Discovery of two clinical histamine H3 receptor antagonists: Trans - N -ethyl-3-fluoro-3-[3-fluoro-4-(pyrrolidinylmethyl)phenyl] cyclobutanecarboxamide (PF-03654746) and trans -3-fluoro-3-[3-fluoro-4- (pyrrolidin-1-ylmethyl)phenyl]- N -(2-methy

ARTICLE

pubs.acs.org/jmc

Discovery of Two Clinical Histamine H₃Receptor Antagonists:

trans-N-Ethyl-3-fluoro-3-[3-fluoro-4-(pyrrolidinylmethyl)-

phenyl]cyclobutanecarboxamide (PF-03654746) and trans-3-

Fluoro-3-[3-fluoro-4-(pyrrolidin-1-ylmethyl)phenyl]-

N-(2-methylpropyl)cyclobutanecarboxamide (PF-03654764)

Travis T. Wager,*^,†Betty A. Pettersen,^†Anne W. Schmidt,^†Douglas K. Spracklin, Scot Mente,

Todd W. Butler, Harry Howard, Jr, Daniel J. Lettiere, David M. Rubitski, Diane F. Wong, Frank M. Nedza,

Frederick R. Nelson, Hans Rollema, Jeﬀrey W. Raggon, Jiri Aubrecht, Jody K. Freeman, John M. Marcek,

Julie Cianfrogna, Karen W. Cook, Larry C. James, Linda A. Chatman, Philip A. Iredale, Michael J. Banker,

Michael L. Homiski, Jennifer B. Munzner, and Rama Y. Chandrasekaran

Groton Laboratories, Pﬁzer Worldwide Research and Development, 558 Eastern Point Road, Groton, Connecticut 06340-5159, United States

S Supporting Information

b

ABSTRACT: The discovery of two histamine H₃antagonist

clinical candidates is disclosed. The pathway to identiﬁcation

of the two clinical candidates, 6 (PF-03654746) and 7

(PF-03654764) required ﬁve hypothesis driven design cycles.

The key to success in identifying these clinical candidates was

the development of a compound design strategy that leveraged

medicinal chemistry knowledge and traditional assays in con-

junction with computational and in vitro safety tools. Overall,

clinical compounds 6 and 7 exceeded conservative safety margins and possessed optimal pharmacological and pharmacokinetic

proﬁles, thus achieving our initial goal of identifying compounds with fully aligned oral drug attributes, “best-in-class” molecules.

’ INTRODUCTION

synthesis and release of histamine (Figure 1A).⁸Other H₃receptors

are heteroreceptors that regulate the release of neurotransmitters

including dopamine (DA) and norepinephrine (NE), such that

an H₃receptor antagonist should increase levels of these other

neurotransmitters in the synapse (Figure 1B).^6,9However, not all

H₃receptor antagonists have the same eﬀects on neurotransmit-

ter release; for example, some H₃antagonists increase DA and

NE levels while others do not.¹⁰In addition, H₃receptors can

regulate the release of acetylcholine (ACh), a neurotransmitter

involved in cognition.⁹

Collectively, the knowledge around histamine and H₃recep-

tors suggests that antagonism of the H₃receptor is a potential

therapeutic target for treatment of attention deﬁcit hyperactivity

disorder (ADHD), narcolepsy, Alzheimer’s disease (AD), and

possibly other cognitive disorders. The cardinal symptoms of

ADHD include impulsive behavior, hyperactivity, and inattention.

Typically, the hyperactivity component seen in school-age children

does not continue into adulthood, whereas the attention and

impulsivity symptoms do persist. The psychostimulant methyl-

phenidate, as well as other agents that successfully treat ADHD

symptoms, elevates extracellular levels of dopamine and norepi-

nephrine in the CNS. Recent microdialysis studies in rats show

Histamine receptors have been attractive and druggable

targets for decades.^1,2Histamine H₁receptor antagonists, such as

diphenhydramine and those that have limited central nervous

system (CNS) exposure such as cetirizine and fexofenadine, are

successful allergy medications. Histamine H₂receptor antago-

nists such as ranitidine and famotidine have proved to be eﬃcacious

gastric ulcer treatments. While the histamine H₃and H₄recep-

tors are also considered attractive and druggable targets, no marketed

agents currently exist for these receptors. Preclinical activity is

high, and early clinical candidates have been disclosed targeting

histamine H₃receptors.^3,4Herein we describe the discovery of

selective histamine H₃receptor antagonists.

Histamine has high aﬃnity for H₃receptors. The highest

density of H₃receptors is in the brain, primarily in prefrontal

regions, where histamine seems to play a key role in a variety of

centrally mediated functions including attention and learning/

memory, as well as arousal and wakefulness.^5,6Histamine levels

oscillate over a 24 h period, with concentrations highest during

the day and lowest at night. Furthermore, CNS-penetrant H₁

receptor antagonists induce marked sedation⁷and impairment of

cognitive performance, suggesting that histamine (HA) is in-

volved in maintaining wakefulness and may promote enhanced

attention and/or vigilance during the waking state. Some hista-

mine H₃receptors are inhibitory autoreceptors that regulate the

Received: July 14, 2011

Published: September 19, 2011

r

2011 American Chemical Society

7602

dx.doi.org/10.1021/jm200939b J. Med. Chem. 2011, 54, 7602–7620

|

Journal of Medicinal Chemistry

ARTICLE

Figure 1. (A) Histaminergic neuron. H₃autoreceptor, endogenous histamine (HA), acts at presynaptic H₃autoreceptors to reduce HA synthesis and

release (a negative feedback loop). (B) Nonhistaminergic neuron. Blockade of the H₃heteroreceptors should modulate the release of other

neurotransmitters.

that methylphenidate also increases levels of histamine and acety-

lcholine.^11,12Although psychostimulants are highly eﬀective agents

for treatment of patients with ADHD, there are several draw-

backs: these agents produce decreased appetite, insomnia, upset

stomach, and headache and additionally have abuse potential,

leading to their classiﬁcation as scheduled drugs. Atomoxetine, a

norepinephrine reuptake inhibitor that is also approved to treat

ADHD, acts by increasing extracellular levels of NE and DA, as

well as HA and ACh, in prefrontal areas of the brain.^11,12Although

atomoxetine is not a scheduled agent, some safety liabilities have

been reported and it bears a black box warning regarding liver

toxicity.^13,14Clearly, an unmet medical need exists in this area.

Herein we describe the discovery of selective histamine H₃

receptor antagonists to treat CNS disorders. Our goal was to

identify H₃receptor antagonists that elevate HA and ACh levels

in rat prefrontal cortex in vivo, possess a predicted pharmacoki-

netic (PK) proﬁle consistent with q.d. or b.i.d. dosing in humans,

and demonstrate a safety proﬁle beyond standard margins in

preclinical safety studies, as we were interested in considering a

pediatric indication for ADHD. The work described here focuses

less on in vitro potency, structureꢀactivity relationships (SARs),

and in vivo potency (typical medicinal chemistry benchmarks)

and more on development of the strategy we used to identify safe

and druglike molecules at the design stage, early in the drug

discovery process.

Figure 2. H₃receptor antagonist 1 emerged from a high-throughput

screen (HTS). Subsequent to our identiﬁcation of HTS hit 1, it was

reported in the literature by Morini et al.²³

intrinsic clearance (CL_int,u) for 1 was determined to be

unacceptably high (CL_int,u= 219 mL min^ꢀ1kg^ꢀ1).^16,17In

addition to this rapid clearance, two other significant issues

with 1 needed to be addressed. First, the compound had been

reported as a curarizing (muscle-relaxant) agent,¹⁸and second,

the planar nature of the biphenyl motif could pose genetic

safety issues. We were concerned with the potential long-term

safety risks inherent in a linear rigid chemotype that could

ultimately result in costly late-stage safety attrition (phase 3 or

postmarketing).

Exposure to polycyclic aromatic hydrocarbon compounds has

been shown to increase genotoxic risk,¹⁹and a crystal structure of

nogalamycin in DNA revealed that bulky dumbbell-shaped mole-

cules can intercalate between DNA base pairs.²⁰Further, dica-

tionic amidine compounds have been reported to bind in the

DNA minor groove.²¹Thus, our primary concern was that the

biaryldiamine template could act as a DNA-binding agent scaf-

fold, producing genetic damage and increasing patients’ cancer

risk. The initial in vitro micronucleus (IVMN) screening result

was negative for 1, but we remained apprehensive about idiosyn-

cratic toxicity for this general chemotype. Further supporting our

safety concerns, Hancock et al. reported positive ﬁndings in an

IVMN assay for several biaryl H₃receptor antagonists.²²To

reduce the potential liability of the biaryl template, we searched

for an aryl isostere that could mimic the distance between the two

basic groups. We hypothesized that a cyclobutyl moiety might act

as such an aryl ring isostere (Figure 3A). Molecular modeling

supported this hypothesis: minimized structures of biaryl and

’ RESULTS AND DISCUSSION

Initial HTS Screen and Analysis of Hits. The original H₃hit 1

(Figure 2) was identified from a high-throughput binding screen

and exhibited high affinity for human H₃receptors (K_i= 1.3 nM).

Structurally related hits (data not shown) were selective versus

other biogenic amine receptors and had weak hERG inhibition as

measured using a whole-cell voltage clamp technique (IC₅₀

>

5600 nM).¹⁵By use of a high-throughput in vitro human liver

microsome (HLM) assay and an in silico model of estimated

microsomal free fraction (cf_u,mic), the unbound microsomal

7603

dx.doi.org/10.1021/jm200939b |J. Med. Chem. 2011, 54, 7602–7620

Journal of Medicinal Chemistry

ARTICLE

Figure 3. (A) Isostere analysis of biaryl moiety vs cyclobutylaryl, where the black oval denotes motifs to be compared (isostere): molecular modeling

and overlap of biaryl (gray) and cyclobutylaryl (green) compounds. The two nitrogen atoms maintain an approximately equal distance from each other

in both molecules. (B) Same molecular overlap matched to surface of DNA structure (in blue, PDB code 302D). As designed, the cyclobutyl ring (green)

is not a matched ﬁt in the minor groove of DNA because of the bulkier cyclobutyl ring relative to the biaryl motif (gray).

arylcyclobutyl cores exhibited good overlap of the basic groups

(Figure 3A).

attribute of the HTS hit 1. Of the 286 dibasic cyclobutyl

analogues synthesized, one compound (2) possessed the profile

we sought for a potential drug candidate (Table 1). Compound 2

had a high affinity for both the human and rat H₃receptors (K_iof

2.2 and 7.7 nM, respectively) and demonstrated antagonist

activity in adenylate cyclase assays in both human and rat cell

lines expressing H₃, with K_iof 0.89 and 3.5 nM, respectively. To

assess selectivity, 2 was screened in a broad pharmacological

panel, both in-house and externally (CEREP), and exhibited

greater than 100-fold selectivity versus other biogenic amine

receptors, enzymes, and transporters (see Supporting In-

formation). In addition, 2 had a good early safety profile, as

characterized by a lack of projected drugꢀdrug interactions

(DDI),²⁵extremely weak hERG inhibition (hERG IC₅₀= 300

μM),¹⁵and negative results in the BiolumAMES,²⁶in vitro

micronucleus (IVMN), and human lymphocyte (HLA) assays.²⁷

Finally, the pharmacokinetic profile of 2 was excellent, with low

HLM clearance in vitro (CL_int,u),¹⁶high MDCK (MadinꢀDarby

canine kidney) passive apparent permeability,²⁸and a low P-gp

efflux ratio (ER) (see Table 1).²⁸Thus, our goal of replacing one

aryl ring with a cyclobutyl moiety while maintaining affinity for

In addition, we hypothesized that because of the three-

dimensional shape (3D-volume) of the cyclobutyl-containing

moiety, this moiety would be less likely to bind in the minor

groove. Using the X-ray crystal structure of a known minor

groove-binding drug (HOECHST 33258) bound to DNA as a

template,²⁴we examined the overlay matched to the surface

volume of DNA (Figure 3B). While the biaryl motif matches well,

the cyclobutyl rotates in a manner that gives a far worse ﬁt.

Although the quality of ﬁt was not quantiﬁed, we felt the

synthetic chemistry investment was worth the risk to test these

hypotheses. If one aryl group could be replaced with a cyclobutyl

isostere, we could potentially accomplish two goals: (1) building

a novel core and (2) increasing our odds of avoiding genetic

toxicity.

Design of the Initial Lead Compound (2). To test whether

replacing the biaryl moiety of the HTS hit 1 could be accom-

plished, an extensive set of dibasic cyclobutyl analogues was

synthesized (see section Chemistry for synthesis). This synthetic

effort yielded numerous compounds that maintained the potency

7604

dx.doi.org/10.1021/jm200939b |J. Med. Chem. 2011, 54, 7602–7620

Journal of Medicinal Chemistry

ARTICLE

Table 1. Selected Physicochemical Properties, in Vitro

Characteristics, and Safety and ADME Data for 2

parameter^a

value

ClogP

ClogD_7.4

MW

3.6

ꢀ0.2

314

TPSA

pK_a

15.7

Figure 4. (A) Electron microscopy image of rat lung, phospholipidosis

“foamy macrophage” induced by 2 in drug-treated group at 400 mg/kg:

multilamellated bodies (/); N = nucleus; TEM 10000ꢁ magniﬁcation.

(B) Electron microscopy image of drug storage. Cytoarchitecture of dog

heart includes the presence of multiple cardiomyocyte vacuoles (CMV, /)

induced by 2within the cytoplasm of cardiomyocytes: ID = intercalated disk

of cardiac muscle; G = glycogen; TEM 1000ꢁ magniﬁcation.

9.8 and 7.7

0.89

H₃K_ifunctional antagonist (human)^b(nM)

hERG IC₅₀^c(μM)

IVMN^d

CL_int,u^e(mL min^ꢀ1kg^ꢀ1

MDCK AB^f(cm/s)

P-gp ER^g

300

negative

<10

>10 ꢁ 10^ꢀ6

<1

)

microscopic examination, which revealed multilamellated mem-

branous lysosomal inclusions commonly referred to as multi-

lamellar bodies or MLBs (Figure 4A). These whorls are the

hallmark of phospholipidosis (PL), a storage disorder resulting in

excessive accumulation of phospholipids in lysosomes of the

tissues. The risk associated with compound-induced PL was

considered manageable provided that an acceptable therapeutic

index (TI) existed and that the effect was reversible.³²On the

basis of these findings, we embarked on an in vivo derisking

strategy requiring longer exploratory toxicology studies (14 days

in rat) that included an arm for reversibility of the PL finding (14-

day washout period), as well as a toxicology study in a second

species (dog). The next step with 2 was to conduct a 14-day study

in rats (7 males/dose) to determine if the threshold for phos-

pholipidosis decreased with increased duration of treatment (14

days versus 4 days). This 14-day study (5ꢀ30 mg/kg) indicated

that the threshold for phospholipidosis did not decrease with

longer treatment. In fact, no PL was observed at the highest

tested dose (30 mg/kg) tested. The safety margin for phospho-

lipidosis was large in rat, g588-fold PCEC (Table 11, Materials

and Methods). Although the safety profile from these rat studies

was favorable, a concurrent 7-day study in dogs (3 dogs at 10 and

25 mg/kg) identified a new finding at all tested doses: cardio-

myocyte vacuoles (CMV) were observed in heart tissue at low

multiples of the PCEC (<17-fold, Table 11). The CMV were

specifically located in the cardiac papillary muscle and character-

ized as drug storage vacuoles (Figure 4B). This safety finding

(CMV) was not manageable, as it had a low TI and no biomarker

to guide clinical safety. The CMV safety findings observed in

dogs precluded 2 from moving into regulatory toxicity studies.

In summary, while the initial design objective was achieved,

two unexpected ﬁndings arose for the dibasic chemotype, both

centered on safety (PL and CMV). Given the fact that our most

advanced chemical matter attrited in later-stage safety studies,

the discovery eﬀort cycled back to the compound design stage

and our initial strategy was re-evaluated. We asked ourselves the

following question: “Was de-risking PL the ideal approach?” Our

conclusion was “no”, that a better strategy, if possible, would be

to design molecules that did not induce PL. We therefore set a

conservative safety standard for compound progression: a com-

pound would not advance if it caused PL in vivo at otherwise

tolerated doses. It was hypothesized that moving to a more

rat f_u,p

0.72

^aCalculated physicochemical properties were obtained using standard

commercial packages: Biobyte for ClogP calculations, ACD/Labs for

ClogD at pH 7.4 and the most basic pK_a. ^bHuman functional assay

measuring cAMP utilizing a reporter gene assay (β-lactamase) in

HEK293 cells stably expressing human H₃receptor. K_iis the geometric

mean of ﬁve independent experiments. ^cBlockade of the hERG potas-

sium channel in HEK293 cells.^{15 d}In vitro micronucleus assay. ^eHuman

liver microsome unbound intrinsic clearance. ^fMS-based quantiﬁcation

of apical f basolateral transfer rate of a test compound at 2 μM across

g

contiguous monolayers of MDCK cells. Ratio of (basal f apical) to

(apical f basal) transfer rate of a test compound at 2 μM across

contiguous monolayers of MDR1-transfected MDCK cells.

the H₃receptor and minimizing the potential genetic toxicity was

realized through the discovery of 2.

