Functionalized 6-(piperidin-1-yl)-8,9-diphenyl purines as inverse agonists of the CB1 receptor – SAR efforts towards selectivity and peripheralization

Article abstract of DOI:10.1016/j.bmc.2019.07.002

Antagonists of type 1 cannabinoid receptors (CB1) may be useful in treating diabetes, hepatic disorders, and fibrosis. Otenabant (1) is a potent and selective CB1 inverse agonist that was under investigation as an anti-obesity agent, but its development was halted once adverse effects associated with another marketed inverse agonist rimonabant (2) became known. Non-tissue selective antagonists of CB1 that have high levels of brain penetration produce adverse effects in a small subset of patients including anxiety, depression and suicidal ideation. Currently, efforts are underway to produce compounds that have limited brain penetration. In this report, novel analogs of 1 are explored to develop and test strategies for peripheralization. The piperidine of 1 is studied as a linker, which is functionalized with alkyl, heteroalkyl, aryl and heteroaryl groups using a connector in the form of an amine, amide, sulfonamide, sulfamide, carbamate, oxime, amidine, or guanidine. We also report more polar replacements for the 4-chlorophenyl group in the 9-position of the purine core, which improve calculated physical properties of the molecules. These studies resulted in compounds such as 75 that are potent inverse agonists of hCB1 with exceptional selectivity for hCB1 over hCB2. SAR studies revealed ways to adjust physical properties to limit brain exposure.

Full text of DOI:10.1016/j.bmc.2019.07.002

Accepted Manuscript
Functionalized 6-(Piperidin-1-yl)-8,9-Diphenyl Purines as Inverse Agonists of
the CB1 Receptor – SAR Efforts Towards Selectivity and Peripheralization
George Amato, Robert Wiethe, Amruta Manke, Vineetha Vasukuttan, Rodney
Snyder, Scott Runyon, Rangan Maitra
PII:
S0968-0896(19)30643-1
https://doi.org/10.1016/j.bmc.2019.07.002
BMC 14986
DOI:
Reference:
To appear in:
Bioorganic & Medicinal Chemistry
Received Date:
Revised Date:
Accepted Date:
17 April 2019
18 June 2019
2 July 2019
Please cite this article as: Amato, G., Wiethe, R., Manke, A., Vasukuttan, V., Snyder, R., Runyon, S., Maitra, R.,
Functionalized 6-(Piperidin-1-yl)-8,9-Diphenyl Purines as Inverse Agonists of the CB1 Receptor – SAR Efforts
Towards Selectivity and Peripheralization, Bioorganic & Medicinal Chemistry (2019), doi: https://doi.org/10.1016/
j.bmc.2019.07.002
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Functionalized 6-(Piperidin-1-yl)-8,9-Diphenyl Purines as
Inverse Agonists of the CB1 Receptor – SAR Efforts Towards
Selectivity and Peripheralization
George Amato, Robert Wiethe, Amruta Manke, Vineetha Vasukuttan, Rodney Snyder, Scott
Runyon and Rangan Maitra*
Discovery Science and Technology, RTI International, 3040 Cornwallis Rd., Research Triangle Park, NC
2
7709-2194, USA
*
Corresponding author
Key Words: CB1, cannabinoid, peripheral, antagonist, otenabant, purine, CB2, blood brain
barrier, endocannabinoid, inverse agonist
Abbreviations Used:
BBB,
Blood-brain
barrier,
BOP,
(benzotriazol-1-yloxy)tris(dimethylamino)phosphonium
hexafluorophosphate; CB1, cannabinoid receptor 1; CB2, cannabinoid receptor 2; CHO-K1, chinese
hamster ovary cells; EtOAc, ethyl acetate; EtOH, ethanol; Ke, apparent affinity constant, MDCK-mdr1,
Madin-Darby canine kidney cells transfected with the human MDR1 gene; IP , inositol phospahatase 3;
3
MeOH, methanol; MRM, multiple reaction monitoring; NaHMDS, sodium bis(trimethylsilyl)amide; TEA,
triethylamine.
1

ABSTRACT
Antagonists of type 1 cannabinoid receptors (CB1) may be useful in treating diabetes, hepatic
disorders, and fibrosis. Otenabant (1) is a potent and selective CB1 inverse agonist that was under
investigation as an anti-obesity agent, but its development was halted once adverse effects associated with
another marketed inverse agonist rimonabant (2) became known. Non-tissue selective antagonists of CB1
that have high levels of brain penetration produce adverse effects in a small subset of patients including
anxiety, depression and suicidal ideation. Currently, efforts are underway to produce compounds that have
limited brain penetration. In this report, novel analogs of 1 are explored to develop and test strategies for
peripheralization. The piperidine of 1 is studied as a linker, which is functionalized with alkyl, heteroalkyl,
aryl and heteroaryl groups using a connector in the form of an amine, amide, sulfonamide, sulfamide,
carbamate, oxime, amidine, or guanidine. We also report more polar replacements for the 4-chlorophenyl
group in the 9-position of the purine core, which improve calculated physical properties of the molecules.
These studies resulted in compounds such as 75 that are potent inverse agonists of hCB1 with exceptional
selectivity for hCB1 over hCB2. SAR studies revealed ways to adjust physical properties to limit brain
exposure.
2

1
. Introduction
Regulation of the endocannabinoid (EC) system, comprised of endocannabinoids, receptors,
transporters and enzymes, has several potential medical applications.1-6 Cannabinoid receptors CB1 and
CB2 are G proteins (G )-coupled G protein-coupled receptors (GPCRs). CB1 receptors are highly
i/o
expressed in the central nervous system (CNS) and in several peripheral organs. Selective attenuation of
CB1 receptor signaling is a validated approach for treating obesity, diabetes, metabolic syndrome,
dyslipidemias and liver diseases.4
The inverse agonist rimonabant (2, Figure 1), which has a high degree of CNS penetration, was
clinically approved for the treatment of obesity in Europe. But this drug had to be withdrawn because of an
increase in psychiatric disorders associated with antagonism of CB1 receptors in the CNS in a subset of
7
patients. The adverse side effects seen with 2 led to discontinuation of clinical development of other such
compounds including 1 (Figure 1). Efforts are now underway to develop compounds that selectively
antagonize the CB1 receptors in the periphery.8-11 These compounds are expected to avoid CNS-related
adverse effects noted with centrally acting ligands. Furthest along among these is TM38837 (Figure 2),
which was tested in humans and demonstrated to have limited brain penetration.12 Two additional
compounds (Figure 2) have completed IND-enabling studies and are poised for clinical development.13, 14
Alternative approaches for this class of therapeutics include development of neutral orthosteric antagonists,
partial agonists, and negative allosteric modulators.15, 16
O
N
Me
NH
N
N
Cl
O
N
N
N
N
N
NH2
Cl
Cl
NH
Cl
Et
Cl
1
, Otenabant
2, Rimonabant
(SR141716A)
(CP-945,598)
Figure 1. Examples of clinical CNS penetrating CB1 inverse agonists.
3

