene; (d) P₂O5, POCl₃; e) N-bromosuccinimide, CH₃CN;

(f) R₉B(OH)₂, K₂CO₃, Pd(PPh₃)₄, dioxane/H₂O, 3:1; (g) urea, CH₃CO₂H; (h) POCl₃; (i) Et₄NCN, DMSO.

Table 1. C9 Phenyl Substituted Analogues

IC₅₀(nM)^a

compd

R₆

Me

R₇

R₉

PDE10A

PDE2

microsomal stability, h/r, t_1/2(min)^b

1

Me

CF₃

Me

CH₂CH₂Me

2-Me-Ph

7.3 ( 0.6

7.4 ( 0.6

24 ( 3.2

270 ( 17.6

47 ( 7.5

3/1

46

47

48

49

50

51

52

53

54

55

56

57

58

59

60

61

62

63

Me

CF₃

CHF₂

CN

18/3

2-CF₃-Ph

135 ( 4.6

11.6 ( 0.8

11 ( 1.9

13/4

2-Cl-Ph

1.9 ( 0.2

7.4 ( 0.7

4.3 ( 0.4

20.5 ( 1.3

14.4 ( 0.8

186 ( 14

1.8 ( 0.2

60 ( 17.9

44 ( 2.3

13/3

2,5-Cl,Cl-Ph

2,4-Cl,F-Ph

2-OMe-Ph

2-OCF₃-Ph

2-CONH₂-Ph

3-CONH₂-Ph

2-Me-Ph

12/6

27.5 ( 3.8

64 ( 7.3

15/4

9/4

237 ( 50.6

11/3

>30/>30

>30/13

27/22

>30/17

>30/24

>30/20

>30/3

>30/12

24/8

37 ( 4.6

19.5 ( 2.4

13.1 ( 0.8

9.5 ( 0.5

29.9 ( 1.4

11.5 ( 0.5

7.4 ( 0.3

108 ( 12.5

5.4 ( 0.7

35.6 ( 3.5

2-Cl-Ph

2,5-Cl,Cl-Ph

2,4-Cl,F-Ph

2,5-Cl,F-Ph

2,5-Cl,OMe-Ph

2-Cl-Ph

173 ( 20.6

102 ( 6.4

87 ( 6.0

114 ( 11.5

197 ( 32.6

7.3 ( 0.8

6.0 ( 0.4

2-Cl-Ph

16/4

^aMean of IC₅₀with standard error of the mean; n = 4. ^bh = human. r = rat.

of synthetic libraries. Palladium-catalyzed cross-coupling reac-

tion of 6 with any number of arylboronic and heteroarylboronic

acids (Suzuki coupling¹⁷) in the presence of Pd(0) or Pd(II)

catalyst under a variety of well-known conditions produced the

desired products 7a (entries 46ꢀ54, 61, and 62 (Table 1) and

64ꢀ80 (Table 2)). The cyano analogue 63 (Table 1) was pre-

pared through an alternative synthetic route b. Imidazole 3b was

treated with urea to produce pyrazone 8, which upon treatment

with phosphorus oxychloride furnished choloropyrazine 9. Dis-

placement of the chlorine of pyrazine 9 with tetraethylammo-

nium cyanide in dimethylsulfoxide aﬀorded cyanopyrazine 10.

Compound 10 was converted to the desired ﬁnal products 7b as

before. The triﬂuoromethyl-substituted analogues (55ꢀ60,

Table 1) were prepared according to synthetic Scheme 2. Ethyl

2-chloro-4,4,4-triﬂuoroacetate 11 was treated with amidine to

produce imidazole 12. Coupling of 12 with 2-chloropyridine 13

7623

dx.doi.org/10.1021/jm2009138 |J. Med. Chem. 2011, 54, 7621–7638

Journal of Medicinal Chemistry

ARTICLE

Table 2. C9 Heteroaryl Substituted Analogues

IC₅₀(nM)^a

compd

R₉

PDE10A

PDE2

microsomal stability, h/r, t_1/2(min)^b

46

64

65

66

67

68

69

70

71

72

73

74

75

76

77

78

79

80

2-Me-Ph

7.4 ( 0.6

2.8 ( 0.1

18.2 ( 1.0

14.2 ( 1.6

12.9 ( 1.1

6.1 ( 0.5

46 ( 3.3

47 ( 7.5

18/3

2-(3-Me-thienyl)

2-(N-Me-pyrrole)

4-pyrazole

62 ( 6.5

6/2

276 ( 20.8

299 ( 33.8

27 ( 2.8

2/1

4/3

3,5-Me,Me-pyrazole

N-Me,3,5-Me,Me-pyrazole

3,5-Me,Me-isoxazole

2,4-Me,Me-thiazole

5-pyrimidine

>30/>30

>30/12

>30/9

27/25

18/19

24/22

>30/25

3/5

24.6 ( 1.9

3.7 ( 0.3

65.3 ( 5.0

13.3 ( 1.4

4.2 ( 0.4

11 ( 1.1

30 ( 5.6

3-pyridine

32 ( 7.1

17 ( 2.4

38 ( 6.4

3-(2 Me-pyridine)

3-(4-Me-pyridine)

3-(6-Me-pyridine)

3-(4-Cl-pyridine)

4-(3-Me-pyridine)

4-(3-Cl-pyridine)

3-(2-MeO-pyridine)

3-(4-MeO-pyridine)

73.4 ( 3.8

7.9 ( 0.7

5.2 ( 0.4

3.6 ( 0.3

13.9 ( 1,2

14.9 ( 2.4

24 ( 2.3

15/6

40.6 ( 6.0

18.4 ( 2.7

116 ( 15.6

21.9 ( 1.7

>30/>30

5/5

>30/26

^aMean of IC₅₀with standard error of the mean; n = 4. ^bh = human. r = rat.

upon treatment with powdered KOH in N,N-dimethylforma-

mide furnished imidazole 14. Reduction of the nitro group of 14

to aniline with sodium bisulfate in acetic acid followed by in situ

intramolecular cyclization of the aniline with the ester function-

ality furnished pyrazone 15. Treatment of 15 with phosphorus

oxychloride gave 2-chloropyrazine 16, which upon treatment

with trimethylaluminum in the presence of tetrakis(triphenyl-

phosphine)palladium catalyst aﬀorded pyrazine 17. The inter-

mediate 17 was converted to the desired products 18 (55ꢀ60,

Table 1) as described in Scheme 1. The monosubstituted phenyl

analogues (Table 3, entries 81ꢀ92) were prepared from com-

mercially available and appropriately substituted 2-ﬂuoronitro-

benzene in a similar manner as described above. The synthetic

approaches depicted in Schemes 3 and 4 were utilized for the

preparation of the highly functionalized phenylpyrazine deriva-

tives contained in Table 3 (entries 93ꢀ122). In Scheme 3, we

have applied the same synthetic strategy as before by creating

building blocks 32b to increase the diversity and eﬃciency for the

preparation of multiple analogues in this series. Building highly

functionalized nitrophenyl intermediates 23, 24, 25b, 26, and

27 allowed the rapid generation of analogues to address several

key SAR objectives related to ligand’s potency, selectivity,

and metabolic stability. The nitrophenyl analogues 23 and

24 were prepared from the corresponding substituted benzenes

19 and 20 upon treatment with nitric acid in dichloromethane.

For the preparation of nitrophenyl 25b, ﬁrst, aniline 21a was

chlorinated at the ortho position with N-chlorosuccinimide and

then was oxidized with sodium perborate. Application of the

same oxidation protocol on aniline 21b yielded compound 26.

The diﬂuoromethoxy intermediate 27 was prepared from nitro-

phenol 22 upon alkylation with sodium chlorodiﬂuoroacetate.

During the execution of this work we were able to introduce the

2-bromo-4-methylimidazole at an earlier step and decrease the

synthetic steps. In some instances, depending on the substitution

of the nitrophenyl intermediate, the more reactive 4-methylimi-

dazole was used in the displacement reaction as described in

Scheme 1. Electron withdrawing groups on the nitrophenyl

intermediate (i.e., 25b) caused a slow coupling reaction and

furnished low yields. Additional functionalization of the phenyl

moiety of the imidazole adducts 28 and 28a was achieved upon

treatment with potassium methoxide in methanol to generate the

methoxy analogues 29 and 30. Imidazoles 28ꢀ30 were con-

verted to the desired ﬁnal products 33 as described before in

Scheme 1. Modiﬁcations of the alkoxy groups at positions C6

and C8 were accomplished as shown in Scheme 4. Methoxy

analogues 34 and 37 were treated with boron tribromide in

dichloroethane to produce phenols 35 and 38a. An alternative

routewas usedfor the preparation of phenols38b. Nitrophenol 40

was ﬁrst alkylated with benzyl bromide in the presence of

potassium carbonate to give compound 41 and then was con-

verted to intermediates 42a, as described earlier. Hydrogenation

of intermediates 42a with ammonium formate in the presence of

Pd/C produced phenols 38b. Phenols 35 and 38a,b were

alkylated with alkyl halides in the presence of cesium carbonate

to aﬀord the desired ﬁnal products 36 (entries 101ꢀ104,

Table 3) and 39 (entries 119ꢀ121, Table 3). Morpholine ana-

logues 42b (entries 109, 110; Table 3) were prepared from

triﬂuoronitrobenzene 43 upon treatment with morpholine to

7624

dx.doi.org/10.1021/jm2009138 |J. Med. Chem. 2011, 54, 7621–7638

Journal of Medicinal Chemistry

ARTICLE

Scheme 2^a

^aReagents: (a) H₂O; (b) KOH, DMF; (c) CH₃CO₂H, NaHSO₃; (d) POCl₃; (e) Me₃Al, Pd(PPh₃)₄; dioxane; (f) N-bromosuccinimide, CH₃CN;

(g) R₉B(OH)₂, K₂CO₃, Pd(PPh₃)₄, dioxane/H₂O, 3:1.

produce 44, which was further converted to the ﬁnal products as

before.

exposure for in vivo evaluation. The phenyl analogues of Table 1

with alkyl, alkoxy, and halogen substituents had desirable TPSA

values (∼50ꢀ60 Å²) and demonstrated excellent brain permea-

tion (>90%); however, their weak metabolic stability and short

terminal half-life (about 0.5 h) excluded them from further

in vivo evaluation. Analogues 46 and 49 were evaluated in the

Ames test (TA98, TA100, TA1537 strains) and were found to be

negative in the assay. Our initial hypothesis to introduce bulky

groups at position C9 of the fused-ring system to alter the

planarity of the parent molecule has resulted in the elimination of

the mutagenicity potential of this class of compounds. Despite

our initial success on breaking the planarity of the core structure

without loss of target aﬃnity and selectivity, our second objective

to improve the microsomal stability of the molecule has remained

unresolved. As we discussed above, the C7 position methyl group

of 1 was identiﬁed as a key site of metabolic instability (Figure 3).

We have masked this position with a triﬂuoromethyl moiety,

which is resistant to metabolic oxidation, in an attempt to

improve the compound’s metabolic stability. All prepared triﬂu-

oromethyl analogues 55ꢀ60 were metabolically stable in both

human and rodent microsomes; however, their potency for

PDE10A was markedly reduced by about 10- to 20-fold com-

pared to the C7 position methyl substituted analogues 46ꢀ50.

Considering that the methyl and triﬂuoromethyl groups are

comparable in size, unfavorable electrostatic repulsions between

the triﬂuoromethyl group of the ligand and the closely positioned

residue Ser667 of the binding site are the likely cause of potency

loss. Next, we investigated the C6 position of 1, where the methyl

group is prone to metabolic oxidation (Figure 3). Replacement of

the C6 methyl group of 1 with a triﬂuoromethyl group (61)

resulted in about 30-fold loss of potency, while the diﬂuoro-

methyl moiety (62), though comparable in size to triﬂuoromethyl

group, exhibited potency similar to that of parent compound 1.

Both of the ﬂuorinated analogues exhibited better microsomal

stability in human microsomes than 1. The cyano analogue 63

was also equipotent to 1. Furthermore, replacement of the

metabolically labile 2-OMe moiety of 1 (Figure 3) with either

a triﬂuoromethoxy or diﬂuoromethoxy group, both of which are

resistant to metabolic oxidation, has resulted in retention of the

PDE10A potency but without any improvement of the micro-

somal stability of the compound (data not shown). A multitude

of C9 position substituted phenyl analogues (data not shown)

with any number of substituents (alkyl, alkoxy, halogens, etc.)

and various substitution patterns (ortho, meta, para, and com-

binations) were prepared and tested against all PDEs. The

’ RESULTS AND DISCUSSION

Inhibitory potencies of our compounds were tested using in

vitro inhibition of human recombinant PDE10A catalyzed cAMP

hydrolysis. The results are compiled in Tables 1ꢀ3.

Our SAR studies encompass three key objectives. First, alter

the planarity of the parent compound 1 to eliminate possible

intercalation with DNA. Second, improve the metabolic stability

and pharmacokinetic properties to produce drugable candidates,

and third, further optimize ligand potency, selectivity, and in vivo

eﬃcacy. Examination of the X-ray crystal structure of 1:PDE10A

revealed a large unoccupied region of the ligand binding pocket

(Figure 2), allowing the introduction of bulky groups at the C9

position of the parent compound to break the planarity of the

tricyclic fused-ring system of 1. Molecular modeling calculations

also suggested that introduction of ortho-substituted aromatic

moieties at the C9 position would produce maximum torque

rotation between the aromatic groups and the fused-ring system.

Such structural arrangement will alter the three-dimensional

planarity of the parent structure 1, thus preventing its align-

ment between the base pairs of DNA. In Tables 1ꢀ3 we have

aggregated selected prepared analogues that support our SAR

objectives. The o-tolyl analogue 46 showed potency similar to

that of 1 with a modest improvement of the microsomal stability

in human microsomes. The triﬂuoromethyl 47 and chloro 48

analogues, carrying more stable groups on the phenyl moiety,

showed metabolic stability properties similar to those of 46;

nonetheless, 47 was 3-fold weaker in potency than 46, while 48

was 3-fold more potent than 46. The dihalogenated analogues 49

and 50 also had a metabolic stability proﬁle similar to that of 46.