The central nervous system eﬀect of 2 was demonstrated

using the R-α-methylhistamine (RAMH) induced dipsogenia

model.^29,30Compound 2dose-dependently blocked RAMH-induced

water drinking in rats with a subcutaneous ID₅₀of 0.78 mg/kg

(N = 1 experiment). The free drug levels were calculated using

plasma protein binding (f_u,p) measurements from a separate in

vitro assay and determined to be 23.0 nM, approximately 3-fold

higher than the K_i(rat H₃binding K_i= 7.7 nM). This result (3-fold the

rat H₃binding K_i) is consistent with antagonist pharmacology

requiring >75% receptor occupancy for eﬃcacy.³¹To allow

comparisons between the safety proﬁles of lead compounds, this

benchmark (3-fold K_i) is used throughout this manuscript as the

projected clinically eﬃcacious concentration (PCEC). One assu-

mption made in this comparison was that all compounds freely

cross the bloodꢀbrain barrier such that the ratio unbound brain

(C_b,u)/unbound plasma (C_p,u) is equal to 1.

Phospholipidosis. Compound 2 met or exceeded our initial

in vitro and in vivo efficacy criteria and was advanced into a 4-day

oral dose study in rats (3/sex per dose) to assess safety. Several

doses were tested in that study (5, 50, and 400 mg/kg). At 5 mg/

kg (mean AUC of 1245 ( 659 ng h/mL) there were no findings.

At g50 mg/kg, phospholipidosis (kidney and lung) was seen

3

(mean AUC of 17375 ( 6983 ng h/mL), while mortality was

3

observed at 400 mg/kg (mean C_maxof 3918 ( 1280 ng/mL).

Phospholipidosis was confirmed in the lung and kidney by electron

7605

dx.doi.org/10.1021/jm200939b |J. Med. Chem. 2011, 54, 7602–7620

Journal of Medicinal Chemistry

ARTICLE

Table 2. Safety Studies of Two Literature Compounds: Dis-

obutamide and Bidisomide^43,44

Figure 5. Development of a design strategy to remove safety liabilities.

The two-pronged strategy uses prospective design (focusing on optimal

physical properties) and the development and validation of an in vitro

screening assay to identify PL liability at an early stage.

of antagonizing the H₃receptor, as the evaluation of alternative

chemotypes did not show this toxicity. Hjelle et al. reported that

disobutamide was a potent inducer of clear cytoplasmic vacuoles

in various tissues (Table 2).⁴³The authors described disobuta-

mide as a molecule that possessed a dibasic motif and they

associated this structural feature with the observed toxicity. The

clear cytoplasmic vacuoles described by Hjelle et al. seemed similar

to those we observed, except that we observed them only in the

cardiac papillary muscle. Their solution to the clear cytoplasmic

vacuole problem was to remove or mask one of the basic amines

(Table 2). All this information suggested to us that elimination of

the CMV safety ﬁnding might be accomplished by simply removing

one basic center. It was crucial, of course, to demonstrate that

potency, safety attributes, and the desired absorption, distribution,

metabolism, and excretion (ADME) proﬁles could be retained in a

monobasic molecule.

Optimized Lead Compound 3. The overall compound de-

sign objective seemed simple enough: find a safe and well-tolerated

H₃receptor antagonist that would test the H₃hypotheses in the

clinic. For the next design cycle we employed the new design

strategy highlighted above (Figure 5). The physicochemical proper-

ties of 2 that we considered optimal were its low MW (314 Da),

moderate ClogP, low ClogD, and moderate to high basicity (pK_a

of basic centers of 7.2 and 9.8). At the time of our initial work we

tracked topological polar surface area (TPSA)⁴⁵and ClogP as a

surrogate for volume of distribution (tissue distribution), which

is thought to be a major factor contributing to in vivo PL.⁴⁶On

the basis of a recent analysis by Hughes et al., this combination

may have even more general utility as an early in vivo safety odds

predictor.³³Topological polar surface area for 2 was undesirable

(15 Å²) based on the cutoff (>75 Å²) provided by Hughes et al.³³

These authors concluded that compounds possessed a greater

safety risk when dual conditions ClogP > 3 and TPSA < 75 Å²

were satisfied (odds of seeing an adverse event at a total plasma

level of 10 μM test compound).³³We used 2 as a baseline for

comparison and to benchmark progress for all later compounds.

(Baseline values for 2 were the following: PL in silico value of

109 (19 higher than the target value 90) and in vitro PL MED =

37.5 μM, target value of g100 μM.) Employing this new design

strategy, we set out to identify a compound with improved

attributes.

favorable physicochemical property space could avoid this safety

ﬁnding.^33,34To increase our odds of success, in addition to phy-

sicochemical property optimization, we also needed to develop a

way to rank compounds according to their PL risk. Ideally, one

would be able to evaluate PL at the design stage and then re-

evaluate after synthesis. Rather than focusing on a single property, in

silico model, or in vitro end point, we employed a multipronged

design strategy (Figure 5). As part of this approach, we needed to

prospectively incorporate medicinal chemistry knowledge around

central nervous system (CNS) drugs into our design.^34ꢀ40First,

at the design stage we worked toward optimizing physicochem-

ical properties by lowering (a) lipophilicity, as assessed by calc-

ulated partition coeﬃcient (ClogP), (b) distribution coeﬃcient

at pH 7.4 (ClogD), (c) molecular weight (MW), and (d) calculated

pK_aof the most basic center. Simultaneously, we focused on

developing and validating both an in silico PL model and an in

vitro PL assay. Although we generated an in-house statistical PL

model (not discussed herein), we chose to use the intuitive PL

model described by Ploemen.⁴¹Ploemen et al. reported that if the

sum of the squares of the pK_aand ClogP values for a compound is

greater than or equal to 90, then the compound is at higher risk of

causing PL in vivo. Given the fact that H₃is a biogenic amine

receptor and our current hits possessed a basic amine, it was clear

that lowering the pK_awas potentially the most direct way to bring

the calculated value into a more desirable range. Indeed, the pK_a

contribution value in Ploemen’s PL model accounted for the

largest part of the sum of the squares (pK_a= 9.8, 9.8²= 96). A

second factor we considered was that a dibasic compound may

doubly impact the PL ﬁnding, which is not accounted for in

Ploemen’s PL model. Thus, in addition to the in silico PL model,

we examined an in vitro PL cell-based assay to serve as a surrogate

for in vivo study and thus improve our odds of identifying a safe

molecule. Generation of the in vitro PL values was accomplished

using a method similar to that recently reported by Morelli

et al.⁴²For the work describe herein, phospholipid accumulation

was assessed in a transformed human liver epithelial (THLE) cell

line and reported as a minimum eﬀect dose (MED) value in μM,

where lower MED values (see Materials and Methods for details)

predict a greater risk of PL in vivo.

The next compound that we identiﬁed with a suitable proﬁle

was 3 (Table 3). Monobasic compound 3 met all of our phar-

macology and ADME criteria (Tables 7ꢀ10) and most of our

early safety targets. The physicochemical properties of 3 were

similar to those of 2, with the most notable improvement being

In addition, we needed to understand the cause of CMV or, at

a minimum, develop a strategy that would remove this safety

liability. Parallel SAR eﬀorts (data not discussed herein) allowed

us to conclude that the CMV safety ﬁnding was not a direct result

7606

dx.doi.org/10.1021/jm200939b |J. Med. Chem. 2011, 54, 7602–7620

Journal of Medicinal Chemistry

ARTICLE

Table 3. Monobasic Cyclobutylaryl H₃Receptor Antagonist 3

Table 4. Monobasic H₃Receptor Antagonist, 4 with Reduced

pK_aRelative to Parent Compound 2

parameter^a

ClogP

value

parameter^a

ClogP

value

4.0

0.8

323

38

3.3

0.8

335

24

ClogD_7.4

MW

ClogD_7.4

MW

TPSA

pK_a

TPSA

pK_a

9.8

9.3

Early PL Safety Assessment

in silico PL

112

PL MED^b(μM)

119.5

in silico PL

PL MED^b(μM)

98

^aCalculated physicochemical properties were obtained using standard

commercial packages: Biobyte for ClogP calculations, ACD/Labs for

ClogD at pH 7.4 and the most basic pK_a. ^bTransformed human liver

epithelial cellular phospholipidosis assay measuring phospholipid accu-

mulation per cell. Compounds were classiﬁed as positive at the lowest

dose for which the induction of phospholipid accumulation was 3-fold

over background. MED value is the geometric mean of three indepen-

dent experiments.

245.3

^aCalculated physicochemical properties were obtained using standard

commercial packages: Biobyte for ClogP calculations, ACD/Labs for

ClogD at pH 7.4 and the most basic pK_a. ^bTransformed human liver

epithelial cellular phospholipidosis assay measuring phospholipid accu-

mulation per cell. Compounds were classiﬁed as positive at the lowest

dose for which the induction of phospholipid accumulation was 3-fold

over background. MED value is the geometric mean of three indepen-

dent experiments.

removal of one of the basic amines. The in silico PL value

increased to a less desirable value (112), driven by higher

lipophilicity relative to 2 (ClogP of 3.6 f 4.0), but the in vitro

PL value shifted upward toward a more favorable value (MED =

119 μM), meeting our target in this regard. Compound 3 was

synthesized on scale (see Chemistry) and tested in a 14-day rat

safety study (3/sex per dose). Given the potential risk of seeing

PL in vivo, we elected to run three doses in this safety study

(5, 50, 350 mg/kg). Phospholipidosis was observed in multiple

tissues (lung and liver) at g50 mg/kg (mean AUC 21,333 (

from 9.8 to 7.6) resulted in a significant drop in affinity for the H₃

receptor. We therefore set out to identify a compound with only a

moderate reduction in basicity (pK_adecreased to ∼9) that

retained a high affinity for the H₃receptor. From this effort, 4

was identified (see Chemistry for synthesis) as the next genera-

tion H₃receptor antagonist (Table 4). Compound 4 met our

pharmacology and ADME criteria (Tables 7ꢀ10) but now with

an improved in silico PL value (98) and in vitro PL assay value

(MED = 245 μM) relative to 2 and 3. Although compound 4 did

not meet the in silico PL target value of e90, the markedly

increased in vitro PL value was encouraging. The physicochem-

ical properties of 4 were desirable, with low MW, reduced ClogP,

increased TPSA, and a reduction in the pK_aof the most basic

center (Figure 6). The half log unit reductioninpK_a(from 9.8 to 9.3)

relative to the parent compound (2) was achieved through

incorporation of a halogen atom at the 2-position of the phenyl

ring. Compound 4 was evaluated in a 14-day toxicology study in

rats (6 males/dose) over a dose range of 10 to 300 mg/kg. All

doses were well tolerated, and most importantly no PL was

7,146 ng h/mL, day 1). As stated earlier, our compound selec-

3

tion criteria precluded compounds that exhibited PL in vivo, so

this compound’s development was terminated (TI g 7 PCEC;

see Table 11). The in silico PL model value (112 versus target

90) suggested an enhanced risk of PL in vivo, and this prediction

was consistent with the 14-day safety study ﬁndings wherein PL

was observed. We therefore placed more weight on this design

parameter in subsequent work. On the other hand, the in vitro PL

assay results had suggested a reduced risk of in vivo PL for 3

versus 2, which was not observed. At this time we did not

abandon the PL assay because we believed that there would still

be beneﬁt in rank-ordering compounds via an in vitro assay.

Having established that a monobasic compound could provide a

candidate proﬁle, the next generation compound was also

designed to be monobasic.⁴³Further, the use of both the in

silico PL model and the in vitro PL assay was retained within the

design strategy to better guide compound selection, and we

anticipated allowing some ﬂexibility in application of the PL

models based on holistic assessment of scores. We felt that using

hard cutoﬀ values would restrict design space and perhaps cause

undue hardship in the identiﬁcation of a potent and safe H₃

receptor antagonist.

observed (mean AUC of e4444 ng h/mL, day 1) (g36 PCEC,

3

Table 11). These findings suggested that the improved design

strategy (monobasic, moderate reduction of pK_a) was a viable

approach for the identification of a PL-free compound and that

the in vitro PL assay provided significant value. Notably, had we

driven decisions using a sharp cutoff in the in silico PL value, we

would not have made this compound (PL = 98). The successful

completion of the 14-day toxicology study in rats justified further

investment in compound 4. A second species toxicology study,

7-day study in dogs (2/sex), was initiated at 50 mg/kg (mean

AUC of 4777 ( 1970 ng h/mL, day 1). This dose was lowered

3

to 30 mg/kg on day 2 because of convulsions in 2 of 4 dogs

following a single dose at 50 mg/kg (C_maxof 3200ꢀ6200 ng/mL).

The dose was maintained at 30 mg/kg for the rest of this study

Next Generation Lead Compound (4). Earlier work from our

laboratories (not described herein) suggested that substantially

lowering the pK_aof the remaining basic amine (pK_adecreased

(mean AUC of 2911 ( 626 ng h/mL, day 7). Given the fact that

3

7607

dx.doi.org/10.1021/jm200939b |J. Med. Chem. 2011, 54, 7602–7620

Journal of Medicinal Chemistry

ARTICLE

Figure 6. Path to the H₃receptor antagonist clinical candidates. The red circles represent areas targeted for improvement, and the green circles

represent improvements made to that molecule. The in silico phospholipidosis (PL) cutoﬀ of e90 and the in vitro PL assay target of g100 μM were used

to identify compounds with reduced risk of observing PL in vivo. Labels + (ﬁnding) and ꢀ (no ﬁndings) denote risk in in vitro micronucleus (IVMN) or

ﬁndings for phospholipidosis (PL) and cardiomyocyte vacuoles (CMV) in an exploratory toxicology study (ETS).

4 was well-tolerated in the 14-day study in rats, we did not anticipate

any problems in the 7-day study in dogs. However, testicular

degeneration and CMV were observed at low multiples of the

projected clinically efficacious concentration (<29 PCEC, Table 11).

The findings in the papillary region of the heart were similar to

the CMV observed in dogs with 2 and precluded further develop-

ment of this compound. On the basis of previous exploratory

toxicology studies of monobasic compounds (data not shown), it

was not clear why CMV were observed with this monobasic

chemotype. One possible explanation was that hydrolysis of the

amide yielded a dibasic metabolite that produced CMV in the

7-day dog safety study. While we had no direct evidence that

metabolism of the amide in compound 4 yielded the free amine,

the clinical use of N-acyl prodrugs of amines makes this a

reasonable hypothesis.⁴⁷In any case, the identification of a

compound without an in vivo PL safety liability was a major

advancement. We attributed this success to a design strategy that

leveraged our understanding of how physicochemical properties

affect in vivo safety and appropriate utilization of in silico PL

model values and in vitro PL assay data.

strategy resulted in discovery of the potent H₃receptor antago-

nist 5 (Table 5). Structurally, 5 differs from 4 in two key features:

first, the exo-amide is internalized, and second, the chlorine atom

has been replaced by a fluorine. Metabolic cleavage of the endo-

amide would provide a carboxylate moiety, avoiding the potential

issues with a dibasic metabolite. We hypothesized that such a

carboxylate metabolite would not cause CMV. The fluorine atom

served the same purpose as the original chlorine atom (reduction

of pK_a) but with the added benefit of reducing MW (335 f 318)

and lipophilicity (ClogP = 3.3 f 2.6). Compound 5’s pharma-

cology, ADME (Tables 7ꢀ10) and early safety attributes were

similar to those of 4; in addition, it displayed a comparable in

silico PL value (94) and in vitro PL assay value (MED = 300 μM).