O
O
S
H N
2
O
O
S
O
N
NH
N
O
N
O
NH
Et
HN
NH
i-Pr
Cl
N
N
N
N
N
N
N
S
S
Cl
Cl
Cl
Cl
Cl
F3C
F3C
JD-5037
TM-38837
DBPR-211
Figure 2. Examples of advanced peripherally restricted CB1 inverse agonists.
Analyses of clinically used compounds show that a topological polar surface area (TPSA) of 80-
40 Å and a molecular weight of 450-600 Da can improve the likelihood of producing orally active
1
compounds that have limited brain exposure.17, 18 Inclusion of one or two hydrogen bond donors may also
limit CNS penetration, but has the potential to negatively impact oral absorption. Incorporation of a mildly
basic group into the structure can enhance oral bioavailability by improving solubility in acidic media, such
1
9
as that of the stomach. To improve the potential for good drug-like properties, it is also desirable to have
a cLogP of less than 5. Compounds with a high cLogP may have poor aqueous solubility, which in turn can
negatively impact absorption following an oral dose.20 While analogs of 2 have been extensively explored,
there are far fewer studies of compounds based on 1, perhaps because 1 was never approved as a drug.
Compound 1 has many desirable characteristics that could enable peripheralization, including a TPSA of
1
02 Å, 3 hydrogen bond donors and a molecular weight of >500 Da.21 Studies indicate that intra-molecular
H-bonding between the primary amide and the ethyl amine portion of 1 (observed in the X-ray structure)
effectively lowers its polarity, allowing for penetration into the CNS.21 Compounds related to 1 are
particularly attractive because the mildly basic 6-piperidinyl purine core can aid in oral absorption.
A detailed analysis of the SAR associated with 1 would facilitate the development of better
compounds. We began such an effort with an investigation of isomeric replacements for the 4-
aminopiperidine group 22 and employment of the piperazine group as a replacement.23, 24 Early studies
4

resulted in compounds 3-5, possessing excellent hCB1 potency (Ke < 20 nM) and good selectivity (>50
fold) against hCB2 (Figure 3).22, 23, 25 The goal of these studies was to further investigate related analogues.
O
Me
O
S
O
N
N
n-Bu
N
NH
N
N
N
N
NH
N
N
N
Cl
N
N
N
N
N
N
Cl
Cl
Cl
Cl
Cl
Cl
4
5
3
Figure 3. Early purine based peripherally selective CB1 inverse agonists.
Recent crystal structures of hCB1 coupled with docking studies with 1 and our past research
indicate that functionalization of the piperidine group at the 4-position could result in access to a binding
pocket near the extracellular surface of the plasma membrane.23, 26, 27 This pocket may accommodate polar
as well as nonpolar groups and differs significantly from hCB2. By targeting this region, we hypothesized
that it is possible to identify compounds with better drug-like properties that have good hCB1 potency and
excellent selectivity versus hCB2.
2
2
.
Results and Discussion
.1
Compound Design and Synthesis
N
N
N
L R
L = NH, NMe, NHCO, NHSO , NHCO ,
2
2
N
N
Ar
NHC(=NH), C(=NOH),NHC(=NH)NH
R = Alkyl, Heteroalkyl, Aryl, Heteroaryl
Ar = 4-SubstitutedPh, Heteroaryl
Cl
6
Figure 4. General strategy for SAR studies
5

In this report, we explored functionalization of the piperidine linker at the 4-position with an amine,
amide, sulfonamide, sulfamide, carbamate, oxime, amidine or guanidine connector (6, Figure 4). Polar and
nonpolar groups were explored to establish the needed SAR to develop better compounds. We also report
our attempt to increase the TPSA and reduce cLogP via replacement of the 4-chlorophenyl in the 9-position
of the purine core with more polar groups, including heterocycles. Combined, these changes allowed us to
design a variety of compounds with properties that were predicted to favor both peripheral selectivity and
oral absorption.
To prepare most of the targeted compounds, we used the 4-aminopiperidine 8a (Scheme 1) as a
key intermediate, the preparation of which was previously described.22 The 4-methylaminopiperidine
analog (9) and analogs of 8a with replacements for the 4-chlorophenyl group of the purine ring were
prepared in a similar way as shown in Scheme 1 and detailed in the experimental section. Of note, in the
conversion of 7f to 8f, the acetyl group was lost during hydrolysis of the boc group.
N
N
N
Me
NH
N
Cl
N
N
NH2
N
N
N
a
N
Cl
NH2
N
N
b
N
N
N
N
Ar
Ar
Cl
Cl
OR
Cl
Cl
Cl
7(a-j)
8(a-j)
9
o
Scheme 1. Reagents and conditions: (a) (1) ArNH , 6 N HCl, n-BuOH, 80 C; (2) 2-ClPhCHO, 15%
2
o
FeCl /silica gel, dioxane, 95 C, then DDQ, CH Cl , rt; (b) (1) 4-N-Boc-aminopiperidine or 4-N-Boc-
3
2
2
o
o
methylaminopiperidine, K CO , NMP, 80 C; (2) 6 N HCl, EtOH, 50 C. Ar = 4-chlorophenyl (7a, 8a), 4-
2
3
fluorophenyl (7b, 8b), 6-(trifluoromethyl)pyridin-3-yl (7c, 8c), 6-(difluoromethoxy)pyridin-3-yl (7d, 8d),
-carboxamidophenyl (7e, 8e), 6-(acetamido)pyridin-3-yl (7f), 6-(amino)pyridin-3-yl (8f), 5-methyl-1,2-
oxazol-3-yl (7g, 8g), 1-methyl-1H-pyrazol-4-yl (7h, 8h), 1-methyl-1H-pyrazol-3-yl (7i, 8i), 5-
trifluoromethyl)-1H-pyrazol-3-yl (7j, 8j).
4
(
6