The less bulky monochloro analogue 48 was about 4-fold more

potent than the dichloro analogue 49. The alkoxy analogues 51

and 52 demonstrated an 2- to 3-fold decrease in potency (51 and

52 vs 1). Carboxamides 53 and 54 were the most metabolically

stable analogues in both human and rodent microsomes; how-

ever, they displayed diﬀerent inhibitory potency for PDE10A.

The ortho analogue 53 was about 90-fold weaker than the para-

substituted analogue 54. Compound 54 exhibited a desirable in

vitro proﬁle for further in vivo evaluation; however, its high

topological polar surface area (TPSA, 93 Å²) property contrib-

uted by the polar carboxamide group resulted in poor brain

permeability (<20%) with concomitant limited central drug

7625

dx.doi.org/10.1021/jm2009138 |J. Med. Chem. 2011, 54, 7621–7638

Journal of Medicinal Chemistry

ARTICLE

Table 3. Phenylpyrazine Analogues

IC₅₀(nM)^a

compd

R₁

R₆

R₈

PDE10A

PDE2

microsomal stability, h/r, t_1/2(min)^b

81

3-(4-Me-pyridine)

4-(3-Me-pyridine)

2-Me-Ph

H

113 ( 7.7

158 ( 6.4

49.2 ( 2.6

14.4 ( 0.3

28.9 ( 4.3

11 ( 1.7

>30/26

>30/>30

>30/14

25/3

82

H

F

83

H

F

84

H

F

75 ( 5.8

40 ( 2.3

14 ( 1.7

89 ( 1.8

85

3-(4-Me-pyridine)

4-(3-Me-pyridine)

3-(4-Me-pyridine)

4-(3-Me-pyridine)

3-(4-Me-pyridine)

4-(3-Me-pyridine)

3-(2 Me-pyridine)

3-(4-Me-pyridine)

4-(3-Me-pyridine)

3-(4-Me-pyridine)

4-(3-Me-pyridine)

3-(2 Me-pyridine)

2,4-Me,Me-thiazole

N-Me,3,5-Me,Me-pyrazole

2-Me-Ph

H

Cl

>30/24

27/14

86

H

Cl

87

H

OMe

OCF₃

CF3

CF₃

F

24.2 ( 1.1

46 ( 5.9

25/10

88

H

23/22

89

H

8.3 ( 0.7

35.7 ( 3.5

7.4 ( 0.6

6.8 ( 0.6

79.5 ( 1.4

22 ( 1.5

23 ( 4.5

>30/12

>30/17

24/15

90

H

91

H

13.6 ( 2.0

14.7 ( 2.2

92

H

>30/28

>30/>30

>30/26

>30/>30

>30/26

18/5

93

F

94

F

153 ( 2.2

305 ( 41

108 ( 17

92 ( 16.7

621 ( 78.4

194 ( 22.3

66 ( 2.4

95

OMe

OCHF₂

F

2.0 ( 0.3

0.7 ( 0.03

0.55 ( 0.07

1.5 ( 0.1

0.85 ( 0.12

0.35 ( 0.03

1.9 ( 1.3

7.5 ( 0.9

11.4 ( 1.5

102 ( 4.8

63.9 ( 7.0

0.5 ( 0.05

0.2 ( 0.02

0.14 ( 0.01

1.7 ( 0.2

0.1 ( 0.01

0.05 ( 0.01

0.06 ( 0.01

0.09 ( 0.01

0.8 ( 0.1

0.7 ( 0.03

11.7 ( 1.8

3.9 ( 0.3

1.5 ( 0.2

3.5 ( 0.1

5.5 ( 0.5

5.7 ( 0.3

2.9 ( 0.3

96

F

97

F

98

F

99

F

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

F

2-Me-Ph

F

138 ( 24.6

300 ( 66.9

619 ( 65

2-Me-Ph

OCH₂CH₃

F

15/5

2-Me-Ph

OCH₂CF₃

F

>30/8

2-Me-Ph

OCH₂-cyclopropane

F

12/6

2-Me-Ph

OH

OMe

Cl

F

6/6

3-(4-Me-pyridine)

4-(3-Me-pyridine)

3-(2 Me-pyridine)

4-(3-Me-pyridine)

3-(2 Me-pyridine)

4-(3-Me-pyridine)

3-(4-Me-pyridine)

4-(3-Me-pyridine)

3-(2 Me-pyridine)

4-(3-Me-pyridine)

N-Me,3,5-Me,Me-pyrazole

OMe

165 ( 18.8

83 ( 10.5

30 ( 3.9

501 ( 50

42 ( 4.9

>30/16

25/13

>30/27

>30/26

>30/>30

3/3

morpholine

OCH₂Ph

OCHF₂

CF3

71 ( 7.8

>30/>30

>30/14

>30/17

>30/>30

17/15

112 ( 9.1

411 ( 22.6

325 ( 21

70 ( 6.5

OMe

Cl

CF3

F

OMe

F

OMe

70.5 ( 10.6

21.6 ( 2.1

36 ( 1.2

F

OMe

>30/28

>30/20

>30/>30

18/9

F

OCHF₂

OCH₂CF₃

OCH₂-cyclopropane

OMe

F

128 ( 4.6

81 ( 2.1

F

22 ( 2.3

>30/>30

^aMean of IC₅₀with standard error of the mean; n = 4. ^bh = human. r = rat.

potency, selectivity, and metabolic stability of these new ana-

logues were similar to those of the above-discussed ortho

analogues. We have routinely evaluated the selectivity proﬁle of

all prepared compounds against all PDEs isoforms. All tested

7626

dx.doi.org/10.1021/jm2009138 |J. Med. Chem. 2011, 54, 7621–7638

Journal of Medicinal Chemistry

ARTICLE

Scheme 3^a

^aReagents: (a) HNO₃, CH₂Cl₂; (b) N-chlorosuccinimide, CH₃CN; (c) NaBO₃.H₂O, CH₃CO₂H; (d) PhCOCClF₂, KOH or ClF₂CO₂Na;

(e) 2-bromo-4-methylimidazole or 4-methylimidazole, K₂CO₃, DMF; (f) KOH, MeOH; (g) Pd/C, HCO₂NH₄, MeOH; (h) AC₂O, toluene;

(i) P₂O₅, POCl₃; (j) N-bromosuccinimide, CH₃CN; (k) R₁B(OH)₂, K₂CO₃, Pd(PPh₃)₄, dioxane/H₂O 3:1.

compounds have demonstrated high selectivity (>500ꢀ1000ꢁ;

data not shown) for all PDEs isoforms with the exception of the

PDE2 isoform. For matter of brevity, we are presenting our SAR

evolution only with respect to PDE10A and PDE2 activity.

However, if unexpected ﬁndings emerged during the screening

against the other PDEs, we present these data appropriately.

Next, we turned our attention to the evaluation of heteroaryl

groups at position C9. The thienyl analogue 64 was 3-fold more

potent than 46, while the N-Me pyrrole 65 and pyrazole 66 were

2-fold weaker in potency than 46. Both of these new analogues

were metabolically unstable. In contrast, the methyl-substituted

pyrazoles 67 and 68 exhibited high resistance to metabolic

oxidation and were stable in both human and rodent microsomes

(67, 68 vs 66). Steric interactions between the bulkier ligands,

where the C9-group projects oﬀ the plane of the fused-ring core

scaﬀold, and the cytochrorme P450 degrading proteins are likely

the cause of the decreased metabolic activity, with concomitant

increase in compound metabolic stability. Similar to the sub-

stituted pyrazoles, both substituted isoxazole 69 and thiazole 70

also exhibited enhanced metabolic stability in human micro-

somes. Pyrimidine 71 showed weak PDE10A potency and

moderate stability (71 vs 46), while pyridine 72 was 5-fold more

potent than 71. Introduction of a methyl moiety at the ortho

position on the C9 pyridine nucleus to increase its size and maxi-

mize rotation in relation to the fused-ring system resulted in a

3-fold increase in potency (73 vs 72). The para-methyl sub-

stituted pyridine 75 was 18-fold less potent and metabolically

unstable (75 vs 73). The para-methyl group most likely generates

a steric conﬂict with residues at the wall of the binding pocket,

which contributes to the marked loss in ligand aﬃnity. The

chloro analogue 76 has retained the PDE10A potency, but it was

metabolically less stable (76 vs 74). The para-pyridine 77 was

2-fold more potent that the meta-pyridine 74. Introduction of an

ortho chlorine substituent (entry 78) resulted in potency reten-

tion but with pronounced microsomal instability (78 vs 77). One

can hypothesize that chlorine’s electron-withdrawing eﬀect re-

duces the electronic richness of the pyridine nucleus (entries 76

and 78), which likely aﬀects the binding aﬃnity between the

ligand and cytochrorme P450 oxidizing proteins, thus inﬂuen-

cing their microsomal stability. The methoxy-substituted pyr-

idines 79 and 80 exhibited potency and stability similar to those

of the methyl analogues 73 and 74. Compounds 68 and 74 were

selected for evaluation in the Ames test (TA98, TA100, TA1537

strains) and were found to be negative in the assay. All prepared

heteroaryl analogues had desirable TPSA values (∼55ꢀ65 Å²)

and demonstrated excellent brain permeability (>90%). In order

to unambiguously conﬁrm the orientation of the C9 aromatic

moiety, we have crystallized compound 74 with human PDE10A

protein. Examination of the 74:PDE10A crystal structure (Figure 4)

revealed that the C9 pyridine projects toward the opened space

of the PDE10A binding pocket, in agreement with our molecular

modeling calculations, and it is positioned practically in a vertical

orientation relative to the fused-ring system. This orientation

breaks the overall three-dimensional planarity of the parent

compound 1 and restricts basically the insertion of the ligand

between the base pairs of DNA. Furthermore, the fused ring

7627

dx.doi.org/10.1021/jm2009138 |J. Med. Chem. 2011, 54, 7621–7638

Journal of Medicinal Chemistry

ARTICLE

Scheme 4^a

^aReagents: (a) BBr₃, ClCH₂CH₂Cl; (b) Cs₂CO₃, alkyl halides; (c) PhCH₂Br or MeI, K₂CO₃; (d) HCO₂NH4, Pd/C; (e) morpholine.

binding pocket with additional water-mediated hydrogen bond-

ing interactions between the imidazole nitrogen and the residues

Tyr514 and Asp664. Compounds 68 and 74 have displayed

desirable in vitro and ADME proﬁles and were selected for

further in vivo evaluation. Both compounds were eﬃcacious in

well-established predictive animal models for the positive symp-

toms of schizophrenia, namely, the reversal of MK-801 induced

locomotor activity (LMA, Table 5) and conditioned-avoidance

responding (CAR; MED values of 3 mg/kg, po, rat) assays. The

pharmacokinetic proﬁles of 68 and 74 in rats (Table 4) and dogs

have shown moderate clearance, low volume of distribution,

good bioavailability in rats (68, 67%; 74, 45% po 2.5 mg/kg) and

dogs (74, 100% po 2 mg/kg), and rather short terminal half-lives

in rats (0.5 h) and dogs (1 h). Despite the robust metabolic

stability of 68 and 74 in both rodent and human microsomes,

in vivo metabolic studies have revealed that the 2-OMe pyridine

group was extensively demethylated after oral administration to

form a 2-pyridone moiety. The in vivo extensive metabolism of

68 and 74 most likely has contributed to the rapid elimination of

the compounds and short terminal half-life, especially in rodents.

Armed with this information, we developed a chemistry strategy

to replace the 2-OMe pyridine moiety of 74 with a substituted

phenyl moiety in an attempt to reduce metabolism and improve

the pharmacokinetic properties of the compound. In parallel, we

envisioned expanding the SAR studies of 74 and improving its

potency and selectivity for PDE10A. Examination of the 74:

PDE10A X-ray crystal structure (Figure 4) revealed that addition

of an electron rich group at C4 position of 74 oﬀered the

potential for additional interactions between the ligand and the

invariant residue Glu716. To delineate these observations, we

Figure 4. Crystal structure (1.9 Å resolution) of PDE10A complexed

with 74, highlighting the key hydrogen-bonding interactions between

ligand and protein at the invariant Glu716, and the water-mediated

contacts with Asp664 and Tyr514. The fused ring system of the ligand is

sandwiched between residues Phe719, Ile682, and Phe686, making π-

stacking interactions with Phe719, π-edge stacking with Phe686, and

hydrophobic contact with Ile682 (not shown). The C9 position

methylpyridine rotates oﬀ the fused tricyclic ring system at about 90ꢀ.

system of the ligand is sandwiched between residues Phe719,

Ile682, and Phe686, making π-stacking interactions with Phe719

and hydrophobic contacts with residues Ile682 and Phe686. Also,

the ligand extends toward the invariant Glu716 residue and

makes a hydrogen-bonding interaction between the pyrazine

nitrogen and Glu716. The ligand is further stabilized within the

7628

dx.doi.org/10.1021/jm2009138 |J. Med. Chem. 2011, 54, 7621–7638

Journal of Medicinal Chemistry

ARTICLE

Table 4. Selectivity of Representative Compounds toward Human PDEs Isoforms

IC₅₀(nM)^a

compd

PDE1

PDE2

PDE3

PDE4

PDE5

PDE6^b

PDE7

3345

PDE8

PDE9

PDE10

PDE11

1

>5000

>1000

>10000

>5000

>10000

>5000

>10000

>500

270 ( 17.6

11 ( 1.9

>5000

>1000

>5000

>10000

6900

>5000

>1000

>10000

6350

920

>1000

2080

>1000

1870

>5000

>1000

>5000

>1000

7.3 ( 0.6

7.4 ( 0.7

4.3 ( 0.4

1.8 ( 0.2

7.5 ( 0.7

2.8 ( 0.1

6.1 ( 0.5

3.7 ( 0.3

11 ( 1.1

690

>1000

2810

49

332

1300

50

27.5 ( 3.8

37 ( 4.6

>1000

54

2930

>10000

>5000

>10000

>1000

46

47 ( 7.5

1240

3100

542

2400

64

62 ( 6.5

1440

1780

266

>10000

>5000

1510

68

24.6 ( 1.9

30 ( 5.6

>10000

5280

>10000

2590

>10000

3110

>10000

1330

1170

>10000

5090

70

824

74

38 ( 6.4

2680

>10000

>5000

>10000

>5000

>10000

2650

1850

1630

2318

>10000

5290

89

23 ( 4.5

>5000

>10000

>500

>10000

>1000

>5000

5165

>10000

>5000

13700

3570

919

8.3 ( 0.7

6.8 ( 0.6

0.7 ( 0.03

0.55 ( 0.07

0.5 ( 0.05

0.2 ( 0.02

0.14 ( 0.01

0.09 ( 0.01

0.8 ( 0.1

0.7 ( 0.03

91

14.7 ( 2.2

108 ( 17

92 ( 16.7

165 ( 18.8

83 ( 10.5

30 ( 3.9

1340

>5000

>10000

2200

96

>100000

>10000

>500

>10000

>500

>10000

>500

97

106

107

108

113

114

115

8780

4870

>5000

2050

3440

3350

2300

2350

112 ( 9.1

411 ( 22.6

325 ( 21

>500

>1000

^aMean of IC₅₀with standard error of the mean; n = 4. IC₅₀values are related to the maximum concentration used in the screen. ^bBovine isoform.