A 14-day toxicology study in rats (6 males/dose) was conducted

with 5 over a dose range of 10ꢀ300 mg/kg, and no PL was

observed (AUC up to 35600 ng h/mL). Further, this compound

3

was well tolerated up to a high multiple of the projected clinically

efficacious concentration (g371 PCEC, Table 11). These data

supported further development. A single-dose study in dogs (1/

sex per dose at 10 and 50 mg/kg; 1 male dog/dose at 75 and 100

mg/kg) identified 50 mg/kg (C_maxof 5875 ( 2227 ng/mL) as

the maximum tolerated dose for a subsequent 7-day study. In this

7-day study in dogs (2/sex), the dose was lowered from 50 mg/kg

(mean C_maxof 12859 ( 4473 ng/mL, day 1) to 25 mg/kg on day 2

because of convulsions in 2 of 4 dogs at 50 mg/kg. The dose was

Addressing CMV. Lead Compound 5. The hypothesis that

in vivo amide cleavage of 4 occurred and that the resulting meta-

bolite was responsible for the CMV finding seemed plausible. In

the next generation molecule, we therefore sought a compound

that possessed all the desired attributes achieved with 4 but

lacked the potential to generate a dibasic metabolite. This design

maintained at 25 mg/kg (mean AUC of 10780 ( 1368 ng h/mL,

3

7608

dx.doi.org/10.1021/jm200939b |J. Med. Chem. 2011, 54, 7602–7620

Journal of Medicinal Chemistry

ARTICLE

Table 5. Monobasic H₃Receptor Antagonist 5 with

Table 6. Clinical H₃Histamine Receptor Antagonists 6 and 7

endo-Amide Replacing the exo-Amide of Compound 4

parameter^a

value

parameter^a

ClogP

value for 6

value for 7

ClogP

ClogD_7.4

MW

2.6

0.4

318

24

2.4

0.03

322

32

3.4

0.9

350

32

ClogD_7.4

MW

TPSA

pK_a

TPSA

pK_a

9.3

9.2

Early PL Safety Assessment

in silico PL

PL MED^a(μM)

94

in silico PL

PL MED^a(μM)

90

96

300.0

302.4

^aCalculated physicochemical properties were obtained using standard

commercial packages: Biobyte for ClogP calculations, ACD/Labs for

ClogD at pH 7.4 and the most basic pK_a. ^aTransformed human liver

epithelial cellular phospholipidosis assay measuring phospholipid accu-

mulation per cell. Compounds were classiﬁed as positive at the lowest

dose for which the induction of phospholipid accumulation was 3-fold

over background. MED value is the geometric mean of three indepen-

dent experiments.

^aCalculated physicochemical properties were obtained using standard

commercial packages: Biobyte for ClogP calculations, ACD/Labs for

ClogD at pH 7.4 and the most basic pK_a. ^aTransformed human liver

epithelial cellular phospholipidosis assay measuring phospholipid accu-

mulation per cell. Compounds were classiﬁed as positive at the lowest

dose for which the induction of phospholipid accumulation was 3-fold

over background. MED value is the geometric mean of three indepen-

dent experiments.

day 7 values) for the remainder of the study. Excellent drug

exposure was achieved in this 7-day study, and there was no

histological evidence of CMV (g777 PCEC, Table 11). The

next-generation lead compound 5 therefore met all of our criteria

to move forward into clinical development. The redesigned H₃

receptor antagonist (endo-amide) yielded a compound that not

only retained the potency attributes of 4 but also improved the

overall safety profile (no CMV). The continual refinement of the

physicochemical properties, in silico PL value, and in vitro PL

value, coupled with further compound design (exo- to endo-

amide), resulted in retention of druglike attributes (potency,

specificity, ADME, and safety alignment), yielding a compound

that moved forward in the drug development process. Unfortu-

nately, genetic toxicology screening revealed a positive signal in

the IVMN.²⁷Follow-up testing indicated that the increased

incidence of micronuclei resulted from an equal contribution

of chromosome breakage and chromosome loss/gain. At this

point we had the option to further characterize the risk of these

findings and attempt to move 5 forward despite its genetic safety

liability. In pursuit of a “best-in-class” molecule, however, we

decided to further improve upon the profile of 5 and eliminate

the genotoxic liability.

Final Drug Candidates 6 (PF-03654746) and 7 (PF-

03654764).⁵²In the context of remaining true to our design

strategy, we considered the possibility that an amplified electro-

static potential within the core of compound 5 might have a

beneficial effect with respect to the IVMN liability. Our next

design cycle focused on identifying where we could incorporate a

fluorine atom without decreasing affinity for the H₃receptor. We

selected the cyclobutyl ring based on the assumption that the

resulting compound would retain affinity for the H₃receptor.

The additional fluorine served two purposes: (1) to perturb any

interactions with DNA that potentially existed with 5 and (2) to

provide a long-range electron-withdrawing effect that should

further reduce the pK_aof the basic amine. A second modification

made in this template served to compensate for the molecular

weight added by the second fluorine atom; we chose to remove

the methyl on the amide nitrogen in the expectation that the re-

duction in molecular weight and increase in polarity (incorporation

of one hydrogen bond donor) would provide the maximum

benefit with respect to overall physicochemical properties and

safety.³³Through these changes in this design cycle, we identified

6, one of the clinical candidates (Table 6). Overall, physico-

chemical properties were optimal from a CNS druglike perspec-

tive: ClogP (2.4), ClogD (0.03), MW (322), pK_a(9.2), and

TPSA (32). These physicochemical property values are in line

with the median values recently reported for a large set of

marketed CNS drugs.^34,48Lowering the ClogP and pK_aresulted

in an improved in silico PL value (90), and the in vitro PL assay

value remained high (MED = 300.0 μM). Compound 6 was

negative in the BiolumAMES and IVMN assays, and the negative

IVMN was confirmed in an exploratory cytogenetics assay.²⁷All

data supported the continued development of 6. In a 14-day

study in rats (4 males/dose), all doses (30, 100, 300 mg/kg) were

well tolerated (mean AUC of 189000 ng h/mL at 300 mg/kg,

3

day 14) and showed no evidence of PL. Treatment-related findings

were limited to transient postdose salivation at g100 mg/kg and

a slight decrease in food intake and body weight at 300 mg/kg

(g4181 PCEC, Table 11). The only remaining hurdle was de-

monstration of safety in a second species. The 7-day study in dogs

(2/sex) was initiated at 25 mg/kg. Because of convulsions observed

on day 3, this dose was lowered on day 4 to 15 mg/kg. There

were no treatment-related findings at 15 mg/kg (g1248 PCEC,

Table 11). The continual refinement of physicochemical proper-

ties, in silico PL value, and in vitro PL value, all while focusing on

alignment of drug attributes (potency, specificity, ADME, and

safety profile) yielded a compound that moved forward into drug

development; 6 has successfully completed regulatory toxicology

7609

dx.doi.org/10.1021/jm200939b |J. Med. Chem. 2011, 54, 7602–7620

Journal of Medicinal Chemistry

ARTICLE

Table 7. In Vitro Receptor Binding Aﬃnity

Table 10. In Vitro PL Values for H₃Antagonists

standard

human H₃

pK_i( SEM^a

rat H₃

pK_i( SEM^a

compd mean (μM)^adeviation (μM) lower^b(μM) upper^b(μM)

species

ratio

1

2

3

4

5

6

7

17.6

11.2

24.6

73.9

135.9

0.01

0.3

9.3

25.9

compd

(K_iin M)

K_i(nM)

(K_iin M)

K_i(nM)^b

37.5

19.3

55.7

1

2

3

4

5

6

7

8.88 ( 0.25

8.66 ( 0.10

8.06 ( 0.11

7.89 ( 0.06

8.40 ( 0.09

8.64 ( 0.09

8.84 ( 0.08

1.3

2.2

8.7

14

4

8.03 ( 0.07

8.11 ( 0.05

9.3

7.7

82

7

119.5

245.3

300.0

302.4

64.8

174.2

346.0

300.03

300.2

370.1

3.5

9.4

11

15

16

14

144.5

300.01

299.8

234.7

6.81 ( 0.05

7.22 ( 0.06

7.43 ( 0.05

7.73 ( 0.09

160

59

91.4

^aAt least three determinations. ^b95% conﬁdence limits.

2.3

37

1.4

19

^aData are expressed as geometric mean ( SEM (at least three

determinations except for rat K_iof 3 which was N = 1). ^bRat K_iwas

determined in rat brain for compound 1. All other compounds were

tested in HEK-293 cells expressing rat H₃receptor.

around antagonism of the histamine H₃receptor, including the

treatment of allergic rhinitis, narcolepsy, and Alzheimer’s disease

(results will be communicated separately). For clinical study

design details of 6, see ClinicalTrial.gov (http://clinicaltrials.

gov/ct2/results?term=pf-03654746). To further increase our

odds of identifying a drug, we continued to probe this highly

desirable drug space in an effort to identify a second high-quality

clinical candidate. The second clinical candidate, 7, was discovered

during follow-up work on the 6 template (Table 6). Compound 7

also has full alignment of the desired attributes (potency, phar-

macokinetics, and safety) and large safety multiples over the

PCEC in both preclinical species profiled (Table 7ꢀ11). While 6

and 7 are structurally similar, they differ in projected human

pharmacokinetics, where 7 was projected to have a longer T_1/2

(∼16 h), and in in vivo biological profiles (details to be described

elsewhere). Further, 7 has successfully completed regulatory

toxicology studies and phase I studies including safety and tole-

rability (results will be communicated separately). For clinical

study design details of 7, see ClinicalTrial.gov (http://clinical-

trials.gov/ct2/results?term=pf-03654764).

Table 8. Whole Cell Assay, H₃Receptor cAMP, β-Lactamase

Functional Aﬃnity^a

human H₃

rat H₃

pK_i( SEM

K_i

pK_i( SEM

K_i

species

ratio

compd

(K_iin M)

(nM)

(K_iin M)

(nM)

2

3

4

5

6

7

9.05 ( 0.23

7.93 ( 0.33

8.48 ( 0.08

8.83 ( 0.12

8.77 ( 0.10

8.98 ( 0.09

0.89

12.0

3.3

8.45 ( 0.07

3.5

79

3.7

6.5

6.4

14

7.68 ( 0.29

7.69 ( 0.05

8.05 ( 0.07

8.10 ( 0.11

21

1.5

21

1.7

8.9

7.9

5.2

6.6

1.2

^aData are expressed as geometric mean ( SEM (at least three

determinations except for compound 3, which was tested only once in

the rat functional assay).

’ CONCLUSION

There are numerous obstacles along the drug discovery

process, some of which slow and can even stop the development

of a compound or drug target. Unexpected ﬁndings such as

phospholipidosis, cardiomyocyte vacuoles, in vitro micronucleus,

or unusual structureꢀactivity relationships must be treated as

information to be exploited in crafting the next design hypoth-

esis. We have developed a design strategy that leverages medic-

inal chemistry knowledge, in conjunction with an in silico PL tool

and an early in vitro safety PL assay, to identify increasingly safer

H₃receptor antagonists that possess fully aligned attributes

(potency, pharmacokinetics, and safety). Critical to the success

in identifying compounds free of PL was identifying compounds

with PL MED > 200 μM and in silico PL of <98. Drug candidates

6 and 7 met or exceeded all of our design and biological screening

criteria and eﬀortlessly moved through both early safety studies

and regulatory toxicology studies. The path to these clinical

candidates was not a smooth one, but each step of the way generated

additional knowledge that was applied in the next design cycle.

In Vitro Pharmacology of Compounds. H₃receptor binding

Table 9. Pharmacokinetic (PK) Properties of Compound Set

parameter

2

3

4

5

6

7

Caco-2 AB ꢁ 10^ꢀ6(cm/s)^a

8.2

32.5 53.4 34.3 44.7 38.8

HLM T_1/2(min)^b

>120 >120 >120 >120 >120 >120

HLM CL_h(mL min^ꢀ1kg^ꢀ1

P-gp ER^d

)

<8

<5

0.96

c

0.95 0.90 1.27 1.29 1.1

plasma protein binding (f_u) (rat) 0.72 0.47

brain/plasma (rat) (1 h, sc)

F (rat) %

0.70 0.70 0.56

2.1

24.6 25.9 20.2

^aMS-based quantiﬁcation of apical f basolateral transfer rate of a test

compound at 2 μM across contiguous monolayers of Caco cells.

^bDerived from human liver microsomes. ^cMicrosomal unbound intrin-

sic clearance using human liver microsomes and an in silico model

estimate for f_u,mic. ^dRatio of (basal f apical) to (apical f basal) transfer

rate of a test compound at 2 μM across contiguous monolayers of

MDR1-transfected MDCK cells.

3

studies and phase I studies including safety, tolerability, and proof

of mechanism (POM) (results will be communicated separately).

The drive for a best-in-class molecule resulted in the identifica-

tion of an ideal molecule to broaden the scope of clinical testing.

The profile of 6 makes it a practical tool to test not only the

primary hypothesis around ADHD but several clinical hypotheses

affinities were determined by displacement of specific H-N-α-

methylhistamine binding to H₃receptor binding sites in HEK-

293 cells expressing full-length human or rat H₃receptors. Clinical

compounds 6 and 7 bound with high affinity to the human H₃

receptor and with lower affinity to the rat H₃receptor (Table 7),

in addition to having >1000-fold selectivity versus other human

7610

dx.doi.org/10.1021/jm200939b |J. Med. Chem. 2011, 54, 7602–7620

Journal of Medicinal Chemistry

ARTICLE

Table 11. In Vivo Assessment of H₃Antagonist Lead Compounds and Clinical Candidates

maximum tolerated dose study duration human protein projected human species C_max

,

species

C_max, (nM)^c

species AUC_(0ꢀ24)safety margin ꢁ

compd

(mg/kg)

(days)

binding (hf_u)

PCEC, C_p, nM^a

(ng/mL)^b

(ng h/mL)

PCEC^d

3

Rat Exploratory Toxicology Studies^e

2

3

4

5

6

7

g5

4

0.78

8.5

95 ( 100

302

1245 ( 659

15513 ( 1451

1278 ( 260

4444^c

12700^c

189000^c

g35

g30

g5

14

0.78

0.44

0.24

0.57

0.66

0.34

8.5

1573 ( 189

130 ( 141

2107 ( 803

2482^c

5002

402

g588

g7

59.3

175

21

g300

g100

g300

6292

7794

43486

22990

g36

g371

g4181

g1869

10.4

12.3

14020 ( 2087

8057 ( 3083

67400^f

Dog Exploratory Toxicology Studies^g

2

3

4

5

6

7

<10

ND^h

7

ND^h

0.78

8.5

46.7 ( 0.9

148.5

ND^h

740 ( 167

ND^h

<17

ND^h

0.44

0.24

0.57

0.66

0.34

59.3

175

21

ND^h

<30

7

1717 ( 295

5197 ( 635

4185 ( 1762

6302 ( 1370

5127

2911 ( 626

10780 ( 1368

18910 ( 10775

18175 ( 222

<29

g25

g15

16320

12980

17982

g777

g1248

g1462

10.4

12.3

^aHuman projected clinical eﬃcacious concentration (PCEC) was calculated (3 ꢁ human H₃binding K_i) and converted to total plasma concentration

(C_p) using measured human plasma protein binding (hf_u). ^bT_max≈ 0.5ꢀ1 h postdose. ^cMean value. ^dAverage multiples over the projected clinical

eﬃcacious concentrations in the species examined at C_maxof maximum tolerated dose (corrected for species diﬀerences in plasma protein binding). ^eRat

oral exploratory toxicology study (ETS) study. ^fDay 1 value. ^gDog oral 7 day ETS. ^hND = not determined.

histamine receptor subtypes (H₁, H₂, and H₄). Compounds 6

and 7 displayed >1000-fold selectivity versus other G-protein-

coupled receptors (GPCRs) and transporters.

Selective deprotection of the benzyl ether via palladium black

catalyzed hydrogenolysis transformed 10 into alcohol 11, which

was converted to a mixture of the corresponding cis and trans

tosylates 12a and 12b, respectively. At this stage, the isomers

were separated by chromatography. When heated to 120 °C with

morpholine in dimethylacetamide, cis-tosylate 12a provided

compound 2 in moderate yield. Detosylation of trans-tosylate

12b was accomplished by heating with magnesium turnings in

methanol to give alcohol 13. S_NAr reaction of 2-chloropyrimi-

dine with the potassium alkoxide of 13 generated pyrimidine 3 in

good yield.

The synthesis of exo-amide H₃receptor antagonist 4 is

depicted in Scheme 2. Pyrrolidine amide 14 was prepared in

excellent yield via coupling of 4-bromo-2-chlorobenzoic acid and

pyrrolidine using propylphosphonic anhydride (T3P). Subse-

quent reduction of the amide with boraneꢀTHF yielded benzylic

amine 15. The conversion of 15 into 16 began with the selective

bromineꢀmetal exchange of 15 with n-BuLi at ꢀ78 °C. An amount

of 2 equiv of this lithiated intermediate was added to 3-oxocy-

clobutanecarboxylic acid, and the resultant cyclobutanoic acid

carbinol was treated in situ with methylamine and T3P to provide

amide 16. Dehydration with TFA aﬀorded oleﬁn 17, which was

directly hydrogenated using Wilkinson’s catalyst to predomi-

nantly yield trans-cyclobutylamide 18. The trans selectivity is

presumed to have arisen from a coordination eﬀect between the

carbonyl group of the amide and the metal of the Wilkinson’s

catalyst, as suggested by Schultz and McCloskey for similar hy-

drogenations of amido-substituted cyclohexanes using Crab-

tree’s catalyst.⁵⁰BoraneꢀTHF reduction of amide 18 gave amine

19, which was acetylated with acetic anhydride to aﬀord H₃

receptor antagonist 4 in high yield.