From intermediate 8a we prepared a variety of amines, sulfonamides, amides and carbamates using
standard procedures as shown in Scheme 2. Intermediate 9 and analogues with a replacement for the 4-
chlorophenyl group of the purine ring were functionalized in the same way. An amidine (23) and a
guanidine (24) were also prepared from 8a (Scheme 2). The heterocyclic analog 78 was prepared from
acetylation of 77 using acetyl chloride, pyridine and tetrahydrofuran. The benzimidazole analog 22 was
prepared directly from 7a by displacement of the purine chloro with 2-(4-piperidinyl)-1H-benzimidazole.
N
NH
N
H
N
N
N
N
R
N
N
NH2
H
N
N
N
SO2R
N
N
N
N
a or b
c
N
N
Cl
Cl
Cl
Cl
Cl
Cl
8a
d or e
h or i
f or g
N
N
O
N
H
H
H
N
N
N R
N
N
N
CO2R
N
N
N
R
N
N
N
N
N
N
Cl
Cl
Cl
Cl
Cl
Cl
.
.
Scheme 2. Reagents and conditions: (a) RC(=NH)OMe HCl, NEt , EtOH, reflux; (b) RNH(C=NH)SMe HI,
3
i-PrOH, reflux; (c) RSO Cl, NEt , THF, rt; (d) RO CCl, NEt , DCE, rt; (e) (1) p-F-PhO CCl, NEt , DCE,
2
3
2
3
2
3
rt; (2) ROH, NaHMDS, THF, rt; (f) RCOCl, NEt , THF, rt; (g) RCO H, BOP, NEt THF, rt; (h) ArF, 125
3
2
3,
o
C; (i) RCHO, NaCNBH , AcOH, MeOH, rt.
4
The oximes 20 and 21 were prepared as shown in Scheme 3. Reaction of 7a with commercially
available 4-(4-fluorobenzoyl)piperidine provided the ketone 10. Reaction of 10 with hydroxylamine
provide a mix to the two isomeric oximes (E and Z isomers), which were separated by chromatography, but
not assigned.
7

OH
N
N
N
O
N
N
Cl
N
N
F
N
N
F
N
N
a
N
N
b
N
N
Cl
Cl
Cl
Cl
Cl
Cl
7a
10
20 & 21
o
Scheme 3. Reagents and conditions: (a) 4-(4-Fluorobenzoyl)piperidine, NEt , EtOH, 80 C; (b)
3
.
o
H NOH HCl, EtOH, 80 C.
2
8

2
.2.
Pharmacological Characterization
All target compounds were evaluated in a FLIPR-based calcium mobilization assay for hCB1
activity (Tables 1-3) as has been described in our previous publications.22, 28 In general, those that
demonstrated apparent antagonist dissociation equilibrium constant (Ke) <50 nM were further tested for
3
affinity to hCB receptors using radioligand displacement of [ H]CP55940 in purified membrane fractions
overexpressing either hCB1 or hCB2. A selection of potent and selective compounds was further tested for
potential peripheral selectivity by calculating % apical (A) to basal (B) permeability in a MDCK-mdr1
2
9
monolayer permeability assay, which is predictive of brain penetration. A lower number predicts less
transport across the blood-brain barrier (BBB). To assess metabolic stability, some compounds were
advanced to a human liver microsome (HLM) stability assay, calculating half-life (HL) and clearance (Cl)
values.
Select compounds with favorable in vitro properties were tested in mice to assess their ability to be
orally absorbed and kept out of the brain. Initial studies entailed cassette dosing of mice and measuring the
brain and plasma levels at multiple time points. Plasma and brain maximum concentrations (Cmax) were
obtained and used to calculate and compare brain/plasma levels. The main goals were to achieve good
plasma levels (>100 ng/mL after an oral dose of 2.5 mg/kg) while keeping the brain to plasma ratio as low
as possible, preferably less than 0.05 (5% penetration) using the maximum concentrations in the brain and
plasma.
Connecting groups to the piperidine via an amine results in compounds that are basic in nature.
Furthermore, the basicity can be tuned by the nature of the group that is connected. This could aid in
achieving good oral bioavailability. In Table 1, in vitro data for a set of amines and a few related analogs
are presented. Aryl amines are mildly basic, a trait that is often found in commercial drugs. The
unsubstituted phenyl amine 11 was found to be highly potent and selective versus hCB1. To improve
9

calculated physical properties (see supplemental data) and avoid the potential metabolic liability associated
with anilines, we prepared heterocyclic versions of 11.
Table 1. In Vitro Data for CB1 Antagonists -Amines & Additional Analogs
N
N
N
R
N
N
Cl
Cl
MDCK
-mdr1
Ke
hCB1
nM)
Ki
hCB1
(nM)
Selectivity
Ki
CB2/CB1
Ki hCB2
#
1
R
a
(nM)
A to B
a
(
b
(%)
Otenabantc
PhNH
0.2
0.7
7700
11000
>1600
ND
0
1
1
2
1.4±0.4
54±7
6±1
>10000
N
1
1
2.4±0.7 3100±100
1300
ND
0.1
NH
N
3
NH
N
43±14
ND
ND
ND
1
1
1
1
4
5
6
7
N
NH
100±30
150±50
ND
ND
ND
ND
ND
ND
ND
64
ND
ND
ND
ND
BnNH
N
CH NH
580±110
34±4
ND
ND
2
BnNMe
42±12
2700±500
1
0