withdrawing and electron donating groups (entries 85ꢀ94) at

position C8 resulted in about 5- to 10-fold enhancement of the

PDE10A potency, which was comparable to that of the pyridineꢀ

pyrazine analogues as described above. The phenylpyrazine ana-

logues demonstrated a 3- to 5-fold potency diﬀerence between

para- and meta-substituted C1 pyridines (88 vs 89, 90 vs 91, 93

vs 94). The increased potency of the C1 substituted para-pyridines

is due to the additional water-mediated hydrogen-bonding inter-

action between the pyridine nitrogen and residues Leu623 and

Thr623 (supportive X-ray crystallography ﬁndings are shown

below in Figure 5). Next, we turned our attention to the C6

position of 74, where molecular modeling calculations and X-ray

observations with the 74:PDE10A crystal structure have sug-

gested that electron donating groups at the C6 position oﬀered

the potential of new contacts between the ligand and residue

Glu716. Introduction of a methoxy moiety at position C6 (entry

Figure 5. Crystal structure (1.8 Å resolution) of PDE10A complexed

95) resulted in about 80-fold improvement of the PDE10A

with 96, highlighting the key hydrogen-bonding interactions between

potency (95 vs 82). In addition, the para- and ortho-pyridines 96

ligand and protein at the invariant Glu716 with both the pyrazine

and 97 exhibited exceptional potency for PDE10A, representing

nitrogen and the 6-OMe group, and the two water-mediated contacts

our ﬁrst compounds with subnanomolar aﬃnity for PDE10A.

with Asp664, Tyr514, Thr623, and Leu625. The additional contacts

The thiazole and pyrazole analogues 98 and 99 were also very

potent PDE10A inhibitors. To conﬁrm our molecular modeling

calculations, compound 96 was crystallized with human PDE10A.

Examination of the 96:PDE10A X-ray crystal structure (Figure 5)

revealed that the ligand orients into the binding pocket similarly

to that of compound 74, with two additional hydrogen-bonding

interactions: (i) those between the C6 methoxy group of 96 and

the invariant Glu716; (ii) water-mediated contacts between the

pyridine nitrogen of the ligand and the residues Thr623 and

Leu625 of the binding pocket. All other observed contacts

between the ligand and residues of the ligand binding pocket

were similar to the contacts observed in the 74:PDE10A X-ray

crystal structure. Next, we explored the size and electronic

between the 6-OMe and Glu716 and the water-mediated interactions

with Thr623 and Leu625 contribute to the ligand's increased aﬃnity for

PDE10A.

have designed an alternative scaﬀold (Table 3) by replacing the

pyridine nucleus with a phenyl group while maintaining at the C1

position of the new scaﬀold the same pyridine groups we have

optimized above for PDE10A potency and microsomal stability.

The initial phenyl analogues 81ꢀ83 were weak against PDE10A,

with the C1 position substituted para-pyridine 83 being about

3-fold more potent than the analogous meta-pyridine 82. The

toluene analogue 84 was about 10-fold better than 82, but

metabolically it was unstable. Employing a variety of electron

7629

dx.doi.org/10.1021/jm2009138 |J. Med. Chem. 2011, 54, 7621–7638

Journal of Medicinal Chemistry

ARTICLE

methoxy group with ﬂuorine and chlorine substituents while

maintaining at the C8 position the above optimized substituents

(114ꢀ122). All analogues were about 3- to 20-fold less potent

than the analogous C6-methoxy analogues (114 vs 107, 117 vs

97). The unexpected high potency of the C6-chloro analogue

115 (115 vs 89), considering that the addition of an electron

donating methoxy group at C6 was a prerequisite to the ligand’s

high aﬃnity, as discussed above (i.e 113 vs 89), prompted us to in-

vestigate this ﬁnding by crystallizing compound 115 with human

PDE10A. Examination of the 115:PDE10A X-ray crystal struc-

ture (Figure 6) revealed that the ligand orients in the binding

pocket similarly to that of compound 96; however, the presence

of the C6-chloro substituent has caused a slight shift of the ligand

within the binding pocket, bringing it closer to residue Tyr514.

This ligand realignment within the binding pocket has triggered

the displacement of the conserved water, present in the previous

96:PDE10A X-ray crystal structure, promoting a direct contact

between the ligand and residue Tyr514. The gained binding

energy of the direct hydrogen-bond interaction between the

ligand and residues Tyr514 is likely the cause of the increased

ligand aﬃnity. The direct contact between ligand and protein

favorably compensates for the loss in desolvation energy neces-

sary to displace the conserved water. Most of the other contacts

between the ligand and PDE10A protein were similar to the

above-discussed X-ray crystal structures.

Figure 6. Ligplot²³representation (2.4 Å resolution) of PDE10A

complexed with 115, highlighting the hydrophobic and hydrogen bond

interactions between the ligand and the amino acid residues involved in

the PDE10A binding pocket. The ligand orients into the binding pocket

similarly to that of compound 96 (Figure 5), making several hydro-

phobic contacts; however, the presence of the C6-chloro substituent has

caused a slight shift of the ligand within the binding pocket, causing

direct interaction with residue Tyr514.

All tested compounds have demonstrated high selectivity for

all PDEs with the exception of PDE2, where a lower level of

selectivity was observed. Nevertheless, the C6,C8-disubstituted

phenyl analogues (i.e., entries 106, 112ꢀ115), which have

exhibited subnanomolar aﬃnity for PDE10A, have demonstrated

high selectivity (>500ꢁ) against PDE2. Table 4 outlines a collec-

tion of representative analogues that were evaluated against the

PDE panel of human isoforms, with the exception of PDE6 which

is the bovine isoform.

The selectivity of key analogues 68, 96, and 106 was assessed

in Nova Screen assays (neurotransmitters, ion channels, growth

factors/hormones, kinases, brain/gut peptides, immunological

factors, and prostaglandins). All compounds were devoid of any

activity at 10 μM in all tested assays, with the exception of 96

which showed activity against adenosine A1 (IC₅₀= 150 nM)

and A2A (IC₅₀= 5.05 μM).

Most of the analogues shown in Table 3 have demonstrated

good microsomal stability in both human and rodent micro-

somes, have desirable TPSA values (∼50ꢀ65 Å²), and exhibit

excellent brain permeability (>90%). The robust brain perme-

ability of this class of compounds was indicative of the molecule’s

low molecular weight, low topological surface area, and lack of

P-glycoprotein (P-gp) mediated eﬄux property. Evaluation of

the PDE10 inhibitors in the MDR1-MDCK permeability assay

revealed that the compounds, while showing a high level of

permeability [apical to basolateral (AꢀB) P_appand basolateral to

apical (BꢀA) P_app], lacked eﬄux properties (data not shown).

The pharmacokinetic properties of several compounds in rats

and dogs after oral administration are summarized in Table 5. All

compounds showed low to moderate clearance, low volume of

distribution, good bioavailability in rats and dogs, low to good

terminal half-lives (especially for the later phenylpyrazines), and

good brain exposure.

requirements of C6 substitution toward potency. The diﬂuor-

omethoxy analogue 101 was 6-fold weaker than the methoxy

analogue 100. Since the diﬂuoromethoxy moiety is practically

equal in size to that of the methoxy group, the increased electro-

negativity of the diﬂuoromethoxy group is likely the cause of

potency loss. One can hypothesize that either the reduced electro-

nic environment of the oxygen atom attached to the diﬂuor-

omethyl electronegative group weakens the hydrogen-bond

interaction between the diﬂuoromethoxy group and the NH

functionality of the Glu617 residue or the electronegative diﬂuoro-

methoxy group creates an electronic repulsive environment with

Glu716. Increasing the size of the C6 group has aﬀected the

ligand’s potency. Ethoxy analogue 102 was 20-fold less potent

than 100, while the more electronegative triﬂuoroethoxy ana-

logue 103 was even less potent (40-fold, 103 vs 100). The bulkier

cyclopropane analogue 104 was markedly 400-fold less potent

(104 vs 100). Bulky groups are likely restricting the hydrogen-

bond contact between the ligand’s pyrazine nitrogen and Glu716

and causing the realignment of the ligand within the binding

pocket, which results in marked loss of the ligand’s aﬃnity.

Demethylation of the C6 methoxy group (entry 105) resulted in

200-fold loss of potency. To further expand the breadth of our

SAR studies, we explored C8 position substitutions (106ꢀ112).

Introduction of a methoxy moiety at C8 resulted in about 3-fold

improvement of the PDE10A potency (108 vs 97). The bulkier

morpholine 110 and benzyl analogues 111 were also very potent

(110, 111 vs 97). Even though the C8 methoxy analogues showed

good metabolic stability, in vivo metabolic studies indicated that

the C8 methoxy group of 106 was prone to demethylation,

resulting in a short terminal half-life (rat t_1/2≈1.5 h) after oral

administration. Replacing the C8 methoxy group with the more

stable diﬂuoromethoxy and triﬂuoromethoxy moieties (entry

112, 113) resulted in (i) about 2-fold increase in potency (112 vs

108), (ii) enhancement of the microsomal stability, and (iii)

longer elimination half-life in rats (112 t_1/2≈ 3.5 h and 113

t_1/2≈ 11.7 h, respectively). Finally, we have replaced the C6

Effect of PDE10A Inhibition on Striatal Cyclic Nucleotide

cAMP and cGMP Levels in Vivo. Given the dual nature of

PDE10A phosphodiesterase activity, we have examined the effect

of PDE10A inhibitor 96 in vivo by assessing striatal cAMP and

7630

dx.doi.org/10.1021/jm2009138 |J. Med. Chem. 2011, 54, 7621–7638

Journal of Medicinal Chemistry

ARTICLE

Table 5. Pharmacokinetic Proﬁle of Selected Compounds

68

rat

74

rat

96

106

112

rat

dog

rat

dog

rat

dog

dose (mg/kg)

F (%)

2.0

67

2.5

27

2.5

73

1.8

38

6

2.5

25

2

0.1

83

3.5

7

0.5

84

3.0

6

25

0.5

31

1

100

3.4

8

t_1/2(h)

CL (mL min^ꢀ1kg^ꢀ1

0.5

10

1.5

14

1.3

14

)

V_ss(L/kg)

0.7

0.9

1.3

1.6

nt

0.6

1.1

nt

AUC_brain/AUC_plasma,^a10 mg/kg po

2941/4128

866/747

2372/3638

nt

3279/8852

4400/12738

^aAUC in h ng/mL.

3

preclinical model predictive of antipsychotic activity. In the rat

CAR model (Figure 8), 96 decreased avoidance responding

with a significant treatment effect at 1 mg/kg. Data are expressed

as mean (with SEM) avoidance, escape, and response failures

observed over 50 trials (n = 8 animals/group). Nonlinear re-

gression analysis calculated an ID₅₀equivalent to 1.19 (

0.29 mg/kg; p = 0.05.

In summary, we have described the evolution of new phenyl-

pyrazine PDE10A inhibitors based on our previously reported

pyridyl-pyrazine 1. Applications of structure-based drug design

techniques and molecular modeling have led to the identiﬁcation

of subnanomolar PDE10A inhibitors with excellent selectivity

against the other members of the PDE family. Exploring the C9

position of 1, we were able to alter the planarity of the parent

compound and eliminate a potential risk of mutagenicity. Also,

the identiﬁcation of a second hydrogen-bond interaction be-

tween the ligand and the invariant residue Glu716 has led to a

marked increase of the ligand’s aﬃnity. A systematic assessment

of the ADME and pharmacokinetic properties of the newly

synthesized compounds has led to the design of several druglike

candidates with high brain permeability and desirable drug kinetics

(t_1/2, bioavailability, clearance). Direct correlation of PDE10A in-

hibition and in vivo functional activity, as assessed by cyclic

nucleotides level changes, was demonstrated with 96. Both striatal

cAMP and cGMP levels were increased by 1.9ꢁ and 4.9ꢁ,

respectively, at a 3 mg/kg oral dose. Compound 96 was highly

potent for PDE10A (IC₅₀= 0.7 nM), demonstrated high

selectivity (>3000ꢁ) for the other PDEs, with the exception of

PDE2 (150ꢁ selective), and was eﬃcacious in animal models of

psychoses; reversal of MK-801 induced hyperactivity and con-

ditioned avoidance responding (CAR).

Figure 7. Dose-dependent eﬀects of 96 upon mouse striatal cyclic

nucleotide levels. Compound 96 was administered 10 or 30 min before

sacriﬁce (n = 4): vehicle-treated; /, p = 0.001; //, p < 0.0005.

cGMP tissue levels after dosing with the compound. Compound

96 showed excellent potency (PDE10A IC₅₀= 0.7 nM) and

selectivity (>3000ꢁ against PDE1, PDE3ꢀ9, PDE11; >150ꢁ

against PDE2). The effect of dosing compound 96 (0.3 and

3.0 mg/kg, ip, 30 min) in CF-1 mouse on striatal cyclic nucleo-

tides was assessed. Dose-dependent increases in both cAMP and

cGMP levels were observed with statistically significant changes

of 1.9- and 4.9-fold in the levels of cAMP and cGMP, respectively,

over the vehicle-treated animals with the 3 mg/kg dose (Figure 7).