Clinical candidates 6 and 7 displayed potent antagonist pro-

perties in functional assays measuring cAMP utilizing a reporter

gene assay (β-lactamase) in HEK293 cells stably expressing full-

length human or rat H₃receptors (Table 8). Results were con-

sistent with the species diﬀerences observed in the binding assays.

Pharmacokinetic (PK) Properties of Compound Set (2, 3, 4,

5, 6, and 7). Compounds 6 and 7 both possessed excellent ADME

properties (Table 9). Both compounds were well absorbed, as evi-

denced by both their in vitro properties (Caco-2 A f B, >38 ꢁ

10^ꢀ6cm/s) and the results of in vivo studies suggesting complete

absorption. The bloodꢀbrain barrier penetration was specifically

tested with 6 (considered representative of the two), and the results

indicated that the compound readily penetrated into the brain

(brain/plasma ratio of 2.1), with no evidence for efflux mechanisms

moderating its central disposition (MDR1 BA/AB = 1.1). In human

hepatic microsomes, both compounds were metabolically stable

(HLM T_1/2> 120 min) and were predicted to have low clearance

(<5 mL min^ꢀ1kg^ꢀ1) in humans.

’ CHEMISTRY

The synthesis of the H₃receptor antagonists 2 and 3 is

outlined in Scheme 1. The preparation of these lead compounds

began with etheriﬁcation of (3,3-dimethoxycyclobutyl)methanol⁴⁹

using benzyl bromide and potassium tert-butoxide to aﬀord the

corresponding benzyl ether. Unmasking of the latent ketone was

accomplished with aqueous p-toluenesulfonic acid, generating 8.

Lithiumꢀhalogen exchange was carried out on 1-(4-bromoben-

zyl)pyrrolidine via treatment with n-butyllithium; reaction with

ketone 8 produced carbinol 9 as a mixture of cis and trans

isomers. Access to cyclobutane 10, as a mixture of isomers, was

achieved by dehydroxylation of 9 with TFA and triethylsilane.

Synthesis of the endo-amide H₃receptor antagonist 5 is

described in Scheme 3. Nucleophilic displacement of 4-bromo-

2-ﬂuorobenzyl bromide with pyrrolidine generated benzylic

amine 20 in high yield. Selective lithiation of 20 using n-BuLi

7611

dx.doi.org/10.1021/jm200939b |J. Med. Chem. 2011, 54, 7602–7620

Journal of Medicinal Chemistry

ARTICLE

Scheme 1. Synthesis of Lead Compounds 2 and 3^a

^aReagents and conditions: (a) KO-t-Bu, benzyl bromide, THF, rt (69%); (b) p-TsOH (cat.), acetone/water, 65 °C (58%); (c) (i)

1-(4-bromobenzyl)pyrrolidine, n-BuLi, THF, ꢀ78 °C, (ii) 8, THF, ꢀ78 °C (71%); (d) triethylsilane, TFA, 75 °C (taken on crude); (e) palladium black,

45 psi of H₂, EtOH, rt (two steps, 79%); (f) p-toluenesulfonyl chloride, pyridine, CH₂Cl₂, 0 °C to rt (93%); (g) 12a, morpholine, DMA, 120 °C (61%);

(h) 12b, Mg turnings, MeOH, rt (73%); (i) KO-t-Bu, THF, 2-chloropyrimidine, rt (89%).

Scheme 2. Synthesis of Lead Compound 4^a

^aReagents and conditions: (a) T3P, pyrrolidine, EtOAc, rt (96%); (b) BH₃THF, THF, rt (74%); (c) (i) n-BuLi, 3-oxocyclobutanecarboxylic acid,

3

THF, ꢀ78 °C to rt; (ii) T3P, CH₃NH₂, rt (48%); (d) TFA, DCE, 75 °C (taken on crude); (e) Rh(Ph₃P)₃Cl, H₂(45 psi), EtOH, 60 °C (two steps, 37%);

(f) BH₃THF, THF, 70 °C (77%); (g) Ac₂O, Et₃N, CH₂Cl₂, 0 °C (>95%).

3

under cryogenic conditions and use of the optimized conditions

described in Scheme 2 aﬀorded carbinol 21. Dehydration of 21

aﬀorded cyclobutene 22, which was not stable to column chro-

matography and was used without further puriﬁcation. Reduc-

tion of the double bond was accomplished via hydrogenation;

using 10% Pd/C as catalyst in the presence of hydrogen yielded

the desired cis-cyclobutane 5 in good yield over two steps.

The syntheses of the two clinical H₃receptor antagonists, 6

and 7, consisted of three linear steps from 20 (Scheme 4) and

leveraged a late coupling step to install diversity (ethyl or isobutyl).

Overall, the synthesis was eﬃcient, requiring only two reaction

vessels to complete the process. In the ﬁrst reaction vessel all the

requisite carbons were installed. Aryl bromide 20 was lithiated

and added to a precooled solution of 3-oxocyclobutanecarboxylic

acid in THF. The resulting cyclobutanoic acid carbinol was con-

verted to the requisite amide using the coupling reagent T3P and

appropriate amine (ethyl or isobutyl) to yield carbinols 23 or 24,

respectively. Conversion of the alcohol group to a ﬂuorine was

7612

dx.doi.org/10.1021/jm200939b |J. Med. Chem. 2011, 54, 7602–7620

Journal of Medicinal Chemistry

ARTICLE

Scheme 3. Synthesis of Compound 5^a

^aReagents and conditions: (a) pyrrolidine, CH₃CN, 10 °C to rt (90%); (b) (i) n-BuLi, 3-oxocyclobutanecarboxylic acid, THF, ꢀ78 °C,

(ii) ethylmethylamine, T3P, EtOAc, rt (60%); (c) TFA, CH₂Cl₂, reﬂux (taken on crude); (d) 10% Pd/C, H₂(45 psi), EtOH, rt (two steps, 71%).

Scheme 4. Synthesis of Clinical Candidates 6 and 7^a

^aReagents and conditions: (a) (i) n-BuLi, 3-oxocyclobutanecarboxylic acid, THF, ꢀ78 °C, (ii) ethylamine or isobutylamine, T3P, EtOAc, rt (R = ethyl,

49%; R = isobutyl, 63%); (b) BAST, CH₂Cl₂, ꢀ78 °C (R = ethyl, 58%; R = isobutyl, 85%).

accomplished using the ﬂuorinating agent BAST (bis(2-methox-

yethyl)aminosulfur triﬂuoride). Thus, reaction of carbinols 23

and 24 with BAST aﬀorded 6 and 7, respectively, with high trans

selectivity. We hypothesize that the observed selectivity was im-

parted either through a stabilized carbocation or simply as a result

of a steric interaction that caused the ﬂuorinating reagent to

approach the face opposite the carbonyl. Removal of the minor

isomers was achieved by either preparative chromatography of

the free base or recrystallization of the HCl salt.

Through extensive salt evaluation, we discovered that the

tosylate salt possessed optimal properties and this form of 6 was

evaluated to determine its acceptability for tablet development.

The material was determined to be a crystalline nonsolvated

monotosylate salt with a melting onset temperature of 165 °C.

Analysis of the moisture sorption isotherm at 25 °C, collected

using the kinetic vacuum method, indicated that this solid-state

form is nonhygroscopic. Compound 6 tosylate salt has a solubi-

lity of greater than 10 mg/mL in 0.1 M phosphate-buﬀered saline

(ﬁnal pH 7) and a solubility of 23.6 mg/mL in unbuﬀered water

(ﬁnal pH 3.8). The solubility in 0.1 M phosphate-buﬀered saline

containing 0.5% sodium taurocholate/pamitoyloleoylphosphati-

dylcholine salts (ﬁnal pH 6.5) was determined to be greater

than 11 mg_A/mL (active component). The apparent solubility of

the 6 tosylate salt in simulated gastric ﬂuid (pH 1) was greater

than 15 mg/mL (active component). Overall, the high aqueous

solubility of this compound in all simulated biological media

suggests that the candidate would remain well solubilized in the

gastrointestinal tract.

number of hydrogen bond donors (HBD), and the calculated most basic

center (pK_a). Calculated physicochemical properties were obtained

using standard commercial packages: Biobyte for ClogP calculations

and ACD/Labs for ClogD at pH 7.4 and the most basic pK_a. TPSA was

calculated using the method developed by Ertl.⁵¹

Chemistry. General Information. All solvents and reagents were

obtained from commercial sources and were used as received. All

reactions were followed by TLC (TLC plates F254, Merck) or LCMS

(liquid chromatographyꢀmass spectrometry) analysis. Proton and ¹³C

NMR spectra were obtained using deuterated solvent on a Varian 300,

400, or 500 MHz instrument. All proton shifts are reported in δ units

(ppm) downfield of TMS (tetramethylsilane) and were measured

relative to the signals for chloroform (7.27 ppm) and methanol (3.31 ppm).

All ¹³C shifts are reported in δ units (ppm) relative to the signals

for chloroform (77.0 ppm) and methanol (49.1 ppm) with ¹H decoupled

observation. All coupling constants (J) are reported in hertz (Hz). NMR

abbreviations are as follows: app, apparent; br, broadened; s, singlet; d,

doublet; t, triplet; q, quartet; p, pentuplet; m, multiplet; dd, doublet of

doublets; ddd, doublet of doublet of doublets; sept, septuplet. GCMS

data were recorded on an HP 6890 GC system equipped with a 5973

mass selective detector (stationary phase, HP-1, fused silica, 12 m ꢁ

0.202 mm, 0.33 μm; temperature limits of ꢀ60 to 325 °C; ramp rate of

30 °C/min; solvent delay = of 0.4 min). Analytical analyses by Ultra

performance liquid chromatography (UPLC) were performed on a

Waters Acquity system with PDA detection (UV 210 nm) at 45 °C, flow

rate of 0.5 mL/min, with a gradient of 95/5 buffer/acetonitrile (0ꢀ7.55 min),

10/90 buffer/acetonitrile (7.55ꢀ7.85 min), 95/5 buffer/acetonitrile

(8.10ꢀ10.30 min) using the following columns and buffers: Waters

BEH C8 column (2.1 mm ꢁ 100 mm, 1.7 μm) with 50 mM sodium

perchlorate/0.1% phosphoric acid or 10 mM ammonium bicarbonate as

buffer; Waters BEH RP C18 column (2.1 mm ꢁ 100 mm, 1.7 μm) or

Waters HSS T3 (2.1 mm ꢁ 100 mm, 1.8 μm) column with 0.1%

methanesulfonic acid buffer. All melting points are uncorrected. Ele-

mental analyses were performed by QTI Development. Mass spectra

were recorded on a Micromass ADM atmospheric pressure chemical

’ MATERIALS AND METHODS

Physicochemical Property Data. This analysis included calcu-

lated lipophilicity (ClogP), calculated lipophilicity at pH 7.4 (ClogD),

molecular weight (MW), topological polar surface area (TPSA), the

7613

dx.doi.org/10.1021/jm200939b |J. Med. Chem. 2011, 54, 7602–7620

Journal of Medicinal Chemistry

ARTICLE

1

ionization instrument (APCI). High resolution mass spectra were

obtained on an Agilent LCMS TOF instrument equipped with a Zorbax

Eclipse column (50 mm ꢁ 4.6 mm, 1.8 μm XDB-C18) using 0.1% aqueous

formic acid as mobile phase A1 and acetonitrile containing 0.1% formic

acid as mobile phase B1. Column chromatography was carried out on

white solid (30.6 g, 99%). H NMR (MeOH-d₄, 400 MHz) ∼1.5:1

mixture of rotamers, δ 7.70ꢀ7.73 (m, 1H), 7.50 and 7.46 (2 d, J = 1.7,

1H total), 7.40 and 7.36 (2 dd, J = 7.9, 1.5, 1H total), 4.57 (s, 2H),

3.70ꢀ3.82 (m, 3H), 3.53ꢀ3.60 (m, 2H), 3.24 (m, 2H), 3.24 and 3.14

(2 s, 3H total), 2.72ꢀ2.86 (m, 1H), 2.17ꢀ2.38(m,9H), 2.02ꢀ2.12(m, 2H).

¹³C NMR (MeOH-d₄, 100 MHz) δ 175.7, 175.4, 151.6, 151.5, 136.2,

134.3, 129.4, 129.3, 127.6, 127.5, 127.4, 127.3, 57.1, 55.9, 55.2, 54.7, 38.3,

37.2, 37.0, 36.0, 32.8, 32.6, 31.0, 30.4, 24.0, 20.2, 19.8. HRMS m/z, calcd

(M + 1) 335.1884, observed 335.1880.

ꢀ

silica gel 60 (32ꢀ60 mesh, 60 Å) or on prepacked Biotage columns. The

purities of final compounds 2ꢀ7 as measured by UPLC and/or

elemental analysis methods were found to be above 95%.

4-({cis-3-[4-(Pyrrolidin-1-ylmethyl)phenyl]cyclobutyl}-

methyl)morpholine (2). Morpholine (33.6 mL, 385 mmol) was

added to 12a (76.3 g, 191 mmol) in dimethylacetamide (DMA) (191 mL),

and the mixture was heated at 120 °C for 1 h. The mixture was cooled

to room temperature and concentrated. EtOAc and concentrated

HCl (32 mL) were added. Solvents were removed in vacuo and two

additional portions of EtOAc were added followed by reconcentra-

tion to help remove residual DMA. EtOAc (600 mL) was added to

the resulting residue, and stirring was continued for 2 days. The

solvent was decanted off, and the gummy residue was dissolved in

3:1 CHCl₃/MeOH and basified with 15% aqueous NaOH. The

organic layer was concentrated and purified by chromatography

using a gradient of 3ꢀ50% MeOH in CH₂Cl₂containing 0.3%

NH₄OH, affording the title compound (36.8 g, 61%) as a solid.

This was dissolved in EtOAc, treated with 1 N HCl in THF, and

filtered to yield the HCl salt as a white solid. ¹H NMR (MeOH-d₄,

400 MHz) δ 7.42 (AB quartet, Δν_AB= 50.7, J_AB= 8.3, 4H), 4.35

(s, 2H), 4.03 (dd, J = 12.9, 3.7, 2H), 3.84 (ddd, J = 13.3, 13.3, 1.7, 2H),

3.49ꢀ3.59 (m, 1H), 3.45 (br d, J = 12.0, 4H), 3.28 (d, J = 7.1, 2H),

3.13ꢀ3.22 (m, 4H), 2.77ꢀ2.87 (m, 1H), 2.61ꢀ2.68 (m, 2H),

2.10ꢀ2.22 (m, 2H), 1.96ꢀ2.05 (m, 4H). ¹³C NMR (MeOH-d₄,

100 MHz) δ 146.7, 130.4, 128.8, 127.2, 63.8, 62.0, 57.7, 53.5, 51.9,

36.5, 34.6, 26.5, 22.6. HRMS m/z, calcd (M + 1) 315.2430, observed

315.2442.