1
1
8
9
F
F
CH NH
210±130
460±300
ND
ND
ND
ND
ND
ND
ND
ND
2
CH NMe
2
HO
N
2
2
0
1
20±12
21±3
5±2
7±2
>10000
>10000
>2000
>1400
ND
ND
F
OH
N
F
N
2
2
12±3
47±26
830±20
18
ND
N
H
NH
NH
2
2
3
4
170±120 780±80
>6000
ND
>7
0.2
F
NH
H
N
300±50
ND
ND
ND
F
NH
Data are reported as mean + SD from three independent experiments.
a
3
Displacement was measured using [ H]CP55940 in CHO cell membrane preparations
overexpressing hCB1 or hCB2 receptors.
b% transported from the apical side (A) to the basal side (B).
cSee reference 21
ND: not done as compound did not meet progression criterion/criteria
One such compound, the 2-pyridyl analog 12 retained good hCB1 potency and selectivity versus
hCB2. Although the calculated physical properties were not optimal (TPSA = 72, cLogP = 6.2), they were
an improvement over 11. In the MDCK-mdr1 assay, compound 11 performed well, with 0.1% permeability
across the polarized cell monolayer. To check how close we were to our pharmacokinetic goals, this
1
1

compound was progressed to an oral pk study in mice at 2.5 mg/kg. While it was orally bioavailable, the
plasma Cmax was low (38 ng/mL) and the peripheral selectivity was not good enough (brain/plasma Cmax
=
0.45). To enhance oral absorption, more basic compounds were of interest and so a set of benzyl amines
was prepared. These compounds (15-19) were less active than the aryl amines and were not pursued.
Interestingly, a set of compounds containing an imine were more successful. The oximes 20 and 21
(regioisomers not assigned) both presented good hCB1 potency and hCB2 selectivity. This was
encouraging as oximes contain a strong hydrogen bond donor, which could favor peripheral restriction.
These compounds were tested for HLM stability. One regioisomer, 21, performed slightly better (HL = 101
minutes, Cl = 12 L/min/mg), with values that are close to our targets (HL > 2 hours, Cl < 10 L/min/mg).
The more compact benzimidazole 22 also has good hCB1 potency, but poor selectivity versus hCB2. The
more polar and basic amidine 23 and guanidine 24 were also tested. While these compounds were not highly
potent, the amidine 23 (hCB1 Ke = 170 nM) is more polar (TPSA = 83), less lipophilic (cLog D = 4.1),
more basic and has an added hydrogen bond donor compared to 12. It was of interest to know if these
changes would lead to an improvement in peripheral selectivity, as expected. In the MDCK-mdr1 assay,
compound 23 was found to have 0.2% permeability, predicting peripheral selectivity. The HLM stability
of 23 was excellent (HL = 166 minutes, Cl = 8 L/min/mg) and the compound was progressed to a mouse
oral pk study at 2.5 mg/kg. The compound was orally bioavailable with a plasma half-life of 6.0 hours and
a plasma Cmax of 50 ng/mL. Compared to 12, the peripheral selectivity of 23 was significantly better, with
a brain to plasma ratio of the Cmax of 0.16, but this is still short of our goal of 0.05 or less.
Generally, secondary amides and carbamates are good connectors for achieving peripheral
selectivity, as they contain a strong hydrogen bond donor and contribute heteroatoms that increase TPSA.
In Table 2, in vitro data for a set of alkyl and aryl amides as well as alkyl carbamates are presented. We
previously reported on aliphatic amides with good hCB1 potency, but physical properties needed to be
improved.25 We tested more polar compounds by incorporating heteroatoms into the alkyl groups. The
previously reported n-butyl amide 25 has good potency against hCB1, but less than desirable selectivity
1
2

versus hCB2 and poor calculated physical properties.25 Replacing one of the carbons of the n-butyl chain
with an oxygen (26 and 27) or a sulfoxide (28) led to more desirable physical properties, but greatly
diminished hCB1 potency. Similar results were obtained in our efforts to incorporate heteroatoms into
cycloalkyl amides. For example, while the cyclohexylmethyl analog 41 has a hCB1 Ke of 2 nM,
replacement of one of the ring atoms with oxygen or nitrogen (see compounds 42-44) resulted in Ke values
>
200 nM. One exception for which an oxygen was tolerated in the alkyl group was the t-butoxyethyl amide
3
2. This compound retained good hCB1 potency (Ke = 1.7 nM and Ki = 6 nM) and was highly selective
versus hCB2. The maintenance of good hCB1 potency for 32 could be a result of the oxygen being shielded
by the large t-butyl group. While this compound has improved calculated physical properties (cLogP of 5.0
and TPSA of 85), the MDCK-mdr1 assay predicted poor peripheral selectivity (A to B of 20%). We also
investigated improving the physical properties of the n-butyl amide 25 by replacing the 4-chlorophenyl
with the 6-trifluoromethypyrid-3-yl group. The resulting compound, 29, had good hCB1 potency and
selectivity versus hCB2. Calculated physical properties also improved, with a lower cLogP (5.2) and higher
TPSA (89), but the MDCK-mdr1 assay predicted poor peripheral selectivity (A to B of 28%). In summary,
for alkyl and cycloalkyl amides, a high TPSA combined with a high MW and a single hydrogen bond donor
were not sufficient for good peripheral selectivity.
Table 2. In Vitro Data for CB1 Antagonists – Amides & Carbamates
O
N
H
N
N
N
R
N
N
R'
Cl
MDCK-
mdr1
Ke
hCB1
nM)
Ki
hCB1
(nM)
Ki
hCB2
(nM)
Selectivity
Ki
CB2/CB1
#
R
R’
A to B
a
a
(
b
(%)
2
5c
n-Bu
4-Cl-Ph
5
19
1000
53
ND
1
3

2
2
6
7
MeOCH CH
4-Cl-Ph
4-Cl-Ph
370±70
240±70
ND
ND
ND
ND
ND
ND
ND
ND
2
2
2
MeCH OCH
2
6
000
2
2
8
9
MeSO CH CH
4-Cl-Ph
ND
3±2
ND
ND
140
ND
28
2
2
2
±
1000
4
20
F3C
n-Bu
4±2
N
±
180
3
3
0
1
t-BuCH2
4-Cl-Ph
4-Cl-Ph
0.3±0.2
0.2±0.1
ND
15±1
ND
75
ND
ND
HO
CH2
1700±300
ND
4
900
3
2
t-BuOCH CH
4-Cl-Ph
1.7±0.4
6±3
820
20
2
2
±
1300
3
3c
c-Pen
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
0.9
150±40
3
1.5
ND
3
60
40
ND
ND
ND
ND
O
3
4
ND
840
ND
ND
250
ND
3
5c
c-Hex
3
6
7
O
440±90
ND
1
900
700
3
S
4-Cl-Ph
26±7
3±2
630
ND
±
1
4