A study of enzyme occupancy with 96 (unpublished data)

indicated about approximately 80% occupancy was required to

achieve these cyclic nucleotide elevations.

Effect of PDE10A Inhibitors in Models Predictive of Anti-

psychotic ActivityReversal of MK-801 Induced Hyperactivity

in Rats. A representative subset of high affinity PDE10A inhibi-

tors was selected for investigation of their antipsychotic-like

potential in vivo. To enable the characterization of the anti-

psychotic activity, we utilized a MK-801 induced hyperactivity

reversal model. MK-801 (0.1 mg/kg ip) significantly increased

horizontal activity (t test, p < 0.001) and induced stereotyped

sniffing (t test, p < 0.001) in female rats. These behaviors were

significantly reversed by the positive control, which in this case

was the atypical antipsychotic clozapine. All evaluated PDE10A

inhibitors also reversed hyperactivity and stereotyped sniffing in

a dose dependent and significant manner. The minimal effective

dose (MED) of the novel PDE10A inhibitors can be seen in

Table 6. There is a clear qualitative correlation between in vitro

potency at PDE10A and MED values in this assay, with the most

potent phenylpyrazines (96, 106, 108, 112, 113) representing

the most efficacious in vivo analogues.

’ EXPERIMENTAL SECTION

Chemistry. Melting points were determined in open capillary tubes

on a Mel-Temp II apparatus and reported uncorrected. ¹H NMR spectra

were determined in the cited solvent on a Varian Unity or Varian Inova

(300ꢀ400 MHz) instrument with tetramethylsilane as an internal

standard. Chemical shifts are given in ppm, and coupling constants

are in hertz. Splitting patterns are designated as follows: s, singlet; br s,

broad singlet; d, doublet; t, triplet; q, quartet; m, multiplet. Mass spectra

were recorded on a Micromass LCT, Waters, spectrometer. Elemental

analyses (C, H, N) were performed on a Perkin-Elmer 240 analyzer, and

all compounds are within (0.4% of theory unless otherwise indicated.

HPLC techniques and high resolution mass spectrometry were used to

determine the purity of compounds outside the range of the elemental

analysis assessment. Purity of all final products was >96% as determined

by HPLC and/or combustion analysis. Purity was determined by HPLC

Disruption of Conditioned Avoidance Responding (CAR)

in Rats. Disruption of the “avoidance” response in CAR is another

7631

dx.doi.org/10.1021/jm2009138 |J. Med. Chem. 2011, 54, 7621–7638

Journal of Medicinal Chemistry

ARTICLE

Table 6. Reversal of MK-801 Induced Hyperactivity and Stereotyped Sniﬃng

compd

MED^ahyperactivity (mg/kg po)

% decrease in hyperactivity^b

MED^asniﬃng (mg/kg po)

% decrease in sniﬃng^b

IC₅₀(nM)^c

68

0.5

57.4

38.2

62.4

39.3

30.4

34.8

33.2

51.1

16.53

0.1

25.1

83.5

42.6

25.2

43.6

23.7

16.8

31.9

16.28

6.1

11

74

1

2.5

96

0.075

0.05

0.0125

0.025

0.01

0.25

20

0.075

0.05

0.025

0.01

0.25

20

0.7

0.5

0.1

0.06

0.09

0.7

ni^d

106

108

112

113

115

clozapine

^aMinimal eﬀective dose that signiﬁcantly reduced MK-801-induced hyperactivity or stereotyped sniﬃng, respectively (p < 0.05). ^bp < 0.001, n = 6.

^cInhibition of PDE10 activity from Tables 1ꢀ3. ^dni: no inhibition at PDE10.

Figure 8. Compound 96 was evaluated in the conditioned avoidance response in male SpragueꢀDawley rats using a 60 min pretreatment interval and

an oral route of administration. Data are expressed as mean (with SEM): (A) avoidance, escape, and response failures observed in over 50 trials (n = 8

animals/group); /, statistical diﬀerence in avoidance and escape response from vehicle-treated controls (p = 0.05). Nonlinear regression analysis (B)

calculated an ID₅₀equivalent to 1.19 ( 0.29 mg/kg; p = 0.05.

analysis using the following protocols. Method A: mobile phase A, H₂O

(0.1% HCOOH); mobile phase B, CAN (0.1% HCOOH); solvent

gradient 85/15 to 5/95 A/B in 2 min, hold 1.25 min, re-equilibrate

0.5 min; flow rate 1.8 mL/min; Onyx Monolithic C18, 0.3 cm ꢁ 10 cm;

temperature 40 ꢀC; detection at 210ꢀ370 nm. Method B: mobile phase

A, 10 mM ammonium formate in water (pH 3.5); mobile phase B, 50:50

ACN/MeOH; solvent gradient 85/15 to 5/95 A/B in 2 min, hold 1.25

min, re-equilibrate 0.5 min; flow rate 1.1 mL/min; column Agilent SB

C18 1.8 μm, 3.0 cm ꢁ 50 mm; temperature 45 ꢀC; detection at 210ꢀ

370 nm. All products, unless otherwise noted, were purified by “flash

chromatography” with use of 220ꢀ400 mesh silica gel. Thin-layer chro-

matography was done on silica gel 60 F-254 (0.25 mm thickness) plates.

Visualization was accomplished with UV light and/or 10% phospho-

molybdic acid in ethanol. The hydration was determined by the Karl

Fischer titration, using a Mitsubishi moisture meter model CA-05.

Unless otherwise noted, all materials were obtained commercially and

used without further purification. All reactions were carried out under an

atmosphere of dried argon or nitrogen.

Representative Synthetic Protocols of the Aminohydan-

toins Shown in Schemes 1ꢀ4. 8-Methoxy-4-methyl-3-triflu-

oromethyl-2,5,9,9b-tetraaza-cyclopenta[a]naphthalene

(17). Step a. 5-Trifluoromethyl-3H-imidazole-4-carboxylic Acid Ethyl

Ester (12). Ethyl 2-chloro-4,4,4-trifluoroacetate (25 g, 0.114 mol) was

combined with amidine (50 g, 1.1 mol) and water (5 mL). The mixture

became warm and was heated to 130 ꢀC for 1.5 h. The mixture was then

cooled to room temperature, and 100 mL of iceꢀwater was added. The

resulting solids were collected and washed with water, then dried to give

5-trifluoromethyl-3H-imidazole-4-carboxylic acid ethyl ester as a brown

solid (5.5 g, 23% yield). MS m/e 209 [M + H]⁺; ¹H NMR (400 MHz,

DMSO) δ ppm 8.41 (s, 1H), 5.6 (s, 1H), 4.12 (q, J = 7.07 Hz, 2H), 3.93

(s, 3H), 1.01 (t, J = 7.19 Hz, 3H).

Step b. 3-(6-Methoxy-3-nitropyridin-2-yl)-5-trifluoromethyl-3H-imida-

zole-4-carboxylic Acid Ethyl Ester (14). 5-Trifluoromethyl-3H-imida-

zole-4-carboxylic acid ethyl ester (5 g, 24 mmol) and 2-chloro-3-nitro-6-

methoxypyridine (4.5 g, 24 mmol) were dissolved in DMF (60 mL). Into

this mixture was added freshly powdered KOH (1.3 g, 24 mmol). The

mixture was heated to 70 ꢀC for 16 h, then cooled to room temperature

and diluted with water. The solution was then extracted twice with ethyl

acetate. The organic layers were combined and washed with water and

brine and dried over MgSO₄. Evaporation and purification by flash chro-

matography on silica gel in hexane/ethyl acetate, 2:1, gave 3-(6-methoxy-

3-nitro-pyridin-2-yl)-5-trifluoromethyl-3H-imidazole-4-carboxylic acid ethyl

1

ester as a yellow oil (3.4 g, 39% yield). MS m/e 361.0 [M + H]⁺; H

NMR (400 MHz, DMSO) δ ppm 8.76 (d, J = 8.9 Hz, 1H), 8.46 (s, 1H),

7.32 (d, J = 9.04 Hz, 1H), 4.12 (q, J = 7.07 Hz, 2H), 3.93 (s, 3H), 1.01

(t, J = 7.19 Hz, 3H).

Step c. 8-Methoxy-3-trifluoromethyl-5H-2,5,9,9b-tetraaza-cyclopenta-

[a]naphthalen-4-one (15). 3-(6-Methoxy-3-nitropyridin-2-yl)-5-triflu-

oromethyl-3H-imidazole-4-carboxylic acid ethyl ester (2.9 g, 8.0 mmol)

was dissolved in glacial acetic acid (45 mL). Into this mixture was added

water (23 mL) followed by sodium hydrogen sulfite (10 g, 80 mmol).

The mixture was heated to 105 ꢀC for 16 h. An additional 2 g of the

hydrogen sulfite is then added every 2 h until all starting material has

been consumed, as indicated by TLC. The mixture was diluted with

7632

dx.doi.org/10.1021/jm2009138 |J. Med. Chem. 2011, 54, 7621–7638

Journal of Medicinal Chemistry

ARTICLE

water, and the solids were filtered and collected. The solids were washed

with water followed by a small amount of chloroform and dried to give

8-methoxy-3-trifluoromethyl-5H-2,5,9,9b-tetraazacyclopenta[a]naph-

thalen-4-one as a gray/white solid (1.8 g, 79% yield). MS m/e 285.1

MeOH (40 mL) was added freshly powdered KOH (2.5 g, 44.6 mmol)

at room temperature under nitrogen. The resulting mixture was stirred

at 55 ꢀC for 2 h, cooled to room temperature, diluted with dichlor-

omethane, and poured into water. The organic extracts were separated

and dried over magnesium sulfate. Evaporation of the solvents provided

2-bromo-1-(3,5-dimethoxy-2-nitrophenyl)-4-methyl-1H-imidazole as

1

[M + H]⁺; H NMR (400 MHz, DMSO) δ ppm 11.68 (s, 1H), 8.95

(s, 1H), 7.66 (d, J = 8.66 Hz, 1H), 6.96 (d, J = 8.67 Hz, 1H), 3.94 (s, 3H).

Step d. 4-Chloro-8-methoxy-3-trifluoromethyl-2,5,9,9b-tetraaza-cy-

clopenta[a]naphthalene (16). 8-Methoxy-3-trifluoromethyl-5H-2,5,9,-

9b-tetraazacyclopenta[a]naphthalen-4-one (1.0 g, 3.5 mmol) was sus-

pended in POCl₃(11 mL) and heated to 120 ꢀC for 3 h. The POCl₃was

removed under reduced pressure and the residue taken in water and

neutralized with sodium bicarbonate. The resulting solids were filtered

and collected and dried to give 4-chloro-8-methoxy-3-trifluoromethyl-

2,5,9,9b-tetraazacyclopenta[a]naphthalene as a pale yellow solid (0.98 g,

99% yield). MS m/e 303.0 [M + H]⁺; ¹H NMR (400 MHz, DMSO) δ

ppm 9.36 (s, 1H), 8.33 (d, J = 8.81 Hz, 1H), 7.21 (d, J = 8.81 Hz, 1H),

4.09 (s, 3H).

1

an off-white solid (3.35 g, 100% yield). MS m/e 342.0 [M + H]⁺; H

NMR (400 MHz, DMSO) δ 7.18 (s, 1H), 7.01 (s, 1H), 6.82 (s, 1H),

3.95 (s, 3H), 3.9 (s, 3H), 2.09 (s, 3H).

8-Fluoro-3,4-dimethyl-1-(3-methylpyridin-4-yl)imidazo-

[1,5-a]quinoxalin-6-ol (35; R₁= 3-methylpyridin-4-yl). Into a

mixture of 8-fluoro-6-methoxy-3,4-dimethyl-1-(3-methylpyridin-4-yl)-

imidazo[1,5-a]quinoxaline (200 mg, 0.595 mmol in dichloroethane

(10 mL) was added boron tribromide (0.5 mL, 5.29 mmol) at 0 ꢀC.

The resulting mixture was warmed to 80 ꢀC overnight, cooled to room

temperature, quenched with aqueous K₂CO₃, and extracted with di-

chloromethane. The extracts were dried over anhydrous MgSO₄.

Evaporation and purification by column purification (ethyl acetate, then

4ꢀ8% MeOH in dichloromethane) provided 8-fluoro-3,4-dimethyl-

1-(3-methylpyridin-4-yl)imidazo[1,5-a]quinoxalin-6-ol as a light yellow

powder (150 mg, 78% yield). MS m/e 323.1 [M + H]⁺; ¹H NMR (400

MHz, DMSO) δ ppm 8.87 (s, 1H), 8.76 (d, J = 5.4 Hz, 1H), 7.80 (d, J =

5.3 Hz, 1H), 6.79 (dd, J = 10.5, 2.7 Hz, 1H), 6.15 (dd, J = 10.4, 1.5 Hz,

1H), 2.87 (s, 3H), 2.74 (s, 3H), 2.14 (s, 3H).

Step e. 8-Methoxy-4-methyl-3-trifluoromethyl-2,5,9,9b-tetraaza-cy-

clopenta[a]naphthalene (17). 4-Chloro-8-methoxy-3-trifluoromethyl-

2,5,9,9b-tetraazacyclopenta[a]naphthalene (0.2 g, 0.66 mmol) was dis-

solved in dry dioxane (4 mL). Into this mixture was added Pd(PPh₃)₄

(0.012 g, 5%mol) followed by trimethylaluminum (2 M/toluene;

1.6 mL, 3.3 mmol). The mixture was heated to 110 ꢀC for 2 h and then

cooled with an ice bath. Dilute HCl (2 mL) was slowly added to the

mixture and then neutralized with dilute sodium hydroxide (4 mL). The

reaction mixture was extracted with ethyl acetate, and the organic layers

were separated and combined. The combined extracts were washed with

water and brine and dried over MgSO₄. Evaporation and purification by

flash chromatography on silica gel in hexane/ethyl acetate 10: 1 gave

8-methoxy-4-methyl-3-trifluoromethyl-2,5,9,9b-tetraazacyclopenta[a]-

naphthalene as a white solid (0.12 g, 60% yield). MS m/e 283.0 [M +

H]⁺; ¹H NMR (400 MHz, DMSO) δ ppm 9.24 (s, 1H), 8.31 (d, J = 8.72

Hz, 1H), 7.18 (d, J = 8.84 Hz, 1H), 4.08 (s, 3H), 2.75 (s, 3H).