2-({trans-3-[4-(Pyrrolidin-1-ylmethyl)phenyl]cyclobutyl}-

methoxy)pyrimidine (3). Potassium tert-butoxide (1.0 M in THF,

22.0 mL, 22.0 mmol) was added to a solution of 13 (3.6 g, 14.7 mmol)

and 2-chloropyrimidine (1.85 g, 16.2 mmol) in THF (14 mL) to give a

cloudy, dark pink mixture. After 10 min, saturated aqueous NH₄Cl was

added and the mixture was extracted with 3:1 CHCl₃/i-PrOH. The

organic layers were concentrated to give 5.2 g of the crude title com-

pound. Chromatography using a gradient of 3ꢀ8% MeOH in CH₂Cl₂

containing 0.2% NH₄OH yielded the title compound (4.2 g, 89%) as a

semisolid. The besylate salt was prepared in MeOH using 1 equiv of

benzenesulfonic acid to yield 6.4 g of a light yellow solid. A portion of

this material was recrystallized from water to provide an analytical

sample. ¹H NMR (MeOH-d₄, 400 MHz) δ 8.58 (d, J = 4.6, 2H), 7.81ꢀ

7.83 (m, 2H), 7.37ꢀ7.45 (m, 7H), 7.10 (t, J = 5.0, 1H), 4.56 (d, J = 7.1,

2H), 4.33 (s, 2H), 3.76 (tt, J = 8.6, 8.2 1H), 3.42ꢀ3.50 (m, 2H),

3.13ꢀ3.31 (m, 2H), 2.75ꢀ2.87 (m, 1H), 2.31ꢀ2.42 (m, 4H), 2.08ꢀ

2.22 (m, 2H), 1.90ꢀ2.06 (m, 2H). ¹³C NMR (MeOH-d₄, 100 MHz)

δ 166.6, 160.9, 149.7, 146.5, 131.7, 131.4, 129.7, 129.5, 128.6, 127.1, 116.6,

72.1, 59.2, 54.9, 37.7, 32.2, 31.4, 23.9. HRMS m/z, calcd (M + 1)

324.2070, observed 324.2062.

cis-N-Ethyl-3-[3-fluoro-4-(pyrrolidin-1-ylmethyl)phenyl]-

N-methylcyclobutanecarboxamide (5). Following the general

procedure described for preparing olefin 17, compound 21 (71.76 g,

214.6 mmol) yielded 182.0 g of crude olefin 22 as the TFA salt, N-ethyl-

3-[3-fluoro-4-(pyrrolidin-1-ylmethyl)phenyl]-N-methylcyclobut-2-ene-

1-carboxamide, as a brown oil. This was dissolved in EtOH (1.5 L) and

hydrogenated (∼45 psi) with 10% Pd/C (10 g) at room temperature for

1.5 h. After filtration and concentration, the resulting oil was dissolved in

EtOAc (500 mL), washed with a solution of K₂CO₃(60 g) in water

(400 mL), then with brine (200 mL), and dried over MgSO₄. Filtration

and concentration afforded 66.26 g of an orange oil, which was purified

by chromatography, flushing first with CH₂Cl₂and then eluting with 5%

MeOH in CH₂Cl₂to yield 41.75 g (58%) of pure title compound and

9.98 g (14%) of ∼85ꢀ90% pure material. The cleaner material (41.75 g,

131.1 mmol) was dissolved in EtOAc (1 L); 2 N HCl in Et₂O (80 mL,

160 mmol) was added over ∼1 min with vigorous stirring. After 30 min,

the light orange precipitate was collected, rinsed with EtOAc, and dried

to yield the corresponding HCl salt of 5 (36.15 g). A portion was

recrystallized from MeOH/EtOAc to provide an analytical sample as a

1

white solid, mp 196ꢀ196.5 °C. H NMR (300 MHz, CDCl₃) ∼1:1

mixture of rotamers, δ 12.69 (br s, 1H), 7.79 (t, J = 7.9, 1H), 7.07ꢀ7.00

(m, 2H), 4.20 (d, J = 5.4, 2H), 3.57ꢀ3.64 (m, 2H), 3.47ꢀ3.53 (m, 2H),

3.16ꢀ3.30 (m, 2H), 2.91 and 2.88 (2 s, 3H total), 2.79ꢀ2.85 (m, 2H),

2.50ꢀ2.60 (m, 2H), 2.34ꢀ2.46 (m, 2H), 2.14ꢀ2.26 (m, 2H), 1.95

ꢀ2.06 (m, 2H), 1.14 and 1.07 (2 t, J = 7.1, 3H total). ¹³C NMR (100

MHz, CDCl₃) δ (mixture of rotamers) 173.3, 162.7, 160.2, 150.6, 150.5,

133.6, 133.6, 124.0, 124.0, 114.4, 114.3, 114.0, 113.8, 52.6, 49.9, 49.9,

44.1, 42.7, 35.3, 34.3, 33.5, 32.8, 32.4, 23.2, 14.1, 12.4. HRMS m/z, calcd

(M + 1) 319.2180, observed 319.2188. Anal. Calcd for C₁₉H₂₇FN₂O

3

HCl: C, 64.30; H, 7.95; N, 7.89. Found: C, 64.36; H, 8.02; N, 7.97.

trans-N-Ethyl-3-fluoro-3-[3-fluoro-4-(pyrrolidin-1-ylmethyl)-

phenyl]cyclobutanecarboxamide (6). A solution of 23 (48.7 g,

152.0 mmol) in CH₂Cl₂(450 mL) was added over 50 min down the

reaction flask walls to a ꢀ78 °C solution of bis(2-methoxyethyl)aminosulfur

trifluoride (BAST) (42.0 mL, 227.8 mmol) in CH₂Cl₂(375 mL). After

2.5 h at ꢀ78 °C, the cooling bath was removed and the mixture was

stirred for 16 h at room temperature. With stirring, aqueous NaHCO₃

was carefully added in portions until all foaming subsided. Solid K₂CO₃

was then added to ensure that the pH was >8. The phases were

separated, and the aqueous phase was extracted with additional CH₂Cl₂

(2 ꢁ 100 mL). The organic phases were combined, dried over MgSO₄,

filtered, and concentrated to yield a dark orange-brown oil (50.2 g). This

crude material was purified by chromatography, eluting with EtOAc and

10% MeOH/EtOAc to yield 28.21 g (58%) of the title compound. The

monotosylate salt of 6 was prepared by addition of a solution of p-

toluenesulfonic acid (2.50 g, 13.14 mmol) in EtOAc (70 mL) to a

stirring solution of the free base (4.18 g, 13.0 mmol) in EtOAc (35 mL).

The resulting mixture was stirred for an additional hour. The white

precipitate was filtered, rinsed with EtOAc, and air-dried to afford 6.32 g

of the tosylate salt of 6. This material was dissolved in MeOH, filtered to

remove particulates, and reconcentrated. The resulting solid was dis-

solved in ∼12ꢀ14 mL of MeOH with gentle heating. EtOAc (75 mL)

was added over 20 min, and then the mixture was stirred for 1 h at room

temperature. The solid was filtered, rinsed with EtOAc, and air-dried to

yield the monotosylate salt of 6 (5.59 g) as a white, crystalline powder.

N-({trans-3-[3-Chloro-4-(pyrrolidin-1-ylmethyl)phenyl]-

cyclobutyl}methyl)-N-methylacetamide (4). Acetic anhydride

(9.2 mL, 97 mmol) was added to a 0 °C solution of 19 (23.7 g, 81.2

mmol) and Et₃N (22.6 mL, 162 mmol) in CH₂Cl₂(230 mL). After the

mixture was stirred for 30 min, 1 N aqueous NaOH (200 mL) was added

and the phases were separated. The aqueous phase was extracted twice

with EtOAc, and the combined organics were dried over MgSO₄,

filtered, and concentrated to yield 29.2 g of crude product. Chromatog-

raphy using a gradient of 5ꢀ20% MeOH in CH₂Cl₂containing 0.1%

NH₄OH yielded the title compound (27 g, 99%) as a colorless, viscous

oil. This was dissolved in EtOAc (200 mL) and treated with 2 M HCl in

Et₂O (81.2 mL) to afford the mono-HCl salt of the title compound as a

7614

dx.doi.org/10.1021/jm200939b |J. Med. Chem. 2011, 54, 7602–7620

Journal of Medicinal Chemistry

ARTICLE

¹H NMR (CDCl₃) δ 7.71 (d, J = 8.3, 2H), 7.65 (t, J = 7.9, 1H),

7.28ꢀ7.22 (m, 2H), 7.15 (d, J = 7.9, 2H), 6.52 (br s, 1H), 4.28 (d, J = 5.4,

2H), 3.68ꢀ3.37 (m, 2H), 3.33ꢀ3.18 (m, 3H), 2.97ꢀ2.88 (m, 2H),

2.84ꢀ2.57 (m, 4H), 2.32 (s, 3H), 2.27ꢀ1.96 (m, 4H), 1.07 (t, J = 7.3,

3H). ¹³C NMR (CDCl₃) δ 173.5, 161.0 (dd, J_CꢀF= 248.9, 1.5), 147.0

(dd, J_CꢀF= 23.6, 7.4), 142.0, 140.1, 133.3 (d, J_CꢀF= 3.0), 128.8, 125.8,

121.5 (dd, J_CꢀF= 8.8, 3.7), 116.1 (dd, J_CꢀF= 14.7, 1.5), 112.6 (dd, J_CꢀF

= 23.6, 8.8), 96.4 (dd, J_CꢀF= 197.6, 2.2), 53.2, 50.4 (d, J_CꢀF= 2.9), 38.6

(d, J_CꢀF= 24.3), 34.4, 32.4, 22.7, 21.3, 14.7. Anal. Calcd for

3-[(Benzyloxy)methyl]-1-[4-(pyrrolidin-1-ylmethyl)phenyl]-

cyclobutanol (9). n-BuLi (2.5 M in hexane, 3.2 mL, 8.0 mmol) was

slowly added to a ꢀ78 °C solution of 1-(4-bromobenzyl)pyrrolidine

(1.9 g, 7.9 mmol) in THF (13 mL). After 30 min, a ꢀ78 °C solution of 8

(1.0 g, 5.3 mmol) in THF (1 mL) was added via cannula and the resulting

solution was stirred at ꢀ78 °C for 30 min. The reaction was quenched

with saturated aqueous NH₄Cl and diluted with EtOAc. The organics

were dried over MgSO₄, filtered, and concentrated. Chromatography

using a gradient of 2% MeOH/CH₂Cl₂containing 0.1% NH₄OH to

30% MeOH/CH₂Cl₂containing 0.1% NH₄OH afforded the title com-

pound (1.3 g, 71%) as a colorless oil. GCMS analysis indicated a ∼3:1

mixture of isomers, presumed to be cis and trans. ¹H NMR (CDCl₃, 400

MHz) δ 7.28ꢀ7.47 (m, 9H), 4.61 and 4.52 (singlets, 2H total), 3.46ꢀ

3.62 (m, 4H), 2.75ꢀ2.97 (m, 2H), 2.48ꢀ2.53 (m, 4H), 2.22ꢀ2.47 (m,

4H), 1.75ꢀ1.82 (m, 4H). MS (APCI) m/z = 352.4 (M + 1).

C₁₈H₂₄F₂N₂O C₇H₈O₃S: C, 60.71; H, 6.52; N, 5.66; F, 7.68; S, 6.48.

3

Found: C, 60.55; H, 6.40; N, 5.58; F, 7.67; S, 6.68. HRMS m/z, calcd

(M + 1) 323.1929, observed 323.1934.

trans-3-Fluoro-3-[3-fluoro-4-(pyrrolidin-1-ylmethyl)phenyl]-

N-(2-methylpropyl)cyclobutanecarboxamide (7). The free

base of this material was prepared in 85% yield as a waxy orange solid

following the general procedure described for compound 6 and sub-

stituting 24 for 23. The free base 7 (13.8 g, 79.0 mmol) was dissolved in

EtOAc (250 mL) and treated with p-toluenesulfonic acid (15.2 g, 79.9

mmol) in EtOAc (150 mL). The resulting solution was stirred overnight,

and the white precipitate was collected and dried under nitrogen purge

to yield 16.5 g of the monotosylate salt. The salt was dissolved in a

mixture of MeOH (20 mL) and EtOAc (40 mL) by application of heat.

Following filtration through a nylon filter, EtOAc (250 mL) was added

to the filtrate over ∼40 min. After the mixture was stirred for an ad-

ditional hour, the resulting white solid was collected and dried under

nitrogen purge. This recrystallization procedure was repeated twice to

1-(4-{3-[(Benzyloxy)methyl]cyclobutyl}benzyl)pyrrolidine

(10). Triethylsilane (49.0 mL, 308 mmol) and TFA (47 mL, 620 mmol)

were added to 9 (21.6 g, 61.5 mmol) and the resulting biphasic mixture

was heated at 75 °C for 1 h, cooled to room temperature and con-

centrated. This material was used without purification in the next step.

For characterization, a portion of the crude material was dissolved in

EtOAc and washed with 1 N aqueous NaOH, dried over MgSO₄, filtered

and concentrated. Chromatography using EtOAc and 10% MeOH/

EtOAc gave a ∼1:1 mixture of cis and trans isomers of the title com-

1

pound as a colorless oil. H NMR (CDCl₃, 400 MHz) δ 7.14ꢀ7.40

(m, 9H), 4.59 and 4.54 (singlets, 2H total), 3.65 and 3.47 (doublets, J =

7.2 and 6.3, 2H total), 3.60 (singlet, 1H), 3.59 (singlet, 1H), 3.55ꢀ3.61

and 3.36ꢀ3.43 (multiplets, 1H total), 2.45ꢀ2.65 (overlapping multi-

plets, 6H total), 2.20ꢀ2.34 and 1.86ꢀ1.94 (multiplets, 3H total),

1.75ꢀ1.82 (m, 4H). MS (APCI) m/z = 336.4 (M + 1).

1

afford 12.2 g of the tosylate salt of 7 as a solid. H NMR (300 MHz,

CDCl₃) δ 11.36 (br s, 1H), 7.79ꢀ7.74 (m, 3H), 7.32ꢀ7.28 (m, 2H),

7.19 (d, J = 7.9, 2H), 5.78 (br s, 1H), 4.32 (d, J = 5.4, 2H), 3.79ꢀ3.73 (m,

2H), 3.29 (tt, J = 8.7, 8.3, 1H), 3.09 (dd, J = 7.1, 6.3, 2H), 2.93ꢀ2.66 (m,

6H), 2.36 (s, 3H), 2.24ꢀ2.18 (m, 2H), 2.08ꢀ2.02 (m, 2H), 1.76 (sept,

J = 6.7, 1H), 0.89 (d, J = 6.6, 6H). ¹³C NMR (100 MHz, CDCl₃) δ 174.0,

161.3 (d, J_CꢀF= 248.0), 147.4 (dd, J_CꢀF= 24.1, 7.5), 142.5, 140.3, 133.3

{3-[4-(Pyrrolidin-1-ylmethyl)phenyl]cyclobutyl}methanol

(11). A sample of 20% palladium black (4.0 g) was added to a stirring

solution of crude 10 (61.5 mmol) in EtOH (250 mL). This mixture was

hydrogenated at 45 psi for 1 h, filtered through Celite, and concentrated.

The residue was dissolved in CH₂Cl₂and washed with saturated

aqueous NaHCO₃. The aqueous layer was extracted with 3:1 CHCl₃/

i-PrOH, and the combined organic layers were dried over MgSO₄,

filtered, and concentrated. Chromatography with a gradient of 2ꢀ20%

MeOH/CH₂Cl₂containing 0.2% NH₄OH provided the desired product

as the TFA salt. This was dissolved in 3:1 CHCl₃/i-PrOH and washed

with aqueous K₂CO₃. Removal of solvent yielded the title compound

(11.9 g, 79%, colorless oil) as a mixture of cis and trans isomers, which

(d, J_CꢀF= 3.0), 128.8, 125.8, 121.5 (dd, J_CꢀF= 9.0, 4.5), 116.1 (d, J_CꢀF

=

13.5), 112.3 (dd, J_CꢀF= 24.1, 9.0), 96.7 (d, J_CꢀF= 196.9), 53.4, 50.0

(d, J_CꢀF= 3.0), 46.9, 38.6 (d, J_CꢀF= 25.6), 32.5, 28.4, 22.7, 21.2, 20.0. Anal.

Calcd for C₂₀H₂₈F₂N₂O C₇H₈O₃S: C, 62.05; H, 6.94; N, 5.36; F, 7.27;

3

S, 6.14. Found: C, 61.85; H, 7.03; N, 5.32; F, 7.21; S, 6.34. HRMS m/z,

calcd (M + 1) 351.2242, observed 351.2236.

{[(3,3-Dimethoxycyclobutyl)methoxy]methyl}benzene

(8a). t-BuOK (1 M in THF, 163 mL, 163 mmol) was added at room

temperature to a solution of (3,3-dimethoxycyclobutyl)methanol⁵⁰

(11.9 g, 81.5 mmol) in THF (100 mL). The mixture was stirred for

30 min. Benzyl bromide (10.2 mL, 85.6 mmol) was added, and the

stirring was continued for an additional 30 min. Water and CH₂Cl₂were

added, and the organic layer was separated, dried over MgSO₄, filtered,

and concentrated. Chromatography using a gradient of 100% hexanes to

5% EtOAc/hexanes furnished the desired product as a colorless oil (13.2

g, 69%). ¹H NMR (CDCl₃, 400 MHz) δ 7.36ꢀ7.25 (m, 5H), 4.52 (s,

2H), 3.48 (d, J = 7.0, 2H), 3.16 (s, 3H), 3.12 (s, 3H), 2.44ꢀ2.26 (m,

3H), 1.90ꢀ1.84 (m, 2H). ¹³C NMR (CDCl₃, 100 MHz) δ 138.8, 128.6,

127.9, 127.8, 101.1, 74.7, 73.2, 48.6, 48.4, 34.9, 24.9. MS (APCI) m/z =

205.3 (M ꢀ OCH₃).

1

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

7.20ꢀ7.24 (m, 2H), 7.09ꢀ7.14 (m, 2H), 5.69 (br s, 1H), 3.76 and 3.75

(singlets, 2H total), 3.24ꢀ3.65 (overlapping doublets and multiplets,

3H total), 2.70ꢀ2.77 (m, 4H), 2.30ꢀ2.43 (m, 2H), 1.73ꢀ1.85 and

2.08ꢀ2.16 (overlapping multiplets, 7H total). MS (APCI) m/z = 246.3

(M + 1).