3
8
O S
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
4-ClPh
4-Cl-Ph
>10000
ND
2
ND
570
ND
ND
280
ND
ND
ND
ND
ND
ND
2
3
9c
c-PenCH2
3
1000±100
2
O
4
0
CH2
ND
2
4
1c
c-HexCH2
2400
ND
1200
ND
4
2
3
4
N CH2
220±90
ND
4
O
CH2
4-Cl-Ph
510±90
ND
ND
ND
ND
4
O
N CH2
4-Cl-Ph
4-Cl-Ph
1100±300
0.7
ND
4
ND
ND
ND
ND
4
5d
Ph
9600
2400
4
6
7
8
Ph
F3C
19±13
140±30
0.6±0.2
8±3
ND
740±60
ND
92
ND
3
N
N
4
4-Cl-Ph
ND
0.1
0
.98
6000
4
2-F-Ph
4-F-Ph
4-Cl-Ph
6000
±
0.02
±1900
1
0000
4000
4
5
9
0
4-Cl-Ph
4-Cl-Ph
1.1±0.6
1.2±0.6
2±1
2±1
5000
0.2
±
2,4-DiF-Ph
>7000
>3500
<0.1
1
5

5
900
5
5
1
2
3,4-DiF-Ph
4-Cl-Ph
4-Cl-Ph
0.7±0.3
2±1
3±1
3±2
2000
50
0.2
±
1500
Me
Ph CH
150±70
ND
<0.1
Me
N
5
3
N
N
4-Cl-Ph
>10000
ND
ND
ND
CH CH2
2
N
5
5
4
5
MeOCH CH O
4-Cl-Ph
4-Cl-Ph
6±1
17±3
>5000
>200
1800
13
2
2
1
050
70
c-BuO
1.0±0.3
0.6±0.3
ND
±
5
5
6
7
O
O
4-Cl-Ph
4-Cl-Ph
92±14
ND
ND
ND
ND
ND
ND
ND
ND
O
O
320±150
O
CH O
2
5
8
4-Cl-Ph
77±15
ND
ND
ND
ND
ND
MeN
CH O
2
5
6
9
0
4-Cl-Ph
4-Cl-Ph
1500±600
3±2
ND
4±1
ND
180
ND
ND
7
10
BnO
±
100
Data are reported as mean + SD from three independent experiments.
a
3
Displacement was measured using [ H]CP55940 in CHO cell membrane preparations overexpressing
hCB1 or hCB2 receptors.
b% transported from the apical side (A) to the basal side (B).
1
6

cSee reference 25
dSee reference 22
ND: not done as compound did not meet progression criterion/criteria
Aryl amides have the potential to be more metabolically stable than alkyl amides and the
previously reported unsubstituted phenyl amide 45 has excellent hCB1 potency and selectivity versus
hCB2.22 Fluorinated analogs of 45 were pursued as such compounds may have better pharmacokinetic
properties, in particular oral bioavailability and plasma half-life. Furthermore, it was perceivable that the
electron withdrawing nature of fluorine might enhance the acidity of the amide, thereby improving
peripheral selectivity. The monofluorinated analogs 48 and 49 and the difluorinated analogs 50 and 51 were
prepared and found to be highly potent hCB1 antagonists with excellent selectivity versus hCB2. The 2-
fluorophenyl amide 48, for example had a hCB1 Ke of ~0.6 nM, a Ki of ~1 nM and ~6000-fold selective
for hCB1 over hCB2. These excellent results prompted us to test fluorinated phenyl amides in the MDCK-
mdr1 assay for potential peripheral selectivity. Each compound tested was found to have <1% permeability
in MDCK-mdr1 transport assay, indicating peripheral selectivity. The fluorinated phenyl amides, however,
have values for cLogP >6 and TPSA <80 and so we sought compounds with better calculated physical
properties. The 2-pyridyl amide 47 has significantly improved calculated physical properties, but a 100-
fold loss in hCB1 potency. Replacement of the 4-chlorophenyl of 45 with the 6-trifluoromethylpyrid-3-yl
group resulted in 46, a compound with better calculated physical properties and acceptable hCB1 potency.
In the MDCK-mdr1 assay, however, 46 did not perform well with ~3% transport.
Compared to amides, carbamates have an additional oxygen atom, which improves calculated
physical properties. The c-butyl carbamate 55 had excellent hCB1 potency and selectivity versus hCB2.
The calculated physical properties were near to our goals, with a TPSA of 85 and cLogP of 5.7. As with
the amides, attempts to further improve the calculated properties without losing potency was difficult. One
exception is the methoxyethyl carbamate 54 (TPSA = 94 and cLogP = 4.7) with a hCB1 Ke of 6 nM. In the
MDCK-mdr1 assay, however, this compound performed poorly. Proton NMR indicated that the
1
7