3,5-Difluoro-4-nitrophenol (22). Into a solution of 3,5-difluor-

ophenol (14.08 g, 108 mmol) and dichlomethane (150 mL) was added

fuming nitric acid (>90%, 15 mL) dropwise under a nitrogen atmo-

sphere at 0 ꢀC. After the addition, the resulting solution was stirred at

0 ꢀC temperature for 2 h. The mixture was then poured into cold water,

and the organic layer was separated. The aqueous layer was extracted

with dichloromethane, and the combined organics were washed with

brine and dried over magnesium sulfate. Condensation under vacuum and

purification by column chromatography using 10ꢀ30% ethyl acetate in

hexane provided 3,5-difluoro-4-nitrophenol as a yellow thick oil (5.1 g,

27% yield) (MS m/e 174 [M ꢀ H]⁺; ¹H NMR (400 MHz, DMSO) δ

ppm 11.75 (brs, 1H), 6.74 (m, 2H)) and the regioisomer 3,5-difluoro-2-

nitrophenol as a yellow solid (8.0 g, 42% yield) (MS m/e 174 [M ꢀ H]⁺;

¹H NMR (400 MHz, DMSO) δ ppm 12.16 (brs, 1H), 7.04 (m, 1H),

6.72 (m, 1H)).

2-Bromo-1-(3,5-dimethoxy-2-nitrophenyl)-4-methyl-1H-

imidazole (29; R₁= Br). Step a. 2-Bromo-1-(3,5-difluoro-2-nitro-

phenyl)-4-methyl-1H-imidazole (28; R₁= Br; R₆, R₈= F). A mixture of

1,3,5-trifluoro-2-nitrobenzene (2.2 mL, 18.6 mmol), 2-bromo-4-methy-

limidazole (3 g, 18.6 mmol), and K₂CO₃(5.66 g, 41 mmol) in DMF

(80 mL) was stirred at room temperature overnight. The mixture was

diluted with ethyl acetate and washed with water and dried over

magnesium sulfate. Evaporation and purification by column purification

using 10% ethyl acetate in dichloromethane as eluent provided the

product 2-bromo-1-(3,5-difluoro-2-nitrophenyl)-4-methyl-1H-imidazole

as a yellow powder (3.18 g, 54% yield). MS m/e 317.9 [M + H]⁺; ¹H

NMR (400 MHz, DMSO) δ ppm 8.01ꢀ8.08 (m, 1H), 7.77ꢀ7.82

(m, 1H), 7.28 (s, 1H), 2.08 (s, 3H).

6-Methoxy-3,4-dimethyl-1-(2-methylpyridin-3-yl)imidazo-

[1,5-a]quinoxalin-8-ol (38b; R₆= OMe, R₁= 2-methylpyridin-

3-yl). A mixture of 8-(benzyloxy)-6-methoxy-3,4-dimethyl-1-(2-methyl-

pyridin-3-yl)imidazo[1,5-a]quinoxaline (500 mg, 1.178 mmol) and Pd/C

(62.7 mg, 0.059 mmol) in a 250 mL flask was vacuumed and refilled with

nitrogen, and then THF (8 mL) and MeOH (8 mL) were added followed

by addition of ammonium formate (371 mg, 5.89 mmol). The final

mixture was stirred at 50 ꢀC for 3 h, cooled to room temperature, and

filtered through Celite. The Celite pad was washed with methanol and

ethyl acetate. The filtrate was evaporated to provide 6-methoxy-3,4-

dimethyl-1-(2-methylpyridin-3-yl)imidazo[1,5-a]quinoxalin-8-ol as an

1

off-white powder (362 mg, 92% yield). MS m/e 335.1 [M + H]⁺; H

NMR (400 MHz, DMSO) δ ppm 8.68 (m, 1H), 8.43 (s, 1H), 7.82 (m,

1H), 7.41 (m, 3H), 6.46 (d, J = 2.3 Hz, 1H), 5.95 (d, J = 2.3 Hz, 1H), 3.82

(s, 3H), 2.71 (s, 3H), 2.68 (s, 3H), 2.12 (s, 3H).

1,3-Difluoro-5-methoxy-2-nitrobenzene (41; R = Me). Into

a mixture of 3,5-difluoro-4-nitrophenol (5.1 g, 29 mmol) and potassium

carbonate (8.0 g, 58 mmol) in N,N-dimethylformamide (45 mL) was

added iodomethane (2.2 mL, 34.8 mmol) at room temperature. The

resulting mixture was stirred at room temperature overnight, and then

most of the solvent was removed under vacuum. The residue was diluted

with water and ethyl acetate. The organic layer was separated, and the

aqueous layer was extracted with ethyl acetate. The combined organics

were washed with brine and dried over magnesium sulfate. Condensa-

tion under vacuum and purification by column chromatography using

20% ethyl acetate in hexane provided 1,3-difluoro-5-methoxy-2-nitro-

benzene as a yellow oil (4.4 g, 81% yield). MS m/e 190 [M + H]⁺; ¹H

NMR (400 MHz, DMSO) δ ppm 7.04 (m, 2H), 3.85 (s, 3H).

5-(Benzyloxy)-1,3-difluoro-2-nitrobenzene (41; R = Bn).

Into a mixture of 3,5-difluoro-4-nitrophenol (14 g, 80 mmol) and

potassium carbonate (24.2 g, 176 mmol) in N,N-dimethylformamide

(150 mL) was added benzyl bromide (10.45 mL, 88 mmol) at room

temperature. The resulting mixture was stirred at room temperature

overnight, and then most of solvent was removed under vacuum. The

residue was diluted with water and ethyl acetate. The organic layer was

separated, and the aqueous layer was extracted with ethyl acetate. The

combined organics were washed with brine and dried over magnesium

sulfate. Evaporation under vacuum and purification by column chroma-

tography using 20% ethyl acetate in hexane provided 5-(benzyloxy)-1,3-

difluoro-2-nitrobenzene as a yellow solid (98% yield). MS m/e 264.8

Step b. 2-Bromo-1-(3, 5-dimethoxy-2-nitrophenyl)-4-methyl-1H-imi-

dazole (29; R₁= Br; R₆, R₈= OMe). Into a solution of 2-bromo-1-(3,5-

difluoro-2-nitrophenyl)-4-methyl-1H-imidazole (2.98 g, 9.4 mmol) in

7633

dx.doi.org/10.1021/jm2009138 |J. Med. Chem. 2011, 54, 7621–7638

Journal of Medicinal Chemistry

ARTICLE

[M]⁺; ¹H NMR (400 MHz, DMSO) δ ppm 7.25ꢀ7.35 (m, 5H), 7.2 (m,

2H), 5.2 (s, 2H).

concentrated to dryness, and the residue was dissolved in 30 mL of ethyl

acetate. The solution was washed twice with brine (2 ꢁ 30 mL), saturated

Na₂SO₃solution (20 mL), and brine (20 mL). All aqueous phases

were combined and extracted with ethyl acetate (2 ꢁ 50 mL). The

organic layers were combined and dried over magnesium sulfate.

Evaporation under vacuum to dryness provided 9-bromo-2-methoxy-

6,7-dimethylimidazo[1,5-a]pyrido[3,2-e]pyrazine as a light yellow

2-Methoxy-6,7-dimethyl-9-(4-methylpyridin-3-yl)imidazo-

[1,5-a]pyrido[3,2-e]pyrazine (74). Step a. 6-Methoxy-2-(4-methyl-

1H-imidazol-1-yl)-3-nitropyridine (3a). To a N,N-dimethylformamide

(500 mL) solution of 4-methylimidazole (8.5 g, 103 mmol) was added

freshly powdered KOH (6.72 g, 120 mmol) in two portions under N₂at

0 ꢀC, followed by addition of 2-chloro-6-methoxy-3-nitropyridine (18.9 g,

100 mmol). The resulting solution was warmed to room temperature and

stirredfor 2h. The majorityof the solvent wasremovedunder vacuum, and

the residue was diluted with water and extracted with ethyl acetate three

times. The organic layer was combined and washed two times with water

to remove additional dimethylformamide and dried over magnesium

sulfate. The solvent was evaporated under vacuum and the residue was

purified by column (15ꢀ25% gradient eluent of ethyl acetate in dichloro-

methane) to provide 6-methoxy-2-(4-methyl-1H-imidazol-1-yl)-3-nitro-

pyridine as a yellow solid (21.9 g, 93% yield, inseparable 4:1 mixture of

imidazole regioisomers). MS m/e 235.0 [M + H]⁺; ¹H NMR (400 MHz,

DMSO) δ ppm 8.48 (d, J = 8.7 Hz, 1H), 8.00 (d, J = 8.7 Hz, 1H), 7.18

(s, 1H), 7.01 (d, 1H), 3.97 (s, 3H), 2.12 (s, 3H).

Step b. 6-Methoxy-2-(4-methyl-1H-imidazol-1-yl)pyridin-3-amine

(3b). To a mixture of 6-methoxy-2-(4-methyl-1H-imidazol-1-yl)-3-nitro-

pyridine (21.4 g, 91.5 mmol) and 10% Pd/C (5.12 g, 4.58 mmol) in a 1 L

round-bottom flask (connected with a condenser) was loaded 240 mL of

THF, followed by slow addition of 240 mL of MeOH under N₂with

stirring. HCOONH₄(34.75 g, 503.25 mmol) was added in two portions

into the stirring mixture, and the final mixture was stirred at room tem-

perature for 10 min (gas released) and then warmed to 50 ꢀC for 1 h. The

mixture was cooled to room temperature and filtered through Celite.

Solvent was evaporated under vacuum to dryness to provide a clean

product as an off-white powder (18.6 g, 99% yield). NMR indicated a 4:1

ratio mixture of two regioisomers with the major product (6-methoxy-

2-(4-methyl-1H-imidazol-1-yl)pyridin-3-amine) as the desired regioi-

somer (confirmed by NOE studies). MS m/e 205 [M + H]⁺; ¹H NMR

(400 MHz, DMSO) δ ppm 7.91 (s, 1H), 7.30 (d, 1H), 7.25 (s, 1H), 6.63

(d, 1H), 4.70 (s, br, 2H), 3.70 (s, 3H), 2.13 (s, 3H).

Step c. N-(6-Methoxy-2-(4-methyl-1H-imidazol-1-yl)pyridin-3-yl)acet-

amide (4). To a solution of (6-methoxy-2-(4-methyl-1H-imidazol-1-yl)-

pyridin-3-amine (8.16 g, 40 mmol, 4:1 mix) in 200 mL of toluene was

added acetic anhydride (18.8 mL, 200 mmol) in dropwise. The resulting

mixture was stirred at room temperature for 3.5 h. The agitation was

stopped for 30 min and the precipitate was filtered to provide a product

as an off-white solid, 5.45 g (70% yield based on the major isomer), as a

single regioisomer. MS m/e 247 [M + H]⁺; ¹H NMR (400 MHz,

DMSO) δ ppm 9.58 (s, 1H), 8.00 (s, 1H), 7.72 (d, 1H), 7.30 (s, 1H),

6.80 (d, 1H), 3.84 (s, 3H), 2.12 (s, 3H), 1.95 (s, 3H).

Step d. 2-Methoxy-6,7-dimethylimidazo[1,5-a]pyrido[3,2-e]pyrazine

(5). To a solution of N-(6-methoxy-2-(4-methyl-1H-imidazol-1-yl)pyridin-

3-yl)acetamide (2.04 g, 8.2 mmol) in 16 mL of POCl₃was added P₂O₅

quickly (minimize the moisture induction). The resulting mixture was

refluxed at 110ꢀ120 ꢀC for 4 h. POCl₃was evaporated, and the residue was

quenched with iceꢀwater very carefully. The mixture was neutralized with

saturated Na₂CO₃solution and extracted with ethyl acetate. The organic

layerwas dried over magnesium sulfate. Condensation followedbycolumn

chromatography using 2ꢀ5% MeOH in dichloromethane as eluent

provided 2-methoxy-6,7-dimethylimidazo[1,5-a]pyrido[3,2-e]pyrazine as

a yellow powder, 1.12 g (55% yield). MS m/e 229.0 [M + H]⁺; ¹H NMR

(400 MHz, DMSO) δ ppm 8.82 (s, 1H), 8.07 (d, 1H), 6.95 (d, 1H), 3.99

(s, 3H), 2.69 (s, 3H), 2.64 (s, 3H).

1

powder (206 mg, 88% yield). MS m/e 306.9 [M + H]⁺; H NMR

(400 MHz, DMSO) δ ppm 8.08 (d, 1H), 7.01 (d, 1H), 4.04 (s, 3H),

2.67 (s, 3H), 2.62 (s, 3H).