{cis-3-[4-(Pyrrolidin-1-ylmethyl)phenyl]cyclobutyl}methyl

4-Methylbenzenesulfonate (12a) and {trans-3-[4-(Pyrrolidin-

1-ylmethyl)phenyl]cyclobutyl}methyl 4-Methylbenzenesul-

fonate (12b). p-Toluenesulfonyl chloride (p-TsCl) (49.9 g, 262 mmol)

was added to a 0 °C solution of 11 (53.5 g, 218.4 mmol) and pyridine

(88.3 mL, 1.09 mol) in CH₂Cl₂(250 mL). After 30 min, an additional

portion of p-TsCl (10.0 g, 52.5 mmol) was added, and the mixture was

stirred overnight at room temperature. The mixture was concentrated to

near dryness. Water, CH₂Cl₂, and solid NaHCO₃and K₂CO₃were added,

and the mixture was stirred for 90 min to consume excess p-TsCl. The

organic layer was separated, dried over MgSO₄, filtered, and concen-

trated. The resulting oil (110 g) was purified by chromatography using a

gradient of 2ꢀ5% MeOH in CH₂Cl₂to provide a 1:2 mixture of cis and

trans isomers of the product (57 g, 66%) as a brown gum. The isomers

were separatedby preparative HPLC (Chiralpak AD column, elutingwith

3-[(Benzyloxy)methyl]cyclobutanone (8). p-Toluenesulfonic

acid monohydrate (2.1 g, 11 mmol) was added to a solution of 8a (13.0

g, 55.1 mmol) in acetone/water (3:1, 200 mL). After heating for 45 min

at 65 °C, the mixture was cooled to room temperature and concentrated.

The residue was dissolved in EtOAc, washed with 5% aqueous NaOH,

dried over MgSO₄, filtered, and concentrated. Chromatography using

10% EtOAc/hexanes yielded the title compound as a colorless oil

1

(6.0 g, 58%). H NMR (CDCl₃, 400 MHz) δ 7.28ꢀ7.38 (m, 5H), 4.56

(s, 2H), 3.59 (d, J = 6.3, 2H), 3.88ꢀ3.18 (m, 2H), 2.84ꢀ2.92 (m, 2H),

2.64ꢀ2.75 (m, 1H). ¹³C NMR (CDCl₃, 100 MHz) δ 207.6, 138.3,

128.7, 127.9, 127.8, 73.4, 73.1, 50.2, 23.9. GCMS m/z = 190 (M⁺).

7615

dx.doi.org/10.1021/jm200939b |J. Med. Chem. 2011, 54, 7602–7620

Journal of Medicinal Chemistry

ARTICLE

85/15 heptane/EtOH with 0.1% diethylamine modifier) to afford the

following compounds.

were added methylamine (2.0 M in THF, 167 mL, 334 mmol) and T3P

(50 wt % solution in EtOAc, 128 mL, 201 mmol). The resulting reaction

mixture was stirred at room temperature for 1 h. Then 1 N aqueous

NaOH (575 mL) and EtOAc (400 mL) were added and the layers were

separated. The aqueous layer was subjected to EtOAc extraction (2 ꢁ

500 mL), and the combined organic layers were dried over MgSO₄,

filtered, and concentrated. Chromatography with a gradient of 5ꢀ15%

MeOH in CH₂Cl₂containing 0.25% NH₄OH provided the title

12a (cis isomer): colorless gum. ¹H NMR (CDCl₃, 400 MHz) δ 7.79

(d, J = 8.3, 2H), 7.32 (d, J = 7.9, 2H), 7.24 (d, J = 7.9, 2H), 7.08 (d, J = 7.9,

2H), 3.99 (d, J = 6.2, 2H), 3.57 (s, 2H), 3.31ꢀ3.40 (m, 1H), 2.53ꢀ2.63

(m, 1H), 2.46ꢀ2.50 (m, 4H), 2.42 (s, 3H), 2.36ꢀ2.41 (m, 2H),

1.78ꢀ1.87 (m, 2H), 1.72ꢀ1.77 (m, 4H). ¹³C NMR (CDCl₃, 100 MHz)

δ 144.4, 143.1, 136.9, 132.8, 129.6, 128.5, 127.5, 125.8, 73.6, 60.0, 53.8,

35.3, 31.9, 29.7, 23.1, 21.3. MS (APCI) m/z = 400.3 (M + 1).

12b (trans isomer): colorless gum. ¹H NMR (CDCl₃, 400 MHz) δ

7.84 (d, J = 8.1, 2H), 7.37 (d, J = 7.8, 2H), 7.26 (d, J = 8.2, 2H), 7.13

(d, J = 8.2, 2H), 4.19 (d, J = 7.4, 2H), 3.59 (s, 2H), 3.49ꢀ3.57 (m, 1H),

2.57ꢀ2.65 (m, 1H), 2.49ꢀ2.52 (m, 4H), 2.47 (s, 3H), 2.23ꢀ2.31 (m,

2H), 2.14ꢀ2.21 (m, 2H), 1.76ꢀ1.80 (m, 4H). ¹³C NMR (CDCl₃, 100

MHz) δ 144.7, 143.9, 137.0, 133.2, 129.8, 128.9, 127.9, 126.1, 73.6, 60.3,

54.1, 35.9, 30.6, 29.9, 23.4, 21.6. MS (APCI) m/z = 400.3 (M + 1).

{trans-3-[4-(Pyrrolidin-1-ylmethyl)phenyl]cyclobutyl}-

methanol (13). Magnesium turnings (4.87 g, 201 mmol) and 12b (8.0 g,

20 mmol) in MeOH (160 mL) were stirred for 4 h to give a milky mixture.

Aqueous NaOH (15%) was added, and the mixture was extracted with 3:1

CHCl₃/i-PrOH. The organic phase was dried over MgSO₄, filtered, and

concentrated to yield the title compound (3.6 g, 73%) as a colorless oil. This

1

compound (25.9 g, 48%). H NMR (CDCl₃, 400 MHz) δ 7.46ꢀ7.48

(m, 2H), 7.34 (dd, J = 7.9, 1.7, 1H), 6.03ꢀ6.11 (br m, 1H), 5.77 (br s,

1H), 3.76 (s, 2H), 2.85 (d, J = 4.9, 3H), 2.73ꢀ2.86 (m, 3H), 2.56ꢀ2.63

(m, 4H), 2.47ꢀ2.54 (m, 2H), 1.76ꢀ1.84 (m, 4H). ¹³C NMR (CDCl₃,

100 MHz) δ 177.3, 145.8, 135.1, 133.8, 130.6, 126.1, 123.3, 74.1, 56.5,

54.1, 40.8, 33.1, 26.6, 23.5. MS (APCI) m/z = 323.1 (M + 1).

trans-3-[3-Chloro-4-(pyrrolidin-1-ylmethyl)phenyl]-N-

methylcyclobutanecarboxamide (18). TFA (48 mL) was added

to a solution of 16 (10.0 g, 31.4 mmol) in DCE (202 mL), and the re-

sulting mixture was heated at 75 °C for 18 h and concentrated. The

crude TFA salt of 3-[3-chloro-4-(pyrrolidin-1-ylmethyl)phenyl]-N-

methylcyclobut-2-ene-1-carboxamide (17) thus obtained was redis-

solved in EtOH (130 mL) and hydrogenated (45 psi) at 60 °C in the

presence of chlorotris(triphenylphosphine)rhodium(I) (Wilkinson’s

catalyst, 1.5 g). After 2 h, the mixture was cooled and concentrated.

The residue was redissolved in 1 N aqueous HCl (100 mL) and washed

twice with EtOAc (2 ꢁ 100 mL). The aqueous layer was then made basic

with 1 N aqueous NaOH (100 mL) and extracted with EtOAc (2 ꢁ

500 mL). The combined organic phases were dried over MgSO₄,

filtered, and concentrated. Chromatography with a gradient of 5ꢀ15%

MeOH in CH₂Cl₂containing 0.25% NH₄OH gave a mixture of cis and

trans isomers of the title compound (4.6 g, 48% yield). These isomers

were separated by preparative chromatography on a Chiralcel OD

(10 cm ꢁ 50 cm) column using heptane/i-PrOH (90/10) to yield

the pure trans (3.6 g) and cis (0.52 g) isomers. trans-3-[3-Chloro-4-

(pyrrolidin-1-ylmethyl)phenyl]-N-methylcyclobutanecarboxamide (18):

¹H NMR (CDCl₃, 400 MHz) δ 7.39 (d, J = 7.8, 1H), 7.22 (d, J = 1.6, 1H),

7.09 (dd, J = 7.8, 1.2, 1H), 5.58 (br s, 1H), 3.69ꢀ3.77 (singlet at 3.72

(2H) overlapping multiplet (1H), 3H total), 2.94ꢀ3.00 (m, 1H), 2.86

(d, J = 5.1, 3H), 2.65ꢀ2.72 (m, 2H), 2.55ꢀ2.60 (m, 4H), 2.31ꢀ2.38

(m, 2H), 1.75ꢀ1.83 (m, 4H). ¹³C NMR (CDCl₃, 100 MHz) δ 175.8,

145.7, 134.4, 133.8, 130.6, 127.1, 124.7, 56.7, 54.2, 36.2, 36.1, 31.9, 26.4,

23.5. MS (APCI) m/z = 307.4 (M + 1).

1

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

MHz) δ 7.28 (d, J = 8.0, 2H), 7.19 (d, J = 8.0, 2H), 3.73 (d, J = 7.4, 2H),

3.60 (s, 2H), 3.52ꢀ3.58 (m, 1H), 3.47 (br s, 1H), 2.49ꢀ2.56 (m, 4H),

2.41ꢀ2.49 (m, 1H), 2.16ꢀ2.29 (m, 4H), 1.70ꢀ1.81 (m, 4H). ¹³C NMR

(CDCl₃, 100 MHz) δ 144.9, 136.2, 128.9, 126.1, 66.2, 60.2, 53.9, 36.3, 32.8,

31.0, 23.2. MS (APCI) m/z = 246.4 (M + 1).

(4-Bromo-2-chlorophenyl)(pyrrolidin-1-yl)methanone (14).

T3P (48.6 g, 153 mmol, 50 wt % in EtOAc) was added to a mixture of

4-bromo-2-chlorobenzoic acid (30 g, 130 mmol), triethylamine (25.8 g,

255 mmol), and pyrrolidine (18 g, 255 mmol) in EtOAc (1.5 L). After 1 h,

1 N aqueous NaOH (200 mL) was added and the mixture was stirred for

10 min. The layers were separated, and the aqueous phase was extracted

with EtOAc (2 ꢁ 500 mL). The combined organic extractswere driedover

MgSO₄, filtered, and concentrated to provide a viscous oil. Chromatogra-

phy using a gradient of 50ꢀ80% EtOAc in hexanes yielded the title com-

pound as a light yellow, viscous oil (35.4 g, 96%). ¹H NMR (CDCl₃, 400

MHz) δ 7.56 (d, J = 1.7, 1H), 7.43 (dd, J = 8.3, 1.7, 1H), 7.17 (d, J = 8.3,

1H), 3.63 (br app t, J = 6.6, 2H), 3.17 (br app t, J = 6.6, 2H), 2.00ꢀ1.84 (m,

4H). ¹³C NMR (CDCl_3,100 MHz) δ 166.0, 136.6, 132.6, 131.3, 130.7,

128.9, 123.3, 48.0, 45.8, 26.1, 24.7. MS (APCI) m/z = 287.9 (M + 1).

1-(4-Bromo-2-chlorobenzyl)pyrrolidine (15). BoraneꢀTHF

complex (1.0 M in THF, 367 mL, 367 mmol) was added dropwise to a

solution of 14 (35.3 g, 122 mmol) in THF (120 mL), and the resulting

mixture was stirred at room temperature for 21 h. The reaction was

quenched with MeOH (120 mL), and the mixture was heated to 80 °C

for 18 h. The mixture was cooled to room temperature and concentrated

and the resulting residue was purified by chromatography using 50%

EtOAc in hexanes to yield the title compound as a colorless, viscous oil

(24.4 g, 74%). ¹H NMR (CDCl_3,400 MHz) δ 7.47 (br s, 1H),

7.38ꢀ7.33 (m, 2H), 3.66 (s, 2H), 2.58ꢀ2.54 (m, 4H), 1.80ꢀ1.76 (m, 4H).

¹³C NMR (CDCl₃,100 MHz) δ 136.4, 134.8, 132.0, 131.9, 130.0, 120.7,

56.6, 54.4, 23.8. MS (APCI) m/z = 275.8 (M + 1).

1-{trans-3-[3-Chloro-4-(pyrrolidin-1-ylmethyl)phenyl]-

cyclobutyl}-N-methylmethanamine (19). BH₃ꢀTHF complex

(1.0 M inTHF, 319 mL, 319 mmol) was added to a solution of 18 (32.5 g,

106.2 mmol) in THF (516 mL), and the resulting solution was heated

at 70 °C for 16 h. MeOH was added to quench the excess borane, and the

heating was continued for another 16 h. The reaction mixture was cooled

to room temperature and concentrated to yield a viscous oil which was

redissolved in EtOAc (500 mL) and extracted into 6 N HCl (260 mL).

The acidic aqueous phase was basified with 15% aqueous NaOH (500 mL)

and extracted with EtOAc (2 ꢁ 750 mL). The extracts were dried over

MgSO₄, filtered, and concentrated to yield 30.3 g of crude product.

Chromatography with a gradient of 5ꢀ100% MeOH in CH₂Cl₂

containing 0.25% NH₄OH afforded the title compound (23.7 g, 77%

1

yield) as a colorless viscous oil. H NMR (CDCl₃, 400 MHz) δ 7.39

3-[3-Chloro-4-(pyrrolidin-1-ylmethyl)phenyl]-3-hydroxy-

N-methylcyclobutanecarboxamide (16). n-Butyllithium (2.5 M

in hexanes, 135 mL, 0.335 mol) was added down the reaction flask walls

over 25 min to a ꢀ78 °C solution of 15 (92.0 g, 0.335 mol) in THF

(400 mL). After the mixturewasstirred atꢀ78 °C for 1 h, aꢀ78 °Csolution

of 3-oxocyclobutanecarboxylic acid (19.0 g, 167.5 mmol) in THF

(400 mL) was cannulated into the reaction mixture over 15 min. The

resulting dark orange solution was slowly warmed to room temperature

over 16 h. To the crude solution of the intermediate 3-[3-chloro-

4-(pyrrolidin-1-ylmethyl)phenyl]-3-hydroxycyclobutanecarboxylic acid

(d, J = 7.9, 1H), 7.23 (d, J = 1.7, 1H), 7.10 (dd, J = 7.9, 1.7, 1H), 3.73 (s, 2H),

3.55 (tt, J = 8.3, 8.2, 1H), 2.79 (d, J = 7.5, 2H), 2.56ꢀ2.61 (m, 4H), 2.48

(s, 3H) overlapping 2.44ꢀ2.51 (m, 1H), 2.21ꢀ2.31 (m, 2H), 2.12ꢀ2.18

(m, 2H), 1.76ꢀ1.83 (m, 4H). ¹³C NMR (CDCl₃, 100 MHz) δ 146.5,

134.1, 133.7, 130.5, 127.2, 124.7, 57.3, 56.7, 54.2, 36.6, 36.2, 32.4, 30.9,

23.5. MS (APCI) m/z = 293.2 (M + 1).

N-({trans-3-[3-Chloro-4-(pyrrolidin-1-ylmethyl)phenyl]-

cyclobutyl}methyl)-N-methylacetamide (4). Acetic anhydride

(9.2 mL, 97 mmol) was added to a 0 °C solution of 19 (23.7 g, 81.2 mmol)

7616

dx.doi.org/10.1021/jm200939b |J. Med. Chem. 2011, 54, 7602–7620

Journal of Medicinal Chemistry

ARTICLE

and Et₃N (22.6 mL, 162 mmol) in CH₂Cl₂(230 mL). After the mixture

was stirred for 30 min, 1 N aqueous NaOH (200 mL) was added and the

phases were separated. The aqueous phase was extracted twice with

EtOAc, and the combined organics were dried over MgSO₄, filtered, and

concentrated to yield 29.2 g of crude product. Chromatography using a

gradient of 5ꢀ20% MeOH in CH₂Cl₂containing 0.1% NH₄OH yielded

the title compound (27 g, 99%) as a colorless, viscous oil. This was

dissolved in EtOAc (200 mL) and treated with 2 M HCl in Et₂O

(81.2 mL) to afford the mono-HCl salt of the title compound as a white

solid (30.6 g, 99%). ¹H NMR (MeOH-d₄, 400 MHz) ∼1.5:1 mixture of

rotamers, δ 7.70ꢀ7.73 (m, 1H), 7.50 and 7.46 (2 d, J = 1.7, 1H total),

7.40 and 7.36 (2 dd, J = 7.9, 1.5, 1H total), 4.57 (s, 2H), 3.70ꢀ3.82

(m, 3H), 3.53ꢀ3.60 (m, 2H), 3.24 (m, 2H), 3.24 and 3.14 (2 s, 3H total),

2.72ꢀ2.86 (m, 1H), 2.17ꢀ2.38 (m, 9H), 2.02ꢀ2.12 (m, 2H). ¹³C NMR

(MeOH-d₄, 100 MHz) δ 175.7, 175.4, 151.6, 151.5, 136.2, 134.3, 129.4,

129.3, 127.6, 127.5, 127.4, 127.3, 57.1, 55.9, 55.2, 54.7, 38.3, 37.2, 37.0,

36.0, 32.8, 32.6, 31.0, 30.4, 24.0, 20.2, 19.8. HRMS m/z, calcd (M + 1)

335.1884, observed 335.1880.