methoxyethyl group is involved in an intramolecular hydrogen bond with the amide, thereby making the
compound significantly less polar and more likely to pass into the CNS.
Purine sulfonamides were of interest as potential peripheral agents because of the inherent strong
hydrogen bond donor and added heteroatoms that make up this connector. In Table 3 is presented in vitro
data for a set of alkyl and aryl sulfonamides. We previously reported that the methyl sulfonamide 61 has
good hCB1 activity, but poor selectivity versus hCB2.22 We proceeded to investigate larger alkyl
sulfonamides to take advantage of the pocket near to the extracellular membrane surface which differs
significantly between hCB1 and hCB2. Our results with analogs 62-64 support our hypothesis. These
compounds were both more potent for hCB1 and more selective versus hCB2 and were subsequently tested
in the MDCK-mdr1 assay for potential peripheral selectivity. Compounds 63 and 64 were both found to
have A to B transport of <0.1%, predicting peripheral selectivity. The sulfamide analog of 64, compound
6
5, was of interest due to the added hydrogen bond donor, increased TPSA (105) and decreased cLogP
(5.4), which favor peripheral restriction. Satisfyingly, the more polar sulfamide 65 maintained the good
potency of the sulfonamide 64. The selectivity versus hCB2, unfortunately, was not as good. Despite this
set-back, it was of interest to know if such compounds would be orally bioavailable and peripherally
restricted and so we proceeded to dose 65 orally to mice. Unfortunately, a low maximum plasma
concentration resulted (4 ng/mL at a dose of 0.3 mg/kg).
Aryl sulfonamides have the potential to be more metabolically stable than alkyl sulfonamides and
the previously reported phenyl sulfonamide 66 was shown to have excellent hCB1 potency and selectivity
versus hCB2.22 Fluorinated analogs of 66 were pursued as such compounds could possess better
pharmacokinetic properties, in particular oral bioavailability and plasma half-life. Furthermore, it was
perceivable that the electron withdrawing nature of fluorine might enhance the acidity of the sulfonamide,
thereby improving peripheral selectivity. The monofluorinated analogs 67 and 71 and the difluorinated
analogs 72 and 73 were prepared and found to be highly potent hCB1 antagonists with excellent selectivity
versus hCB2. The 2-fluorophenyl sulfonamide 67 was among the most potent and selective hCB1
1
8

antagonists (Ke = 0.8 nM, Ki = 0.38 nM; Ki hCB2/hCB1 = 22,000). These excellent results prompted us to
test fluorinated phenyl sulfonamides in the MDCK-mdr1 assay for potential peripheral selectivity. Each
compound tested was found to have an A to B of 0.1% or less, favoring peripheral selectivity. With these
results, the 3,4-difluorophenyl sulfonamide 73 was dosed orally to mice to assess its pharmacokinetic
profile. The maximum plasma concentration of 73 was found to be low (12 ng/mL with a 2.5 mg/kg dose)
and the peripheral selectivity suboptimal (brain/plasma Cmax of 0.31). While the TPSA value for this
compound is 96, favorable for peripheral selectivity and oral bioavailability, the cLogP is high (6.3). To
enhance the potential for desirable pharmacokinetics, it was of interest to investigate compounds with a
lower cLogP. Replacements for the 4-chlorophenyl group in the 9-position of the purine core were
investigated. Swapping the chloro for a fluoro (67 vs 68) resulted in a reduction of cLogP from 6.1 to 5.8,
accompanied by a 100-fold loss in hCB1 Ke activity and a 10-fold loss in binding, highlighting the
importance of the 4-chlorophenyl group for good potency. Replacement of the 4-chlorophenyl group with
the 6-trifluoromethylpyrid-3-yl group, resulted in compounds with comparable potency (see 73 vs 74). This
was encouraging as the cLogP for 74 was reduced from 6.3 to 5.7 and the TPSA was increased from 96 to
1
06. Similarly, the 6-difluoromethoxypyrid-3-yl group was favorable for potency (see 73 vs 75) and once
again, the cLogP was reduced to 5.7 while the TPSA was increased to 116, prompting us to select this
compound for additional studies. The stability of 75 to human liver microsomes (HLM) was tested and
found to be excellent, with a half-life of 274 minutes and a clearance of 5 L/min/mg. We then proceeded
to dose 75 orally to mice at 2.5 mg/kg for a pharmacokinetic analysis. Encouragingly, a significant
improvement over 73 was found (maximum plasma concentration of 62 ng/mL and brain/plasma Cmax of
0
.18). Even more polar replacements for the 4-chlorophenyl were tried, keeping the TPSA over 110, while
reducing the cLogP below 5 (see compounds 76-82), but this resulted in a significant loss of hCB1 potency
as we surpassed the polarity limit of this region of the binding pocket.
Table 3. In Vitro Data for CB1 Antagonists – Sulfonamides
1
9

N
O O
S R
N
N
N
H
N
N
R'
Cl
MDCK-
mdr1
Ke
hCB1
nM)
Ki
hCB1
(nM)
Ki
hCB2
(nM)
Selectivity
Ki
CB2/CB1
#
R
R’
A to B
a
a
(
b
(%)
6
1c
Me
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
4-Cl-Ph
71
29
1100
37
105
63
ND
1
6
2
3
4
5
CF CH CH
2
12±10
29±20
31±17
8±4
4±1
9±4
9±4
5±2
1.2
420±50
570±60
>10000
330±20
5900
3
2
6
6
6
n-Pent
c-HexCH2
c-HexNH
Ph
<0.1
<0.1
ND
1
>1100
66
6
6c
0.3
5100
0
.38
8500
6
7
8
9
2-F-Ph
2-F-Ph
2-F-Ph
4-Cl-Ph
0.8±0.4
21±13
22000
>3000
100
ND
ND
ND
±
0.01
±2200
6
4-F-Ph
2.6±0.1 >10000
N
400
4±1
6
7
0.5±0.1
F3C
±
120
N
F2HCO
0
2-F-Ph
0.7±0.2
1.6±0.9 190±20
120
ND
2
0

9
000
7
1
4-F-Ph
4-Cl-Ph
4-Cl-Ph
1.1±0.5
3.3±2.8
2±1
>2000
0.1
±
2000
7
7
2
3
2,4-DiF-Ph
3,4-DiF-Ph
2±1
>10000
>5000
>5000
0.1
4-Cl-Ph
0.5±0.4
1.7±0.7 >10000
<0.1
N
7
7
4
5
3,4-DiF-Ph
3,4-DiF-Ph
1.0±0.4
1.6±0.7
10±5
3±1
>6000
>600
290
ND
ND
F3C
N
F2HCO
860±40
O
H N
7
7
7
6
7
8
3,4-DiF-Ph
3,4-DiF-Ph
3,4-DiF-Ph
2
330±190
>10000
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
N
H N
2
N
2
700
AcHN
±
1200
N
O
7
9
3,4-DiF-Ph
56±28
420±90 >10000
>20
ND
Me
N
8
8
0
1
3,4-DiF-Ph
3,4-DiF-Ph
140±130
120±30
ND
ND
ND
ND
ND
ND
ND
ND
MeN
MeN
N
2
1