Step f. 2-Methoxy-6,7-dimethyl-9-(4-methylpyridin-3-yl)imidazo[1,

5-a]pyrido[3,2-e]pyrazine (74). A mixture of 9-bromo-2-methoxy-6,7-

dimethylimidazo[1,5-a]pyrido[3,2-e]pyrazine (2.0 g, 6.53 mmol),

4-methylpyridine-3-boronic acid (1.79 g, 13.06 mmol), K₂CO₃(2.70

g, 19.60 mmol), and Pd(PPh₃)₄(150 mg, 0.1306 mmol) in a 250 mL

flask was vacuumed and flushed with nitrogen, followed by addition of p-

dioxane (120 mL) and water (40 mL). The final mixture was stirred at

90 ꢀC for 4 h and then cooled to room temperature. The reaction was

quenched with NH₄Cl solution, and the aqueous portion was extracted

with ethyl acetate. The combined organic layer was washed with brine

and dried over magnesium sulfate. Column chromatography using 50%

ethyl acetate in dichloromethane as eluent provided 2-methoxy-6,7-

dimethyl-9-(4-methylpyridin-3-yl)imidazo[1,5-a]pyrido[3,2- e]pyrazine

as an off-white powder (1.68 g, 81% yield). MS m/e 320 [M + H]⁺; ¹H

NMR (400 MHz, DMSO) δ ppm 8.55 (s, 1H), 8.50 (m, 1H), 8.10 (d,

1H), 7.40 (m, 1H), 6.85 (d, 1H), 3.10 (s, 3H), 2.75 (s, 3H), 2.70 (s, 3H),

2.05 (s, 3H).

8-Fluoro-6-methoxy-3,4-dimethyl-1-(3-methylpyridin-4-yl)-

imidazo[1,5-a]quinoxaline (96). Strep a. 1,5-Difluoro-3-meth-

oxy-2-nitrobenzene (24). To a solution of 3,5-difluoroanisole (30 g,

208 mmol) in 260 mL of dichloromethane was added HNO₃(>90%

fuming, 60 mL) dropwise at 0 ꢀC. The resulting solution was stirred at

0 ꢀC for 3 h and then washed withwater. The aqueous phase was extracted

with dichloromethane. The combined organics were washed with brine

and dried over magnesium sulfate. Evaporation and crystallization in

hexane and ethyl acetate provided 1,5-difluoro-3-methoxy-2-nitrobenzene

as an off-white powder (28.5 g, 72% yield). MS m/e 190.1 [M + H]⁺; ¹H

NMR (400 MHz, DMSO) δ ppm 7.22 (m, 2H), 3.91 (s, 3H).

Step b. 2-Bromo-1-(5-fluoro-3-methoxy-2-nitrophenyl)-4-methyl-1H-

imidazole (28; R₁= Br, R₆= OMe, R₈= F). A mixture of 1,5-difluoro-3-

methoxy-2-nitrobenzene (9 g, 47.6 mmol), 2-bromo-4-methylimidazole

(7.7 g, 47.6 mmol), and potassium carbonate (14.5 g, 104.7 mmol) in

DMF (240 mL) was stirred at room temperature overnight. The

majority of solvent was removed by rotovap, and the residue was diluted

with ethyl acetate and washed with water. The aqueous phase was

extracted with ethyl acetate, and the combined organics were washed

with water and brine and dried over magnesium sulfate. Condensation

and purification using 5ꢀ10% ethyl acetate in dichloromethane as eluent

provided of 2-bromo-1-(5-fluoro-3-methoxy-2-nitrophenyl)-4-methyl-

1H-imidazole as a yellow solid (4.8 g, 31% yield). MS m/e 330.0 [M +

H]⁺; ¹H NMR (400 MHz, DMSO) δ ppm 7.60 (m, 1H), 7.38 (m, 1H),

7.21 (s, 1H), 3.98 (s, 3H), 2.09 (s, 3H).

Step c. 4-(1-(5-Fluoro-3-methoxy-2-nitrophenyl)-4-methyl-1H-imi-

dazol-2-yl)-3-methylpyridine (28a; R₁= 3-methylpyridin-4-yl, R₆=

OMe, R₈= F). A mixture of 2-bromo-1-(5-fluoro-3-methoxy-2-nitro-

phenyl)-4-methyl-1H-imidazole (2.45 g, 7.4 mmol), 3-picoline-4-boronic

acid (1.52 g, 11.1 mmol), K₂CO₃(3.06 g, 22.2 mmol), and Pd(PPh₃)₄

(420 mg, 0.37 mmol) in a 250 mL flask was vacuumed and flushed with

nitrogen, followed by the addition of p-dioxane (120 mL) and water

(40 mL). The final mixture was stirred at 90 ꢀC for 8 h and then cooled

to room temperature. The reaction was quenched with NH₄Cl solution,

and the aqueous portion was extracted with ethyl acetate. The combined

organics were washed with brine and dried over magnesium sulfate.

Evaporation and purification by column chromatography using 50%

Step e. 9-Bromo-2-methoxy-6,7-dimethylimidazo[1,5-a]pyrido-

[3,2-e]pyrazine (6). To a mixture of 2-methoxy-6,7-dimethylimidazo

[1,5-a]pyrido[3,2-e]pyrazine (172 mg, 0.75 mmol) and NBS (200 mg,

1.13 mmol) was added anhydrous CH₃CN (6 mL) under N₂. The

resulting solution was stirred in the dark for 24 h. The mixture was

7634

dx.doi.org/10.1021/jm2009138 |J. Med. Chem. 2011, 54, 7621–7638

Journal of Medicinal Chemistry

ARTICLE

ethyl acetate in dichloromethane and 4% MeOH in dichloromethane as

eluent provided of 4-(1-(5-fluoro-3-methoxy-2-nitrophenyl)-4-methyl-

1H-imidazol-2-yl)-3-methylpyridine as a yellow solid (1.83 g (72%

dried over MgSO₄and filtered. The solvent was removed under reduced

pressure and the crude product purified by flash chromatography on silica

gel in hexane to provide 1-chloro-3-fluoro-2-nitro-5-(trifluoromethyl)-

benzene as a brown oil (1.17 g). MS m/e 242.8 [M]⁺; ¹H NMR

(400 MHz, DMSO) δ ppm 8.23ꢀ8.29 (m, 2H).

1

yield). MS m/e 343.1 [M + H]⁺; H NMR (400 MHz, DMSO) δ

ppm 8.49 (s, 1H), 8.26 (d, J = 5.3 Hz, 1H), 7.44 (m, 1H), 7.23 (m, 1H),

7.21 (s, 1H), 6.94 (d, 1H), 3.89 (s, 3H), 2.31 (s, 3H), 2.18 (s, 3H).

Step d. N-(4-Fluoro-2-methoxy-6-(4-methyl-2-(3-methylpyridin-4-

yl)-1H-imidazol-1-yl)phenyl)acetamide (31b; R₁= 3-methylpyridin-4-

yl, R₆= OMe, R₈= F, R = COMe). Into a mixture of 4-(1-(5-fluoro-3-

methoxy-2-nitrophenyl)-4-methyl-1H-imidazol-2-yl)-3-methylpyridine

(1.83 g, 5.3 mmol) and 10% Pd/C (300 mg, 0.265 mmol) was added

30 mL of THF, followed by the addition of MeOH (30 mL) under a

nitrogen atmosphere. Ammonium formate (2.0 g, 29.15 mmol) was

added to the mixture, and the mixture was stirred at 50 ꢀC for 2 h. The

mixture was filtered through a Celite pad, and the solvent was removed

under vacuum. The off-white solid residue was taken in toluene (45 mL),

and then acetic anhydride (5 mL, 53 mmol) was added. The resulting

mixture was stirred at 80 ꢀC for 4 h and cooled to room temperature.

The reaction was quenched with sodium bicarbonate solution, and the

aqueous portion was extracted 3ꢁ with dichloromethane. Evaporation

of the solvents and purification by column purification using 3%

methanol in dichloromethane as eluent provided N-(4-fluoro-2-meth-

oxy-6-(4-methyl-2-(3-methylpyridin-4-yl)-1H-imidazol-1-yl)phenyl)acet-

amide as an off-white solid (1.58 g 84% yield). EI MS m/e 355.1 [M +

H]⁺; ¹H NMR (400 MHz, DMSO) δ ppm 9.06 (s, 1H), 8.44 (s, 1H),

8.15 (d, J = 5.5 Hz, 1H), 7.05 ꢀ 6.95 (m, 3H), 6.38 (m, 1H), 3.78 (s,

3H), 2.35 (s, 3H), 2.15 (s, 3H), 1.82 (s, 3H).

Step e. 8-Fluoro-6-methoxy-3,4-dimethyl-1-(3-methylpyridin-4-yl)-

imidazo[1,5-a]quinoxaline (96). Into a mixture of N-(4-fluoro-2-meth-

oxy-6-(4-methyl-2-(3-methylpyridin-4-yl)-1H-imidazol-1-yl)phenyl)-

acetamide (1.58 g, 4.46 mmol) and P₂O₅(3.16 g, 22.3 mmol) was added

POCl₃(60 mL) under nitrogen. The resulting mixture was stirred at

120 ꢀC for 6 h and then cooled to room temperature. The majority of the

solvent was removed under vacuum, and the residue was diluted with

ethyl acetate. The mixture was transferred slowly into iced 50% NaOH

solution. The mixture was stirred for 10 min while maintaining the pH

above 7 and then extracted 3ꢁ with ethyl acetate. Condensation and

column purification using 4% MeOH in dichloromethane as eluent pro-

vided 8-fluoro-6-methoxy-3,4-dimethyl-1-(3-methylpyridin-4-yl)imidazo-

[1,5-a]quinoxaline as an off-white powder (880 mg 60% yield). MS

m/e 337.1 [M + H]⁺; ¹H NMR (400 MHz, DMSO) δ ppm 8.71 (s, 1H),

8.64 (d, J = 5.34 Hz, 1H), 7.52 (d, J = 4.89 Hz, 1H), 7.48 (d, 1H), 7.00

(dd, J = 11.17, 2.56 Hz, 1H), 6.19 (dd, J = 10.47, 2.56 Hz, 1H), 3.91 (s,

3H), 2.75 (s, 3H), 2.69 (s, 3H), 2.00 (s, 3H).

6-Chloro-3,4-dimethyl-1-(3-methylpyridin-4-yl)-8-(trifluoro-

methyl)imidazo[1,5-a]quinoxaline (115). Step a. 2-Chloro-6-

fluoro-4-(trifluoromethyl)aniline (25a). 2-Fluoro-4-(trifluoromethyl)-

aniline (5 g, 27.9 mmol) was dissolved in acetonitrile (100 mL). To this

was added N-chlorosuccinimide (4 g, 30.7 mmol). The mixture was

heated to 75 ꢀC for 16 h, then poured into water and extracted with

ether. The organic layer was separated and washed with saturated

aqueous sodium bicarbonate, then water and brine and dried over

MgSO₄. The solution was filtered and the solvent removed under

reduced pressure to provide 2-chloro-6-fluoro-4-(trifluoromethyl)aniline

as a yellow oil (5.2 g). MS m/e 212.8 [M]⁺; ¹H NMR (400 MHz, DMSO)

δ ppm 7.44 (m, 2H), 6.18 (brs, 2H).

Step b. 1-Chloro-3-fluoro-2-nitro-5-(trifluoromethyl)benzene (25b).

Sodium perborate tetrahydrate (7.3 g, 46.8 mmol) was suspended in

glacial acetic acid (30 mL) and heated to 50 ꢀC. Into this mixture was

added dropwise a solution of 2-chloro-6-fluoro-4-(trifluoromethyl)aniline

(2 g, 9.37 mmol) dissolved in glacial acetic acid (20 mL). The mixture

was stirred for 16 h at 50 ꢀC and then poured into water and extracted

with ethyl ether. The organic layer was separated and washed with water,

aqueous bicarbonate solution, and brine. The organic layer was then

Step c. 1-(3-Chloro-2-nitro-5-(trifluoromethyl)phenyl)-4-methyl-1H-

imidazole (28; R₁= H, R₆= Cl, R₈= CF₃). 1-Chloro-3-fluoro-2-nitro-

5-(trifluoromethyl)benzene (1.1 g, 4.5 mmol) and 4-methyl-1H-imida-

zole (0.371 g, 4.5 mmol) were dissolved in DMF (10 mL). To this was

added potassium carbonate (1.2 g, 9.0 mmol). The mixture was stirred at

room temperature for 16 h and then was poured into water and extracted

with ethyl acetate. The organic layer was separated and washed with

water, brine and dried over MgSO₄. The solution was then filtered and

the solvent removed under reduced pressure. The crude was purified by

flash chromatography on silica gel in 10:2 hexane/ethyl acetate to

provide 1-(3-chloro-2-nitro-5-(trifluoromethyl)phenyl)-4-methyl-1H-

imidazole as a tan solid (0.74 g). MS m/e 306.0 [M + H]⁺; ¹H NMR

(400 MHz, DMSO) δ ppm 8.46 (s, 1H), 8.24 (s, 1H), 7.82 (s, 1H), 7.11

(s, 1H), 2.12 (s, 3H).

Step d. 2-Chloro-6-(4-methyl-1H-imidazol-1-yl)-4-(trifluoromethyl)-

aniline (31a; R₁= H, R₆= Cl, R₈= CF₃). 1-(3-Chloro-2-nitro-5-

(trifluoromethyl)phenyl)-4-methyl-1H-imidazole (0.74 g, 2.4 mmol)

was dissolved in a solution of glacial acetic acid and ethanol (10 mL

each). Into this mixture was added iron powder (0.81 g, 14.4 mmol).

The mixture was heated to 100 ꢀC for 1 h. The mixture was poured into

aqueous 1 N sodium hydroxide and extracted with ethyl acetate. The

organic layer was separated and then washed with water, brine and dried

over MgSO₄. The resulting solution was filtered and the solvent removed

under reduced pressure to provide 2-chloro-6-(4-methyl-1H-imidazol-1-

yl)-4-(trifluoromethyl)anilineasanoff-whitesolid(0.68g). MSm/e 274.1

[M ꢀ H]⁺; ¹H NMR (400 MHz, DMSO) δ ppm 7.68ꢀ7.70 (m, 2H),

7.36 (s, 1H), 7.07 (s, 1H), 5.8 (brs, 2H), 2.15 (s, 3H).