1-(4-Bromo-2-fluorobenzyl)pyrrolidine (20). 4-Bromo-2-

fluorobenzyl bromide (375 g, 1.40 mol) was added in six portions to a

10 °C solution of pyrrolidine (363 g, 426 mL, 5.10 mol) in CH₃CN

(2.75 L) while maintaining the temperature below 20 °C. After the mixture

was stirred at room temperature for 4 h, the mixture was concentrated, water

(500 mL) and saturated aqueous Na₂CO₃(2 L) were added, and the

mixture was extracted with CH₂Cl₂(3 ꢁ 700 mL). The extracts were dried

over Na₂SO₄, filtered, and concentrated, and the resulting light yellow oil

was purified by vacuum distillation (∼1 mm, bp 125 °C) to yield the title

compound (324.5 g, 90%) as a colorless oil. ¹H NMR (400 MHz, CDCl₃) δ

7.19ꢀ7.31 (m, 3H), 3.60 (d, J = 1.7, 2H), 2.48ꢀ2.52 (m, 4H), 1.74ꢀ1.77

(m, 4H). ¹³C NMR (CDCl₃, 100 MHz) δ 160.8 (d, J_CꢀF= 250), 132.3 (d,

(t, J = 7.9, 1H), 7.18ꢀ7.24 (m, 2H), 6.10 (br s, 1H), 3.79 (s, 2H), 3.12

(t, J = 6.4, 2H), 2.77ꢀ2.83 (m, 3H), 2.62ꢀ2.70 (m, 4H), 2.45ꢀ2.55

(m, 2H), 1.75ꢀ1.86 (m, 4H), 0.87ꢀ0.99 [d at 0.92 (J = 6.6, 6H)

overlapping m (1H)]. ¹³C NMR (CDCl₃,100 MHz) δ 176.8, 161.0 (d,

J_CꢀF= 245.9), 147.6 (d, J_CꢀF= 7.9), 131.7 (d, J_CꢀF= 4.4), 122.5 (d,

J_CꢀF= 15.4), 120.6 (d, J_CꢀF= 4.7), 112.2 (d, J_CꢀF= 23.5), 74.1 (d, J_CꢀF

= 1.5), 53.6, 51.9 (br s), 47.2, 41.0, 33.3, 28.5, 23.3, 20.0. MS (APCI) m/z

= 349.4 (M + 1).

H₃Receptor Binding Protocol. Cell paste was homogenized in

50 mM Tris-HCl buffer (pH 7.4 at 4 °C) containing 2.0 mM MgCl₂

using a Polytron at medium speed and spun in a centrifuge at 40000g for

10 min. The supernatant was discarded and the pellet suspended in

50 mM Tris-HCl buffer (pH 7.4 at 25 °C) containing 2 mM MgCl₂.

Incubations were initiated by the addition of tissue homogenate to 96-

well plates containing test drugs and radioligand (1.0 nM ³H-N-α-

methylhistamine). Nonspecific binding was determined by radioligand

binding in the presence of a saturating concentration of ciproxifan (10 μM).

After a 60 min incubation at room temperature, assay samples were

rapidly filtered through Whatman GF/B filters that had been presoaked

and dried in 0.5% polyethylenimine and rinsed with ice-cold 50 mM

3

Tris-HCl buffer (pH 7.7 at 4 °C). Membrane-bound H-N-α-methyl-

histamine levels were determined by liquid scintillation counting of

the filters in BetaScint on a Betaplate scintillation counter. The IC₅₀

(concentration at which 50% inhibition of specific binding occurs) was

calculated by linear regression of the concentrationꢀresponse data (logit

transformation ofthepercentinhibitionversuslogoftheconcentration).K_i

values were calculated based on the ChengꢀPrusoff equation, K_i= IC₅₀/

(1 + (L/K_d)), where L is the concentration of the radioligand used in the

experiment and the K_dis the dissociation constant for the radioligand

(determined previously by saturation analysis). Geometric mean values

are expressed as pK_i( SEM.

J_CꢀF= 6.0), 127.1 (d, J_CꢀF= 3.7), 125.3 (d, J_CꢀF= 14.7), 120.6 (d, J_CꢀF

=

9.6), 118.7 (d, J_CꢀF= 25.7), 53.9, 52.2 (d, J_CꢀF= 1.5), 23.4. MS (APCI)

m/z = 258.5 (M + 1).

H₃cAMP Assay Protocol. Cell Preparation. HEK 293 cells stably

expressing the human or rat histamine H₃receptor and cAMP response

element β-lactamase reporter gene were maintained in minimal essential

media supplemented with 10% (v/v) fetal bovine serum, 2 mM gluta-

mine, 0.1 mM nonessential amino acids, 1 mM sodium pyruvate, 25 mM

HEPES, 300 μg/mL Geneticin, and 50 μg/mL Zeocin at 37 °C and 5%

CO₂. Cells were harvested at 70ꢀ80% confluence and washed. A

suspension of cells was added to each well of a 384-well, black clear-

bottom poly-^D-lysine coated plate such that each well received 25 000

cells (in 75 μL). Plates were incubated at 37 °C/5% CO₂overnight.

cAMP Assay. Following overnight incubation of the assay plate in a

37 °C/5% CO₂incubator, 12.5 μL of compound dilutions and max-

imum and minimum controls were added to the assay plate 10 min prior

to addition of 12.5 μL of forskolin or 12.5 μL of forskolin/imetit. An

amount of 25 μL of phosphate-buffered saline was added to the blank

wells. Plates were incubated for 4.5 h in a 37 °C/5% CO₂incubator.

Following the 4.5 h incubation, 50 μL of MEM + 10 μL of dye solution

[loading dye solution was prepared at 6ꢁ final concentration: 6 μL of

solution A (1 mM CCF4-AM in dry dimethylsulfoxide) was added to

60 μL of solution B (100 mg/mL Pluronic-F127 surfactant in dimethyl-

sulfoxide and 0.1% acetic acid); the resulting solution was vortexed and

added to 940 μL of solution C (24% w/w PEG 400, 18% TR-40 by

volume in water)] was added to each well and the assay plates were

incubated for 90ꢀ120 min at room temperature and kept in the dark.

The assay plates were then read using a Tecan SpectraFluor plate reader,

with an excitation wavelength of 405 nm and emission wavelengths at

450 and 530 nm.

N-Ethyl-3-[3-fluoro-4-(pyrrolidin-1-ylmethyl)phenyl]-

3-hydroxy-N-methylcyclobutanecarboxamide (21). This ma-

terial was prepared as a waxy orange solid in 60% yield following the

general procedure described for preparation of compound 16, by using

20 and substituting ethylmethylamine for methylamine. ¹HNMR(300MHz,

CDCl₃) ∼1:1 mixture of rotamers, δ 7.38 (t, J = 7.7, 1H), 7.18ꢀ7.25 (m,

2H), 5.10 and 4.84 (2 overlapping br s, 1H total), 3.68 (s, 2H), 3.48 and

3.33 (2 q, J = 7.2, 2H total), 3.13ꢀ3.23 (m, 1H), 3.00 and 2.97 (2 s, 3H

total), 2.80ꢀ2.86 (m, 2H), 2.53ꢀ2.60 (m, 6H), 1.76ꢀ1.80 (m, 4H),

1.12ꢀ1.19 (m, 3H). ¹³C NMR (CDCl₃, 100 MHz) ∼1:1 mixture of

rotamers, δ 175.9, 175.4, 162.3, 159.8, 146.9, 146.8, 131.3, 131.2, 124.6,

124.4, 120.4, 120.3, 112.3, 112.0, 73.9, 73.9, 73.7, 73.7, 53.9, 52.5, 52.4,

44.7, 43.0, 41.0, 40.8, 34.7, 33.5, 28.9, 28.3, 23.4, 14.0, 12.2. MS (APCI)

m/z = 335.5 (M + 1).

cis-N-Ethyl-3-[3-fluoro-4-(pyrrolidin-1-ylmethyl)phenyl]-

3-hydroxycyclobutanecarboxamide (23). This material was pre-

pared as a waxy orange solid in 49% yield from 20 following the general

procedure described for compound 16 and substituting ethylamine for

1

methylamine. H NMR (400 MHz, CDCl₃) δ 7.36 (t, J = 7.9, 1H),

7.14ꢀ7.22 (m, 2H), 6.01 (br s, 1H), 5.77 (br s, 1H), 3.66 (d, J = 1.3, 2H),

3.28ꢀ3.35 (m, 2H), 2.71ꢀ2.82 (m, 3H), 2.47ꢀ2.55 (m, 6H),

1.74ꢀ1.79 (m, 4H), 1.15 (t, J = 7.3, 3H). ¹³C NMR (CDCl₃,100 MHz)

δ 176.5, 161.0 (d, J_CꢀF= 245.9), 146.8 (d, J_CꢀF= 7.4), 131.3 (d, J_CꢀF

5.2), 124.1 (d, J_CꢀF= 15.5), 120.3 (d, J_CꢀF= 2.9), 112.0 (d, J_CꢀF

=

22.8), 74.0 (d, J_CꢀF= 1.5), 53.8, 52.3 (d, J_CꢀF= 1.5), 40.9, 34.7, 33.2,

23.4, 14.7. MS (APCI) m/z = 321.3 (M + 1).

Data Analysis. Unconstrained sigmoidal curve fitting was used to

estimate an IC₅₀using GraphPad Prism. The K_ivalues were calculated

from the IC₅₀according to the ChengꢀPrusoff equation K_i= (IC₅₀)/

(1 + ([C]/EC₅₀)), where [C] is the concentration of the agonist (1 nM

final imetit) used in the experiment and the EC₅₀value is the EC₅₀for

the imetit determined in each experiment.

cis-3-[3-Fluoro-4-(pyrrolidin-1-ylmethyl)phenyl]-3-hydroxy-

N-(2-methylpropyl)cyclobutanecarboxamide (24). This ma-

terial was prepared in 63% yield as a waxy orange solid from 20 following

the general procedure described for compound 16 and substituting

1

isobutylamine for methylamine. H NMR (400 MHz, CDCl₃) δ 7.46

7617

dx.doi.org/10.1021/jm200939b |J. Med. Chem. 2011, 54, 7602–7620

Journal of Medicinal Chemistry

ARTICLE

Phospholipidosis Assay Protocol. Phospholipidosis Assay.

Transformed human liver epithelial cells (THLE) were purchased from

American Type Culture Collection. Cells were kept in culture in bronchial

epithelial growth media (Lonza) supplemented with 5 ng/mL human

epidermal growth factor and 70 ng/mL phosphoethanolamine. All cell

culture flasks and plates were precoated with bronchial epithelial basal

medium supplemented with 0.01 mg/mL human fibronectin, 0.01 mg/mL

BSA and 0.03 mg/mL collagen. Cells were maintained at 37 °C under

3.5% CO₂. Cells in logarithmic growth phase were seeded at a density of

1 ꢁ 10⁴per well in 96-well microplates. Test articles and the positive

control, amiodarone (Sigma), were diluted in DMSO to 30 mM. Serial

dilutions of all 30 mM stock concentrations were carried out in micro-

plates, and 2 μL of each concentration was transferred from the compound

dilution plate to a new microplate. A fluorophore-labeled phospholipid,

N-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)-1,2-dihexadecanoyl-sn-glycero-3-

phosphoethanolamine, triethylammonium salt (NBD-PE Invitrogen,

CA, U.S.) was diluted with ethanol (1 mL) and sonicated for 10 min.

The dye was then transferred into 250 mL of THLE medium and

sonicated for an additional 30 min. The dye-supplemented medium was

then filtered through a Millipore 22 μm filter flask, and 200 μL aliquots

were added to each well of the intermediate plates containing 2 μL of the

test article and control dilutions to achieve final concentrations of 300,

100, 30, 10, 3, and 1 μM test article or control.

At 24 hours after cell plating, the medium in each microplate was

removed and compound dilutions in NBD-PE supplemented medium

were added. Plates were returned to the incubator under the same

conditions. At 24 hours after dosing, the medium was removed and cells

were washed twice with 100 μL of PBS. Following the wash, nucleic acid

stain Hoechst 33342, diluted 1:1000 in PBS, was added and plates were

incubated for 30 min at room temperature. Following incubation, cells

were washed twice with 100 μL of PBS. A ﬁnal volume of 100 μL of

PBS was added, and plates were covered and left overnight at room

temperature. The microplates were scanned in a Cellomics ArrayScan

for quantitation of uptake of both Hoechst 33342, to identify cell

nuclei, and NBD-PE, to determine phospholipid accumulation per cell.

Compounds were classiﬁed as positive at the lowest dose for which the

induction of phospholipid accumulation was 3-fold over background

(Table 10).

Safety Studies. In Vivo Toleration Studies. In vivo toleration studies

up to 2 weeks in duration were completed in SpragueꢀDawley rats and

beagle dogs. Drug formulations were prepared in 0.5% methylcellulose and

administered at a dosage volume of 10 mL/kg in rats and 1 mL/kg in dogs.

Monitored end points in these studies included clinical signs (daily), food

intake (daily), body weight (daily), clinical pathology (hematology, serum

chemistry, urinalysis; end of study), vital signs (heart rate, respiratory rate;

dogs, predose and postdose at about T_maxat end of study), electrocardio-

gram (dogs, predose and postdose at about T_maxat end of study), plasma

drug levels (day 1 and end of study; 4ꢀ5 time points on each sampling day),

and gross/histopathology (end of dosing phase in repeat dose studies).

All procedures performed on these animals were in accordance with

regulations and established guidelines and were reviewed and approved

by an Institutional Animal Care and Use Committee or through an

ethical review process. Convulsions in dogs were either self-limiting or

required intervention with a standard anticonvulsant. Exposure and

estimated safety margins derived from ETS in rats (14-day) and in dogs

(single dose and 7-day) and are summarized in Table 11.

’ AUTHOR INFORMATION

Corresponding Author

*Phone: 860-715-4059. Fax: 860-686-6052. E-mail: travis.t.

wager@pﬁzer.com.

Author Contributions

^†T.T.W., B.A.P., and A.W.S. contributed to the writing of this

manuscript.

’ ACKNOWLEDGMENT

The authors thank Joe Brady, Christopher O’Donnell, and

Anabella Villalobos for helpful discussions; Ana Cobani, Elaine

Greer, Stephen Anderson, and Margaret Landis for salt form and

formulation evaluations; Daniel Virtue and Robert DePianta, Jr.,

Benjamin Hritzko, Chris Wood, Christopher Foti, Justin Ringl-

ing, Chris Caldwel, Marina Shalaeva, and Randy Smith for

compound puriﬁcations, analytical assessments, and property

measurements; James Candee and Jon Bordner for X-ray powder

and single crystal crystallography support; Chester J. Siok, Darcy

G. Izzarelli, Elizabeth E. Rubitski, David E. Johnson, Francis J.

Sweeney, Holly D. Soares, Leslie K. Chambers, Bernard Fermini,

Michelle A. Vanase-Frawley, Mihaly Hajos, Patricia A. Seymour,

Weldon E. Horner, and William E. Hoﬀmann for providing in vitro

selectivity and in vivo and biomarker data for the team; Bishop

Wlodecki, Mark Elliot, Stanton McHardy, and Vinod Parikh for

helpful discussions during the hunt for the best-in-class molecule; and

Katherine Brighty for her insightful comments on this manuscript.