F3C
5000
8
2
3,4-DiF-Ph
ND
ND
ND
ND
HN
±1000
N
Data are reported as mean + SD from three independent experiments.
a
3
Displacement was measured using [ H]CP55940 in CHO cell membrane preparations overexpressing
hCB1 or hCB2 receptors.
b% transported from the apical side (A) to the basal side (B).
cSee reference 22
ND: not done as compound did not meet progression criterion/criteria
2
.3
Inverse Agonism
Compounds 1 and 2 are inverse agonists. Compounds 48, 67 and 75 were further characterized
using the calcium assay to establish whether these compounds are neutral antagonists or inverse agonists
of the hCB1 receptor. As shown in Figure 5, these compounds are inverse agonists of hCB1, suppressing
signaling through hCB1 with increasing concentration. The EC for 48, 67 and 75 were 94.8, 287.4 and
5
0
1
31.9 nM respectively.
2
2

2000
1500
1000
Compound 75
Compound 48
Compound 67
Vehicle
500
0
-4
-2
0
2
4
6
log [ligand], nM
Figure 5. Compounds 48, 67 and 75 are inverse agonists of hCB1. CHO-CB1 cells were loaded with
calcium indicator dye for 60 min as has been described in Methods. Cells were then stimulated by
various concentrations of each compound and fluorescence change recorded using FLIPR Tetra
(Molecular Devices) instrument. Data reported as Mean ± SEM from 3 independent measurements.
3
.
Summary and Conclusions
Peripherally restricted CB1 antagonists could become important tools in treating obesity, metabolic
syndromes, dyslipidemias and liver diseases such as NASH and fibrosis. Compound 1 is a potent and
selective CB1 inverse agonist developed by Pfizer, but its advancement in clinical trials was halted because
of its ability to penetrate the CNS, which is associated with an increased risk of psychiatric disorders.
Structure-activity relationships of 1 would be useful in developing better compounds with limited CNS
penetration. Recent crystal structures of hCB1 and docking studies with otenabant indicate that the 4-
postion of the piperidine group could be functionalized to access a binding pocket that may accommodate
both polar and nonpolar groups. The same binding pocket differs significantly from that of hCB2 and hence
could be used to obtain highly selective compounds. In this report, the piperidine group was functionalized
in the 4-position with alkyl, heteroalkyl, aryl and heteroaryl groups using a connector in the form of an
amine, amide, sulfonamide, sulfamide, carbamate, oxime, amidine or guanidine (6). We also reported in
2
3

this paper the investigation of heterocyclic replacements for the 9-phenyl group, which served to decrease
the cLogP and increase the TPSA of the molecules.
Gratifyingly, our efforts have produced compounds with exceptional hCB1 potency and selectivity
versus hCB2. The aryl amide 48 and the aryl sulfonamide 67 are among the most potent and selective
compounds tested by our group to date. Systematic adjustment to the functionalization of the piperidine
showed that polarity is tolerated close to the 4-position of the piperidine ring. Potent compounds were
obtained by functionalization with polar connectors such as sulfonamides, sulfamides, amides, carbamates,
amidines and oximes. On the other hand, the small alkyl or aryl groups attached to these polar connectors
required a less polar nature to maintain good hCB1 potency. Attempts to further adjust molecular physical
properties by replacing the 4-chlorophenyl group at the 9-position of the purine core resulted in favorable
improvements with the 6-trifluoromethylpyrid-3-yl and 6-difluoromethoxypyrid-3-yl groups. We identified
several compounds with good hCB1 potency, good selectivity versus hCB2 and calculated physical
properties favoring peripheral restriction. Two such compounds, the aryl sulfonamide 75 and the amidine
2
3, were further studied and found to have excellent HLM stability. In a mouse oral pk study, they were
orally bioavailable and peripherally selective, approaching our goals. These compounds have <20% brain
penetration based on Cmax values. While not meeting our stringent criterion of <5% CNS penetration,
1
9
these compounds are far less brain penetrant than 1, which has ~75% CNS penetration. Most importantly,
these foundational studies provide us with strategies for compound development to produce peripherally
restricted CB1 receptor inverse agonists in the future. In particular, future studies will focus on polarity and
hydrogen bond donors near to the piperidine linker to attain the desired level of peripheral selectivity.
Deeper into piperidine linked pocket, less polar groups will be used to attain high levels of CB1 potency
and selectivity versus CB2.
4
.
Experimental
2
4

Chemistry General. Purity and characterization of compounds were established by a combination of MS,
1
NMR, HPLC and TLC analytical techniques, as described below. H spectra were recorded on a Bruker
Avance DPX-300 (300 MHz) spectrometer and were determined in chloroform-d (7.26 ppm) or methanol-
d4 (3.31 ppm) with tetramethylsilane (0.00 ppm) or solvent peaks as the internal reference unless otherwise
noted. Chemical shifts are reported in ppm relative to the solvent signal and coupling constant (J) values
are reported in hertz (Hz). Thin-layer chromatography (TLC) was performed on EMD precoated silica gel
6
0 F254 plates. TLC spots were visualized with UV light or I detection. Low-resolution mass spectra were
2
obtained using a single quadrupole PE Sciex API 150EX (ESI). Unless stated otherwise, all test compounds
were at least 95% pure as determined by HPLC. HPLC method: an Agilent-Varian system equipped with
Prostar 210 pumps, a Prostar 335 Diode UV detector and a Phenomenex Synergi 4 m Hydro RP 80A C18
2
50 x 4.6 mm column using a 20-min gradient elution of 5-95% solvent B at 1 mL/min followed by 5 min
at 95% solvent B (solvent A, water with 0.1% TFA; solvent B, acetonitrile with 0.1% TFA and 5% water;
absorbance monitored at 220 and 254 nm).
General Procedure A: Sulfonamides and Amides of 8 from Sulfonyl Chlorides and Acid Chlorides
Respectively. To a solution of 8 (0.2 mmol, 1 equiv) in CH Cl (2 ml) was added DMAP (0.02 mmol, 0.1
2
2
equiv), followed by TEA (0.4 mmol, 2 equiv). The sulfonyl chloride or acid chloride (0.24 mmol, 1.2 equiv)
was added and the reaction stirred at rt under nitrogen for 16 h. The mixture was concentrated and then
chromatographed using a silica gel column with a 0-100% gradient of EtOAc/hexanes to provide the
purified sulfonamide or amide.
General Procedure B: Sulfonamides, Sulfamides and Amides of 8 from Sulfonyl Chlorides,
Sulfamoyl Chlorides and Acid Chlorides Respectively. To a solution of 8 (0.2 mmol, 1 equiv) in THF
(1 mL) was added the sulfonyl chloride, sulfamoyl chloride or acid chloride (0.24 mmol, 1.2 equiv),
followed by TEA (0.30 mmol, 1.5 equiv). The mixture was stirred at rt for 3 h. Water (0.4 mL), EtOAc (3
mL) and then saturated NaHCO solution (0.8 mL) were added. After 10 min, the aqueous layer was
3
removed. Celite (600 mg) was added to the organic layer and the solvent was evaporated. Flash
2
5