Step e. N-(2-Chloro-6-(4-methyl-1H-imidazol-1-yl)-4-(trifluoromethyl)-

phenyl)acetamide (31b; R₁= H, R₆= Cl, R₈= CF₃). 2-Chloro-6-(4-

methyl-1H-imidazol-1-yl)-4-(trifluoromethyl)aniline (2 g, 7.26 mmol)

was dissolved in a solution of glacial acetic acid (20 mL) and acetic

anhydride (10 mL). Into this mixture was added 10 drops of concen-

trated sulfuric acid. The mixture was stirred at room temperature for 48 h

and then poured into iceꢀwater and extracted with ethyl acetate. The

organic layer was separated and washed with water, then saturated

Na₂CO₃and then water, brine, dried over MgSO₄, and filtered. The

solvent was removed under reduced pressure to afford N-(2-chloro-

6-(4-methyl-1H-imidazol-1-yl)-4-(trifluoromethyl)phenyl)acetamide

as a tan solid (1.7 g). MS m/e 318.1 [M + H]⁺; ¹H NMR (400 MHz,

DMSO) δ ppm 9.9 (s, 1H), 8.06 (s, 1H), 7.8 (s, 1H), 7.71 (s, 1H), 7.07

(s, 1H), 2.15 (s, 3H), 1.95 (s, 3H).

Step f. 6-Chloro-3,4-dimethyl-8-(trifluoromethyl)imidazo[1,5-a]qui-

noxalines (32a; R1 = H, R₆= Cl, R₈= CF₃). N-(2-Chloro-6-(4-methyl-

1H-imidazol-1-yl)-4-(trifluoromethyl)phenyl)acetamide (1.0 g, 3.15 mmol)

was suspended in phosphorus oxychloride (10 mL). Into this mixture

was added phosphorus pentoxide (3 g, 12.6 mmol). The mixture was

heated to 115 ꢀC for 16 h and then poured slowly into iceꢀwater/

methanol and was basified with 50% sodium hydroxide. The resulting

solution was extracted with ethyl acetate and the organic layer separated

and washed with water, brine and dried over MgSO₄. The solution was

then filtered and the solvent removed under reduced pressure. The

crude product was purified by flash chromatography on silica gel in 1:1

hexane/ethyl acetate to provide 6-chloro-3,4-dimethyl-8-(trifluoro-

methyl)imidazo[1,5-a]quinoxaline as a tan solid (0.57 g). MS m/e

1

300.0 [M + H]⁺; H NMR (400 MHz, DMSO) δ ppm 9.28 (s, 1H),

8.69 (s, 1H), 7.95 (s, 1H), 2.78 (s, 3H), 2.66 (s, 3H).

Step g. 1-Bromo-6-chloro-3,4-dimethyl-8-(trifluoromethyl)imidazo-

[1,5-a]quinoxalines (32b; R1 = Br, R₆= Cl, R₈= CF₃). 6-Chloro-3,4-

dimethyl-8-(trifluoromethyl)imidazo[1,5-a]quinoxaline (0.57 g, 1.902

7635

dx.doi.org/10.1021/jm2009138 |J. Med. Chem. 2011, 54, 7621–7638

Journal of Medicinal Chemistry

ARTICLE

mmol) was suspended in acetonitrile (50 mL). Into this mixture was

added N-bromosuccinimide (1.3 g, 7.6 mmol). The mixture was pro-

tected from light and stirred at room temperature for 48 h and then

poured into water and extracted with ethyl acetate. The organic layer was

separated and washed with water, then brine and dried over MgSO₄and

filtered. The solvent was removed under reduced pressure and the crude

product purified by flash chromatography on silica gel in 10:2 hexane/

ethyl acetate to provide 1-bromo-6-chloro-3,4-dimethyl-8-(trifluoro-

methyl)imidazo[1,5-a]quinoxaline as a tan solid 90.63 g). MS m/e

standard conditions on a 12 h light/dark cycle (lights on at 06:00 h) with

ad libitum access to water and food (ssniff M/R 15, Spezialdiaten

GmbH, Soest/Westfalen, Germany). Experiments were approved by the

Committee for Animal Care and Usage of the Federal State of Saxony

and carried out in accordance with German animal protection laws.

Experimental Procedure. Elevated locomotor activity and stereotypies

induced by the NMDA antagonist MK-801 are generally accepted as

relevant readouts in rodents for the positive symptoms of schizophrenia.

MK-801 induced stereotypies and hyperactivity in rats after intraper-

itoneal administration.²⁰Locomotor activity of the rats was recorded by

the MotiTest apparatus (TSE, Bad Homburg, Germany). The test area

consisted of a square arena (45 cm ꢁ 45 cm) with protective Plexiglas

walls (20 cm high) in which rats can move freely. Horizontal movements

were recorded by 32 infrared photocells arranged along the bottom of

each wall of the arena. The duration of activity was analyzed using the

“ActiMot” computer program (TSE, Bad Homburg, Germany). Stereo-

typed sniﬃng was scored by the experimenter every 5 min over a period

of 1 h (12 intervals) according to the method described by Andine

et al.²⁰The scores (0, no stereotyped sniﬃng; 1, discontinuous sniﬃng

with free interval of >5 s; 2, continuous sniﬃng) of the 12 intervals were

added together for each parameter. On the day of experimentation, rats

receive the test compound or vehicle 30 min prior to test onset; 0.1 mg/

kg MK-801 was administered intraperitoneally 10 min prior to testing.

The rats were placed in the center of the square arena of the MotiTest

apparatus 30 min before the start of the experiment to allow them to

become accustomed to the environment. The behavior of the rats was

then recorded for 1 h.

1

378.0 [M + H]⁺; H NMR (400 MHz, DMSO) δ ppm 9.28 (s, 1H),

8.11 (s, 1H), 2.79 (s, 3H), 2.68 (s, 3H).

Step h. 6-Chloro-3,4-dimethyl-1-(3-methylpyridin-4-yl)-8-(trifluoro-

methyl)imidazo[1,5-a]quinoxaline (115). 1-Bromo-6-chloro-3,4-dimeth-

yl-8-(trifluoromethyl)imidazo[1,5-a]quinoxaline (0.1 g, 0.264 mmol)

was suspended in a solution of dioxane (4 mL) and water (1 mL). Into

this mixture was added potassium carbonate (71 mg, 0.53 mmol),

followed by 3-methylpyridin-4-ylboronic acid (0.054 g, 0.396 mmol).

Argon was bubbled through the mixture, and tetrakis(triphenyl-

phosphine)palladium(0) (10%mol, 30 mg) was then added. The

mixture was sealed and heated to 110 ꢀC for 3 h and then diluted

with water and extracted with ethyl acetate. The organic layer was

separated, washed with brine, and dried over MgSO₄. Evaporation and

purification by flash chromatography on silica gel in ethyl acetate

provided 6-chloro-3,4-dimethyl-1-(3-methylpyridin-4-yl)-8-(trifluoro-

methyl)imidazo[1,5-a]quinoxaline as a white solid (70 mg, 45% yield).

MS m/e 391.1 [M + H]⁺; ¹H NMR (400 MHz, DMSO) δ ppm 8.64ꢀ8.66

(m, 1H), 7.99 (s, 1H), 7.5ꢀ7.6 (m, 2H), 7.11 (s, 1H), 2.9 (s, 3H), 2.78 (s,

3H), 2.11 (s, 3H).

8-(Cyclopropylmethoxy)-6-fluoro-3,4-dimethyl-1-(3-meth-

ylpyridin-4-yl)imidazo[1,5-a]quinoxaline (121). Step a. 6-Fluoro-

3,4-dimethyl-1-(3-methylpyridin-4-yl)imidazo[1,5-a]quinoxalin-8-ol

(38a; R₆= F, R₁= 3-methylpyridin-4-yl). Into a solution of 6-fluoro-8-

methoxy-3,4-dimethyl-1-(3-methylpyridin-4-yl)imidazo[1,5-a]quinoxaline

(92 mg, 0.274 mmol) in ClCH₂CH₂Cl (5 mL) was added dropwise boron

tribromide (0.129 mL, 1.368 mmol) at 0 ꢀC. The resulting orange sus-

pension was warmed to 80 ꢀC overnight. The reaction was quenched with

sodium carbonate solution, and the aqueous portion was extracted 3ꢁ with

ethyl acetate and dried over MgSO₄. Evaporation of the solvents and

purification by column chromatography (50ꢀ100% ethyl acetate in dichlo-

romethane, followed by 10% methanol in dichloromethane) provided

6-fluoro-3,4-dimethyl-1-(3-methylpyridin-4-yl)imidazo[1,5-a]quinoxalin-

8-ol as an off-white solid (80 mg, 91% yield). MS m/e 323.1 [M + H]⁺; ¹H

NMR (400 MHz, DMSO) δ ppm 10.39 (s, 1H), 8.69 (s, 1H), 8.61 (d, J =

4.8 Hz, 1H), 7.45 (d, J = 5.0 Hz, 1H), 6.72 (dd, J = 11.9, 2.5 Hz, 1H), 6.27

(dd, J = 2.2, 1.4 Hz, 1H), 2.75 (s, 3H), 2.70 (s, 3H), 2.00 (s, 3H).

Step b. 8-(Cyclopropylmethoxy)-6-fluoro-3,4-dimethyl-1-(3-methyl-

pyridin-4-yl)imidazo[1,5-a]quinoxaline (121). Into a mixture of 6-

fluoro-3,4-dimethyl-1-(3-methylpyridin-4-yl)imidazo[1,5-a]quinoxalin-

8-ol (60 mg, 0.186 mmol) and cesium carbonate (91 mg, 0.28 mmol) in

DMF (2 mL) was added a solution of (bromomethyl)cyclopropane

(40 mg, 0.296 mmol) in 1 mL of DMF. The resulting mixture was stirred

at 100 ꢀC for 1 h and cooled to room temperature. The mixture was

poured into water extracted with ethyl acetate, washed with water, and

then dried over magnesium sulfate. Evaporation and purification by

column chromatography (50ꢀ80% ethyl acetate in dichloromethane)

provided 8-(cyclopropylmethoxy)-6-fluoro-3,4-dimethyl-1-(3-methyl-

pyridin-4-yl)imidazo[1,5-a]quinoxaline as an off-white solid (13 mg).

MS m/e 377.1 [M + H]⁺; ¹H NMR (400 MHz, DMSO) δ ppm 8 0.72 (s,

1H), 8.63 (d, J = 5.0 Hz, 1H), 7.51 (d, J = 4.9 Hz, 1H), 7.00 (m, 1H), 6.22

(m, 1H), 3,38 (m, 2H), 2.78 (s, 3H), 2.72 (s, 3H), 2.02 (s, 3H), 1.02 (m,

1H), 0.56 (m, 2H), 0.18(m, 2H).

Statistics. Results were analyzed by one way analysis of variance

(ANOVA) when several groups were compared and by t test when

two groups were compared. Tukey test was used for individual compar-

isons. P < 0.05 was regarded as signiﬁcant.^21,22

Conditioned Avoidance Responding in Rats and Mice.

Male SpragueꢀDawley CD rats (350ꢀ450 g) or male CF-1 mice

(40ꢀ60 g) were individually housed and either maintained (rats) on a

food restricted schedule (15 g of standard rodent feed each day after

training/testing) or allowed free access to chow (mice). The testing

apparatus consists of a stainless steel and Plexiglas test chamber with two

8 in. ꢁ 6.5 in. ꢁ 8 in. compartments separated by an arched doorway

measuring 3.5 in. ꢁ 4.5 in. (MED Associates, Burlington, VT). Each of

the two floors of the shuttle box comprises a series of stainless steel grid

rods and is wired for the presentation of an electric foot shock (0.5 mA).

In addition, each side of the chamber is equipped with a stimulus light and

tone (Sonalert; Mallory Sonalert, Indianapolis, IN) and multiple infrared

beam source/detectors tolocate the subjects within thechamber. Subjects

that are trained to avoid the foot shock were placed in the experimental

chambers for a 4-min habituation period followed by 50 trials presented

on a 15 s variable interval schedule (range of 7.5ꢀ 22.5 s). Each trial

consisted of a 10 s warning tone and stimulus light (conditioned

stimulus) followed by a 10 s shock (unconditioned stimulus), presented

through the grid floor on the side where the animal was located, in the

presence of the tone and light. If an animal crossed through the archway

during the initial 10 s of the trial, the tone and light were terminated, and

the response was considered an avoidance response. If an animal crossed

through the archway after a foot shock was initiated, the tone, light, and

shock were terminated, and the response was considered an escape

response. If a response was made during an interval, the animal was

punished with a 0.5 s shock (0.5 mA). A MED Associates computer with

MEDSTATE Notation software controlled the test session and counted

the number of trials in which the animal avoided shock, escaped shock,

and did not respond. Only animals displaying stable performance

(approximately 90% avoidance responding on the training session

before test day for rats and 84% avoidance responding for mice) were

considered “trained” and included on test day. Training was maintained

by at least one nondrug test session each week. Eight animals received

Biology. MK-801-Induced Psychosis in Rats. Animals. Female

Wistar rats (Crl: (WI) BR from Charles River, Germany) weighing

160ꢀ180 g were used. They were housed in groups of five under

7636

dx.doi.org/10.1021/jm2009138 |J. Med. Chem. 2011, 54, 7621–7638

Journal of Medicinal Chemistry

ARTICLE

each dose of the test compounds, as well as a vehicle injection (within

subject design).

Structural Biology: Expression and Purification of the

Catalytic Domain of PDE10A. The catalytic domain of PDE10A,

spanning residues 339ꢀ479, was expressed from E. coli at 15 ꢀCfor16hwith

a thrombin-cleavable N-terminal His₆tag. The cells were collected and lysed

in 50 mM potassium phosphate (pH 7.5), 300 mM NaCl, 10% glycerol,

15 mM imidazole, 5 mM β-mercaptoethanol, and 2 mM benzamidine.

Following clarification, the lysate was bound to nickel-NTA (Qiagen) and

eluted with lysis buffer supplemented with 125 mM imidazole. The eluate

was diluted 6-fold with 50 mM Tris (pH 8.0), 1 mM DTT, and 10% glycerol

and loaded to a HiTrap Q HP column (GE Biosciences). The anion

exchange resin was subsequently washed in dilution buffer supplemented

with 50 mM NaCl, and PDE10A was eluted over a 20 column volume

gradient by competing with dilution buffer containing 1 M NaCl. Appro-

priate fractions were pooled, and the His tag was removed via thrombin

cleavage (Sigma). Following cleavage, thrombin was removed with p-

aminobenzamide agarose (Sigma) to terminate the reaction. PDE10A was

loaded to a Superdex 75 16/60 sizing column (GE Biosciences) pre-

equilibrated in 10 mM Hepes (pH 7.5), 150 mM NaCl, and 1 mM DTT.