’ ABBREVIATIONS USED

ACh, acetylcholine; AD, Alzheimer’s disease; ADHD, attention

deﬁcit hyperactivity disorder; ADME, absorption, distribution,

metabolism, and excretion; cAMP, cyclic adenosine monopho-

sphate; CMV, cardiomyocyte vacuoles; CNS, central nervous

system; DA, dopamine; DCE, dichloroethane; DDI, drugꢀdrug

interactions; DMA, dimethylacetamide; DME, 1,2-dimethox-

yethane; DMF, dimethylformamide; ER, eﬄux ratio; EtOAc, eth-

yl acetate; ETS, exploratory toxicology study; GPCR, G-protein-

coupled receptor; HA, histamine; HBD, hydrogen bond donor;

hERG, human ether-a-go-go-related gene; HLA, human

lymphocyte assay; HLM, human liver microsome; HTS, high-

throughput screening; IVMN, in vitro micronucleus; MDCK,

MadinꢀDarby canine kidney; MDR, multidrug resistance;

MED, minimally eﬀective dose; MW, molecular weight; nAChR,

nicotinic acetylcholine receptor; NaOH, sodium hydroxide; NE,

norepinephrine; PCEC, projected clinically eﬀective concentra-

tion; P-gp, P-glycoprotein; PK, pharmacokinetic; PL, phospholi-

pidosis; RAMH, R-α-methylhistamine; RLM, rat liver microsome;

rt, room temperature; SAR, structureꢀactivity relationship; sc,

subcutaneous; t-BuOK, potassium tert-butoxide; TEA, triethyla-

mine; T3P, propylphosphonic anhydride; TEM, transmission

electron microscope; TFA, triﬂuoroacetic acid; THF, tetrahydro-

furan; THLE, transformed human liver epithelial; TI, therapeutic

index; TPSA, topological polar surface area

’ REFERENCES

’ ASSOCIATED CONTENT

(1) Ash, A. S. F.; Schild, H. O. Receptors mediating some actions of

histamine. Br. J. Pharmacol. Chemother. 1966, 27, 427–439.

(2) Black, J. W.; Duncan, W. A. M.; Durant, C. J.; Ganellin, C. R.;

Parsons, E. M. Deﬁnition and antagonism of histamine H₂-receptors.

Nature 1972, 236, 385–390.

S

Supporting Information. Cerep proﬁle of lead com-

b

pounds and spectral data for intermediates and ﬁnal compounds.

This material is available free of charge via the Internet at http://

pubs.acs.org.

7618

dx.doi.org/10.1021/jm200939b |J. Med. Chem. 2011, 54, 7602–7620

Journal of Medicinal Chemistry

ARTICLE

(3) Celanire, S.; Wijtmans, M.; Talaga, P.; Leurs, R.; de Esch, I. J. P.

Keynote review: histamine H₃receptor antagonists reach out for the

clinic. Drug Discovery Today 2005, 10, 1613–1627.

(4) yaz_ewska, D.; Kieꢀc-Kononowicz, K. Recent advances in hista-

mine H₃receptor antagonists/inverse agonists. Expert Opin. Ther. Pat.

2010, 20, 1147–1169.

(5) Passani, M. B.; Lin, J.-S.; Hancock, A.; Crochet, S.; Blandina, P.

The histamine H₃receptor as a novel therapeutic target for

cognitive and sleep disorders. Trends Pharmacol. Sci. 2004, 25,

618–625.

antiobesity histamine H₃receptor antagonists. Basic Clin. Pharmacol.

Toxicol. 2004, 95, 144–152.

(23) Morini, G.; Comini, M.; Rivara, M.; Rivara, S.; Lorenzi, S.;

Bordi, F.; Mor, M.; Flammini, L.; Bertoni, S.; Ballabeni, V.; Barocelli, E.;

Plazzi, P. V. Dibasic non-imidazole histamine H₃receptor antagonists

with a rigid biphenyl scaﬀold. Bioorg. Med. Chem. Lett. 2006, 16, 4063–4067.

(24) Clark, G. R.; Squire, C. J.; Gray, E. J.; Leupin, W.; Neidle, S.

Designer DNA-binding drugs: the crystal structure of a meta-hydroxy

analogue of Hoechst 33258 bound to d(CGCGAATTCGCG)₂. Nucleic

Acids Res. 1996, 24, 4882–4889.

(6) Barbier, A. J.; Bradbury, M. J. Histaminergic control of

sleepꢀwake cycles: recent therapeutic advances for sleep and wake

disorders. CNS Neurol. Disord.: Drug Targets 2007, 6, 31–43.

(7) Levander, S.; Haegermark, O.; Staahle, M. Peripheral antihista-

mine and central sedative eﬀects of three H₁-receptor antagonists. Eur. J.

Clin. Pharmacol. 1985, 28, 523–529.

(8) Arrang, J. M.; Garbarg, M.; Schwartz, J. C. Autoinhibition of

brain histamine release mediated by a novel class (H₃) of histamine

receptor. Nature 1983, 302, 832–837.

(9) Brown, R. E.; Stevens, D. R.; Haas, H. L. The physiology of brain

histamine. Prog. Neurobiol. (Oxford, U. K.) 2001, 63, 637–672.

(10) Esbenshade, T. A.; Browman, K. E.; Bitner, R. S.; Strakhova, M.;

Cowart, M. D.; Brioni, J. D. The histamine H₃receptor: an attractive

target for the treatment of cognitive disorders. Br. J. Pharmacol. 2008,

154, 1166–1181.

(11) Tzavara, E. T.; Bymaster, F. P.; Overshiner, C. D.; Davis, R. J.;

Perry, K. W.; Wolﬀ, M.; McKinzie, D. L.; Witkin, J. M.; Nomikos, G. G.

Procholinergic and memory enhancing properties of the selective

norepinephrine uptake inhibitor atomoxetine. Mol. Psychiatry 2006,

11, 187–195.

(25) Miller, V. P.; Stresser, D. M.; Blanchard, A. P.; Turner, S.;

Crespi, C. L. Fluorometric high-throughput screening for inhibitors of

cytochrome P450. Ann. N.Y. Acad. Sci. 2000, 919, 26–32.

(26) Aubrecht, J.; Osowski, J. J.; Persaud, P.; Cheung, J. R.; Ackerman,

J.; Lopes, S. H.; Ku, W. W. Bioluminescent Salmonella reverse mutation

assay: a screen for detecting mutagenicity with high throughput attributes.

Mutagenesis 2007, 22, 335–342.

(27) Homiski, M. L.; Muehlbauer, P. A.; Dobo, K. L.; Schuler, M. J.;

Aubrecht, J. Concordance analysis of an in vitro micronucleus screening

assay and the regulatory chromosome aberration assay using pharma-

ceutical drug candidates. Environ. Mol. Mutagen. 2010, 51, 39–47.

(28) Feng, B.; Mills, J. B.; Davidson, R. E.; Mireles, R. J.; Janiszewski,

J. S.; Troutman, M. D.; de Morais, S. M. In vitro P-glycoprotein assays to

predict the in vivo interactions of P-glycoprotein with drugs in the

central nervous system. Drug Metab. Dispos. 2008, 36, 268–275.

(29) Fox, G. B.; Pan, J. B.; Esbenshade, T. A.; Bitner, R. S.; Nikkel,

A. L.; Miller, T.; Kang, C. H.; Bennani, Y. L.; Black, L. A.; Faghih, R.;

Hancock, A. A.; Decker, M. W. Diﬀerential in vivo eﬀects of H₃receptor

ligands in a new mouse dipsogenia model. Pharmacol., Biochem. Behav.

2002, 72, 741–750.

(12) Horner, W. E.; Johnson, D. E.; Schmidt, A. W.; Rollema, H.

Methylphenidate and atomoxetine increase histamine release in rat

prefrontal cortex. Eur. J. Pharmacol. 2007, 558, 96–97.

(13) FDA. Atomoxetine (Marketed as Strattera) Information, 2005.

http://www.fda.gov/Drugs/DrugSafety/PostmarketDrugSafetyInfor-

mationforPatientsandProviders/ucm107912.htm

(14) Wernicke, J. F.; Kratochvil, C. J. Safety proﬁle of atomoxetine in

the treatment of children and adolescents with ADHD. J. Clin. Psychiatry

2002, 63 (Suppl. 12), 50–55.

(15) Fermini, B.; Fossa, A. A. The impact of drug-induced QT

interval prolongation on drug discovery and development. Nat. Rev.

Drug Discovery 2003, 2, 439–447.

(16) Hosea, N. A.; Collard, W. T.; Cole, S.; Maurer, T. S.; Fang,

R. X.; Jones, H.; Kakar, S. M.; Nakai, Y.; Smith, B. J.; Webster, R.;

Beaumont, K. Prediction of human pharmacokinetics from preclinical

information: comparative accuracy of quantitative prediction approaches.

J. Clin. Pharmacol. 2009, 49, 513–533.

(17) Gao, H.; Yao, L.; Mathieu, H. W.; Zhang, Y.; Maurer, T. S.;

Troutman, M. D.; Scott, D. O.; Ruggeri, R. B.; Lin, J. In silico modeling

of nonspeciﬁc binding to human liver microsomes. Drug Metab. Dispos.

2008, 36, 2130–2135.

(30) McLeod, R. L.; Rizzo, C. A.; West, R. E., Jr.; Aslanian, R.;

McCormick, K.; Bryant, M.; Hsieh, Y.; Korfmacher, W.; Mingo, G. G.;

Varty, L.; Williams, S. M.; Shih, N.-Y.; Egan, R. W.; Hey, J. A.

Pharmacological characterization of the novel histamine H₃-receptor

antagonist N-(3,5-dichlorophenyl)-N⁰-[[4-(1H-imidazol-4-ylmethyl)-

phenyl]-methyl]-urea (SCH 79687). J. Pharmacol. Exp. Ther. 2003, 305,

1037–1044.

(31) Smith, D. A.; van de Waterbeemd, H.; Walker, D. K. Pharma-

cokinetics and Metabolism in Drug Design, 2nd ed.; Wiley-VCH Verlag

GmbH & Co. KGaA: New York, 2006; pp 19ꢀ37.

(32) Price, D. A.; Blagg, J.; Jones, L.; Greene, N.; Wager, T.

Physicochemical drug properties associated with in vivo toxicological

outcomes: a review. Expert Opin. Drug Metab. Toxicol. 2009, 5, 921–931.

(33) Hughes, J. D.; Blagg, J.; Price, D. A.; Bailey, S.; DeCrescenzo,

G. A.; Devraj, R. V.; Ellsworth, E.; Fobian, Y. M.; Gibbs, M. E.; Gilles,

R. W.; Greene, N.; Huang, E.; Krieger-Burke, T.; Loesel, J.; Wager, T.;

Whiteley, L.; Zhang, Y. Physiochemical drug properties associated

with in vivo toxicological outcomes. Bioorg. Med. Chem. Lett. 2008, 18,

4872–4875.

(34) Wager, T. T.; Chandrasekaran, R. Y.; Hou, X.; Troutman,

M. D.; Verhoest, P. R.; Villalobos, A.; Will, Y. Deﬁning desirable central

nervous system drug space through the alignment of molecular proper-

ties, in vitro ADME, and safety attributes. ACS Chem. Neurosci. 2010,

1, 420–434.

(35) Wenlock, M. C.; Austin, R. P.; Barton, P.; Davis, A. M.; Leeson,

P. D. A comparison of physiochemical property proﬁles of development

and marketed oral drugs. J. Med. Chem. 2003, 46, 1250–1256.

(36) Veber, D. F.; Johnson, S. R.; Cheng, H.-Y.; Smith, B. R.; Ward,

K. W.; Kopple, K. D. Molecular properties that inﬂuence the oral

bioavailability of drug candidates. J. Med. Chem. 2002, 45, 2615–2623.

(37) Proudfoot, J. R. The evolution of synthetic oral drug properties.

Bioorg. Med. Chem. Lett. 2005, 15, 1087–1090.

(18) Walker, J. Some new curarizing agents. J. Chem. Soc. 1950,

193–197.

(19) Sasaki, J. C.; Arey, J.; Eastmond, D. A.; Parks, K. K.; Grosovsky,

A. J. Genotoxicity induced in human lymphoblasts by atmospheric

reaction products of naphthalene and phenanthrene. Mutat. Res., Genet.

Toxicol. Environ. Mutagen. 1997, 393, 23–35.

(20) Smith, C. K.; Davies, G. J.; Dodson, E. J.; Moore, M. H.

DNAꢀnogalamycin interactions: the crystal structure of d(TGATCA)

complexed with nogalamycin. Biochemistry 1995, 34, 415–425.

(21) Feigon, J.; Denny, W. A.; Leupin, W.; Kearns, D. R. Interactions

of antitumor drugs with natural DNA: proton NMR study of binding

mode and kinetics. J. Med. Chem. 1984, 27, 450–465.

(22) Hancock, A. A.; Diehl, M. S.; Faghih, R.; Bush, E. N.; Krueger,

K. M.; Krishna, G.; Miller, T. R.; Wilcox, D. M.; Nguyen, P.; Pratt, J. K.;

Cowart, M. D.; Esbenshade, T. A.; Jacobson, P. B. In vitro optimization

of structure activity relationships of analogues of A-331440 combining

radioligand receptor binding assays and micronucleus assays of potential

(38) Pajouhesh, H.; Lenz, G. R. Medicinal chemical properties of

successful central nervous system drugs. NeuroRx 2005, 2, 541–553.

(39) Lipinski, C. A. Filtering in drug discovery. Annu. Rep. Comput.

Chem. 2005, 1, 155–168.

(40) Hitchcock, S. A.; Pennington, L. D. Structureꢀbrain exposure

relationships. J. Med. Chem. 2006, 49, 7559–7583.

7619

dx.doi.org/10.1021/jm200939b |J. Med. Chem. 2011, 54, 7602–7620

Journal of Medicinal Chemistry

ARTICLE

(41) Ploemen, J.-P. H. T. M.; Kelder, J.; Hafmans, T.; van de Sandt,

H.; van Burgsteden, J. A.; Salemink, P. J. M.; van Esch, E. Use of

physicochemical calculation of pKa and CLogP to predict phospholipi-

dosis-inducing potential: a case study with structurally related piper-

azines. Exp. Toxicol. Pathol. 2004, 55, 347–355.

(42) Morelli, J. K.; Buehrle, M.; Pognan, F.; Barone, L. R.; Fieles, W.;

Ciaccio, P. J. Validation of an in vitro screen for phospholipidosis using a

high-content biology platform. Cell Biol. Toxicol. 2006, 22, 15–27.

(43) Hjelle, J. T.; Ruben, Z. Investigations in intracellular drug

storage: localization of disobutamide in lysosomal and nonlysosomal

vesicles. Toxicol. Appl. Pharmacol. 1989, 101, 70–82.

(44) Rorig, K. J.; Ruben, Z.; Anderson, S. N. Structural determinants

of cationic amphiphilic amines which induce clear cytoplasmic vacuoles

in cultured cells. Proc. Soc. Exp. Biol. Med. 1987, 184, 165–171.

(45) Ertl, P.; Rohde, B.; Selzer, P. Fast calculation of molecular polar

surface area as a sum of fragment-based contributions and its application

to the prediction of drug transport properties. J. Med. Chem. 2000, 43,

3714–3717.

(46) Hanumegowda, U. M.; Wenke, G.; Regueiro-Ren, A.; Yordanova,

R.; Corradi, J. P.; Adams, S. P. Phospholipidosis as a function of basicity,

lipophilicity, and volume of distribution of compounds. Chem. Res. Toxicol.

2010, 23, 749–755.

(47) Simplício, A. L.; Clancy, J. M.; Gilmer, J. F. Prodrugs for amines.

Molecules 2008, 13, 519–547.

(48) Wager, T. T.; Hou, X.; Verhoest, P. R.; Villalobos, A. Moving

beyond rules: the development of a central nervous system multi-

parameter optimization (CNS MPO) approach to enable alignment of

druglike properties. ACS Chem. Neurosci. 2010, 1, 435–449.

(49) Stevens, C.; De Kimpe, N. A new entry into 2-azabicyclo-

[2.1.1]hexanes via 3-(chloromethyl)cyclobutanone. J. Org. Chem. 1996,

61, 2174–2178.

(50) Schultz, A. G.; McCloskey, P. J. Carboxamide and carbalkoxy

group directed stereoselective iridium-catalyzed homogeneous oleﬁn

hydrogenations. J. Org. Chem. 1985, 50, 5905–5907.

(51) Ertl, P. Polar surface area. Methods Princ. Med. Chem. 2008,

37, 111–126.

(52) Wager, T. T.; Chandrasekaran, R. Y.; Butler, T. W. Preparation

of Pyrrolidinylmethylphenylcyclobutanecarboxamides as Histamine H3

Receptor Antagonists. PCT Int. Appl. WO2007049123, 2007.

7620

dx.doi.org/10.1021/jm200939b |J. Med. Chem. 2011, 54, 7602–7620

Article Doi

Discovery of two clinical histamine H3 receptor antagonists: Trans - N -ethyl-3-fluoro-3-[3-fluoro-4-(pyrrolidinylmethyl)phenyl] cyclobutanecarboxamide (PF-03654746) and trans -3-fluoro-3-[3-fluoro-4- (pyrrolidin-1-ylmethyl)phenyl]- N -(2-methy

DOI: 10.1021/jm200939b

Source and publish data:

Authors:

Article abstract of DOI:10.1021/jm200939b

Full text of DOI:10.1021/jm200939b

Products guided by the article

R&D Labs maybe for 935841-04-6

Relevant to this article

Hot Product