Residual contaminants were removed by briefly batching binding to POROS

HS resin and collecting the unbound material containing PDE10A.

Data Deposition. The atomic coordinates of the PDE10A crystal

structure for compounds 74 (3SNI),¹⁹96 (3SN7), and 115 (3SNL)

have been deposited in the Protein Data Bank, Research Collaboratory

for Structural Bioinformatics, Rutgers University, New Brunswick, NJ.

Molecular Modeling Calculations. Docking calculations were

performed using the Schrodinger software package Maestro (www.

schrodinger.com). After the binding site model was validated by redock-

ing of the X-ray ligands, docking of analogues was performed using

the GLIDE or CombiGlide with SP (standard precision) scoring. Con-

formational analysis was performed using Macromodel with the OPLS-

AA force fields and a GBSA solvation model implemented in Maestro

(www.schrodinger.com).

In Vitro Phosphodiesterase Assays. Phosphodiesterases 1B,

2A, 3A, 4A, 5A, 7B, 8A, 9A, 10A, and 11A were generated from full-

length human recombinant clones. PDE6 was isolated from bovine

retina as described previously.¹⁸PDE activity was measured with the

preferred substrates. For PDE1B, 2A, 3A, 4A, 7B, 8A, 10A, and 11A,

cAMP was used, and for PDE5A, 6, and 9A, cGMP was used at a

concentration at or below the K_m. PDE1B was activated with calcium/

calmodulin and PDE2A with cyclic GMP. Inhibition of PDE activity was

measured by scintillation proximity assay at varied compound concen-

trations and optimized fixed enzyme concentrations for each enzymatic

assay. For determination of IC₅₀values, the Hill plot two-parameter

model was used.

Screening. Inhibition of Recombinant PDE10A (Baculovirus/SF21

System). TheDNA of PDE10A1 (AB 020593, 2340 bp) was synthesized

and cloned into vector pCR4.TOPO (Entelechon GmbH, Regensburg,

Germany). The gene was then inserted into a baculovirus vector, ligated

with the baculovirus DNA, and the enzyme protein was expressed in

SF21 cells. To isolate the enzyme, cells were first harvested by centrifu-

gation at 500g followed by resuspension in 50 mM Tris-HCl/1 mM

EDTA/250 mM sucrosebuffer, pH 7.4. (Sigma, Deisenhofen, Germany;

Merck, Darmstadt, Germany) and lysis by sonication (3 ꢁ 15 s,

Labsonic U, Fa. Braun, Degersheim, Switzerland, level setting “high”).

Cytosolic PDE10A was obtained by centrifugation at 48000g for 1 h. The

supernatant containing the active enzyme was stored at ꢀ70 ^οC. PDE

activity was determined by applying a one-step procedure in microtiter

plates. The reaction mixture contained 50 mM Tris-HCl/5 mM MgCl₂

buffer (pH 7.4. Sigma, Deisenhofen, Germany; Merck, Darmstadt,

Germany), 0.1 μM [³H]cAMP (Amersham, Buckinghamshire, U.K.),

and the enzyme in a total volume of 100 μL. Nonspecific activity was

determined in the absence of enzyme. The reaction was initiated by

addition of the substrate solution and carried out at 37 ꢀC for 30 min.

Enzymatic activity was stopped by addition of 25 μL Ysi-SPA beads

(Amersham-Pharmacia). After 1 h, the mixture was quantified in a liquid

scintillation counter for microtiter plates (Microbeta Trilux). Pipetting

of the incubation mixture was routinely performed using the Biomek

2000 (Beckman). The optimal amount of enzyme to use per assay was

determined for each enzyme preparation batch prior to usage in com-

pound testing. IC₅₀values were determined using the Hill plot, two-

parameter model.

’ ASSOCIATED CONTENT

S

Supporting Information. Elemental analysis data, analy-

b

tical HPLC purity data, and high resolution mass spectrometry

characterization data of intermediates and ﬁnal compounds not

listed in the Experimental Section. This material is available free

of charge via the Internet at http://pubs.acs.org.

Measurement of Striatal Cyclic Nucleotides. Male CF-1

(20ꢀ30 g) mice were dosed intraperitoneally with either 96 (0.3 and

3 mg/kg) with 2% Tween 80/0.5% hydroxyethylcellulose in water as

vehicle (n = 4 or 8). Control groups received vehicle at equal volumes

(10 mL/kg). Mice were sacrificed by focused microwave irradiation

(TMW-6402C, 4.0 kW, 1.15 s; Muromachi microwave applicator;

Muromachi Kikai Co., Ltd., Tokyo, Japan) 30 min after dosing with

indicated. The striata were removed and immediately frozen on dry ice.

Samples were stored at ꢀ80 ꢀC. For measurement of cyclic nucleo-

tides, frozen striatal tissue was pulverized at dry ice temperature,

weighed, and homogenized (Polytron) in 10 volumes of ice-cold 5%

trichloroacetic acid in water. The samples were centrifuged at 1500g at

4 ꢀC. The supernatants were extracted three times with 5 volumes of

water-saturated ether, and the aqueous phase was placed in a 70 ꢀC

heat block for 5 min to evaporate the residual ether. Samples were then

diluted 2- or 3-fold with ether-extracted 5% trichloroacetic acid in

water. Cyclic nucleotide concentrations (pmol/mg tissue) were de-

termined for the diluted samples with cAMP and cGMP enzyme

immunoassay kits (Cayman Chemical Company, Ann Arbor, MI).

Values are expressed as the fold change of the compound treatment

divided by the average of the vehicle-treated samples. Data were

graphed as the mean ( SEM using Prism software (GraphPad Soft-

ware, Inc., San Diego, CA).

Accession Codes

^†The atomic coordinates of the PDE10A crystal structure for

compounds 74 (3SNI), 96 (3SN7), and 115 (3SNL) have been

deposited in the Protein Data Bank, Research Collaboratory for

StructuralBioinformatics, Rutgers University, New Brunswick, NJ.

’ AUTHOR INFORMATION

Corresponding Author

*Telephone: 240-529-0866. Facsimile: 301-846-0794. E-mail:

malamas.michael@gmail.com.

Present Addresses

^{^}Lundbeck Research, 215 College Road Paramus, N.J., USA.

^#Vitae Pharmaceuticals, 502 West Oﬃce Center Drive, Fort

Washington, PA, USA.

^∞AstraZeneca, Mereside, Alderley Park SK10 4TG, UK.

’ ABBREVIATIONS USED

PDE10A, phosphodiesterase 10A; cAMP, cyclic adenosine mono-

phosphate; cGMP, cyclic guanine monophosphate; MSN, medium

spiny neuron; MK-801, dizocilpine maleate; MED, minimal

7637

dx.doi.org/10.1021/jm2009138 |J. Med. Chem. 2011, 54, 7621–7638

Journal of Medicinal Chemistry

ARTICLE

eﬀective dose;PDE, cyclic nucleotide phosphodiesterase; LMA,

locomotor activity; CAR, conditioned avoidance response;

TPSA, topological polar surface area; ADME, absorption, dis-

tribution, metabolism, and excretion; SAR, structureꢀactivity

relationship

(19) Data were collected at Southeast Regional Collaborative Access

Team (SER-CAT) 22-ID (or 22-BM) beamline at the Advanced Photon

Source, Argonne National Laboratory, IL. Supporting institutions may

be found at www.ser-cat.org/members.html. Use of the Advanced

Photon Source was supported by the U.S. Department of Energy, Oﬃce

of Science, Oﬃce of Basic Energy Sciences, under Contract No. W-31-

109-Eng-38.

(20) Andinꢀe, P.; Widermark, N.; Axelsson, R.; Nyberg, G.; Olofsson,

U.; Martensson, E.; Sandberg, M. Characterization of MK-801-induced

behavior as a putative rat model of psychosis. J. Pharmacol. Exp. Ther.

1999, 290, 1393–1408.

(21) SigmaStat, version 3.11; Systat Software, Inc.: Chicago, IL, 2004.

(22) SigmaPlot, version 9.0; Systat Software, Inc.: Chicago, IL, 2004.

(23) Wallace, A. C.; Laskowski, R. A.; Thornton, J. M. LIGPLOT: a

program to generate schematic diagrams of proteinꢀligand interactions.

Protein Eng. 1995, 8, 127–134.

’ REFERENCES

(1) Manallack, D. T.; Hughes, R. A.; Thompson, P. E. The next

generation of phosphodiesterase inhibitors: structural clues to ligand

and substrate selectivity of phosphodiesterases. J. Med. Chem. 2005,

48, 3449–3462.

(2) Odingo, J. O. Inhibitors of PDE4: a review of recent patent

literature. Expert Opin. Ther. Pat. 2005, 15, 773–787.

(3) Rotella, D. Phosphodiesterase 5 inhibitors: current status and

potential applications. Nat. Rev. Drug Discovery 2002, 1, 674–682.

(4) Houslay, M. D.; Schafer, P.; Zhang, K. Y. Keynote review:

phosphodiesterase-4 as a therapeutic target. Drug Discovery Today

2005, 10, 1503–1519.

(5) Bender, A. T.; Beavo, J. A. Cyclic nucleotide phosphodiesterases:

molecular regulation to clinical use. Pharmacol. Rev. 2006, 58, 488–520.

(6) Loughney, K.; Snyder, P. B.; Uher, L.; Rosman, G. J.; Ferguson,

K.; Florio, V. A. Isolation and characterization of PDE10A, a novel

human 30, 50-cyclic nucleotide phosphodiesterase. Gene 1999, 234,

109–117.

(7) Fujishige, K.; Kotera, J.; Omori, K. Striatum- and testis-speciﬁc

phosphodiesterases PDE10A: isolation and characterization of a rat

PDE10A. Eur. J. Biochem. 1999, 266, 1118–1127.

(8) Fujishige, K.; Kotera, J.; Michibata, H.; Yuasa, K.; Takebayashi,

S.; Okumura, K.; Omori, K. Cloning and characterization of a novel

human phosphodiesterase that hydrolyzes both cAMP and cGMP

(PDE10A). J. Biol. Chem. 1999, 274, 18438–18445.

(9) Soderling, S.; Bayuga, S. J.; Beavo, J. A. Isolation and character-

ization of a dualsubstrate phosphodiesterase gene family: PDE10A. Proc.

Natl. Acad. Sci. U.S.A. 1999, 96, 7071–7076.

(10) Soderling, S.; Beavo, J. A. Phosphodiestarase inhibitors. Curr.

Opin. Cell Biol. 1999, 12, 174–179.

(11) Zhang, K.; Card, G.; Suzuki, Y.; Artis, R.; Fong, D.; Gillette, S.;

Hsieh, D.; Neiman, J.; West, B.; Zhang, C.; Milburn, M.; Kim, S.-H.;

Schlessinger, J.; Bollag, G. A glutamine switch mechanism for nucleotide

selectivity by phosphodiesterases. Mol. Cell 2004, 15, 279–286.

(12) Seeger, T. F.; Bartlett, B.; Coskran, T. M.; Culp, J. S.; James,

L. C.; Krull, D. L.; Lanfear, J.; Ryan, A. M.; Schmidt, C. J.; Strick, C. A.;

Varghese, A. H.; Williams, R. D.; Wylie, P. G.; Menniti, F. S. Immuno-

histochemical localization of PDE10A in the rat brain. Brain Res. 2003,

985, 113–126.

(13) Kaiya, H. Second messenger imbalance hypothesis of schizo-

phrenia. Prostaglandins, Leukotrienes Essent. Fatty Acids 1992, 46, 33–38.

(14) Muly, E. C.; Smith, Y.; Allen, P.; Greengard, P. Signal transduc-

tion abnormalities in schizophrenia: the cAMP system. Psychopharmacol.

Bull 2002, 36, 92–105.

(15) Garver, D. L.; Carl Johnson, C.; Kanter, D. R. Schizophrenia

and reduced cyclic AMP production: evidence for the role of receptor-

linked events. Life Sci. 1982, 31, 1987–1992.

(16) Hoefgen, N.; Stange, H.; Rudolf Schindler, R.; Lankau, H.-J.;

Grunwald, C.; Langen, B.; Egerland, U.; Tremmel, P.; Pangalos, M. N.;

Marquis, K. L.; Hage, T.; Harrison, B. L.; Malamas, M. S.; Brandon, N. J.;

Kronbach., T. Discovery of imidazo[1,5-a]pyrido[3,2-e]pyrazines as a

new class of phosphodiesterase 10A inhibitiors. J. Med. Chem. 2010,

53, 4399–4411.

(17) Miyaura, N.; Suzuki, A. Palladium-catalyzed cross-coupling

reactions of organoboron compounds. Chem. Rev. 1995, 95, 2457–2483.

(18) Grauer, S.; Pulito, V.; Navarra, R.; Kelly, M.; Kelley, C.; Graf, R.;

Langen, B.; Logue, S.; Brennan, J.; Jiang, L.; Charych, E.; Egerland, U.;

Liu, F.; Marquis, K.; Malamas, M. S.; Hage, T.; Comery, T.; Brandon, N.

PDE10A inhibitors show anti-psychotic, pro-cognitive and negative

symptom activities suggesting broad-spectrum utility for the treatment

of schizophrenia. J. Pharmacol. Exp. Ther. 2009, 331, 574–590.

7638

dx.doi.org/10.1021/jm2009138 |J. Med. Chem. 2011, 54, 7621–7638

Article Doi

Highly potent, selective, and orally active phosphodiesterase 10A inhibitors

DOI: 10.1021/jm2009138

Source and publish data:

Authors:

Article abstract of DOI:10.1021/jm2009138

Full text of DOI:10.1021/jm2009138

Products guided by the article

R&D Labs maybe for 1159794-62-3

Relevant to this article

Hot Product