chromatography using silica gel with an EtOAc/hexanes gradient provided the purified sulfonamide,
sulfamide or amide.
General Procedure C: Amides of 8 from Carboxylic Acids. To a solution of 8 (0.2 mmol, 1 equiv) in
CH Cl (2 ml) were added TEA (0.6 mmol, 3 equiv), N,N,N′,N′-tetramethyl-O-(1H-benzotriazol-1-
2
2
yl)uronium hexafluorophosphate (HBTU, 0.26 mmol, 1.3 equiv) and the appropriate carboxylic acid (1.25
equiv). The reaction was stirred overnight and then concentrated. Flash chromatography, using a silica gel
column with a gradient of 0−100% EtOAc/hexanes, provided the purified amide.
General Procedure D: Carbamates of 8 from Carbamoyl Chlorides. To an ice-cold solution of 8 (0.2
mmol, 1 equiv) in 1,2-dichloroethane (1 mL) was added TEA (0.24 mmol, 1.2 equiv), followed by slow
addition of the chloroformate (0.24 mmol, 1.2 equiv). The ice bath was removed and stirring continued for
2
h. Ethyl acetate (3 mL) was added, followed by saturated NaHCO solution (0.8 mL) and water (0.4 mL).
3
After 10 minutes, the aqueous layer was removed. Celite (600 mg) was added to the organic layer and the
solvent evaporated. Flash chromatography using silica gel with an EtOAc/hexanes gradient provided the
purified carbamate.
General Procedure E: Carbamates of 8a from Alcohols. To an ice-cold solution of the alcohol (0.3
mmol, 1.5 equiv) in THF (1 mL) was slowly added NaHMDS (1 M/THF, 0.24 mL, 1.2 equiv). After 10
min, the 4-fluorophenyl carbamate of 8a (0.2 mmol; prepared using general procedure C) was added. After
1
5 min, the ice bath was removed and the mixture stirred at rt for 15 h. Ethyl acetate (3 mL) was added,
followed by brine (1 mL) and 2 N NaOH (0.4 mL). After 10 min, the aqueous layer was removed and the
organic layer was washed with 0.8 M NaHCO solution (2x1 mL). Celite (600 mg) was added to the organic
3
layer and the solvent evaporated. Flash chromatography using silica gel with an EtOAc/hexanes gradient
provided the purified carbamate.
General Procedure F: Amines of 8a or 9 from Aldehydes. To a solution of 8a or 9 (0.2 mmol, 1 equiv)
in MeOH (2.0 ml) was added the appropriate aldehyde (0.25 mmol, 1.25 equiv), then sodium
2
6

cyanoborohydride (0.4 mmol, 2 equiv), and lastly acetic acid (0.1 mL). The mixture was stirred at rt
overnight, concentrated and then dissolved in EtOAc (2 mL). The organic solution was washed with
saturated aqueous NaHCO (1 mL), washed with brine (1 mL), dried with Na SO (20 mins), filtered and
3
2
4
then concentrated. The resulting residue was chromatographed with a 4 g silica gel column using a 0-100%
EtOAc/hexanes gradient.
General Procedure G: Heteroaryl Amines of 8a from Heteroaryl Fluorides. A 5 ml sealed tube
containing 8a (0.2 mmol, 1 equiv) and the appropriate heteroaryl fluoride (5 mmol, 25 equiv) was sparged
with vacuum and nitrogen and then heated to 125 °C for 48 h in a dry heating block. The resulting residue
was flash chromatographed with a 12 g silica gel column using a 0-100% EtOAc/hexanes gradient.
General Procedure H: Preparation of 7 from Aryl Amines. To a heterogeneous mixture of 5-amino-
4
,6-dichloropyrimidine (1.6 g, 10 mmol) and an aryl amine (10 mmol, 1 equiv) in n-BuOH (20 mL) was
added 6 N HCl (0.84 mL, 0.5 equiv). The mixture was heated at 100 °C for 15 h. At rt, EtOAc (20 mL) and
N NaOH (8 mL) were added and the mixture stirred vigorously for 15 min. Additional EtOAc (80 mL)
2
and saturated NaHCO solution (40 mL) were added. The aqueous layer was removed and the organic layer
3
washed with 0.8 M NaHCO solution (25 mL). Celite (10 g) and toluene (10 mL) were added to the organic
3
layer and most of the solvent evaporated. Purification by flash chromatography, using EtOAc/hexanes,
provided the intermediate 5-amino-4-arylamino-6-chloropyrimidine. To a solution of the intermediate (10
mmol, 1 equiv) and 2-chlorobenzaldehyde (2.3 mL, 2 equiv) in dioxane (30 mL) was added 15%
FeCl /silica gel (6.0 g, 600 mg/mmol). The mixture was stirred at rt for 10 min and then at 95 °C for 20 h.
3
At rt, the mixture was filtered using a sintered glass funnel (sand on sodium sulfate on celite with a medium
frit) and washed with chloroform (30 mL). Toluene (3 mL) was added and most of the solvent evaporated.
The resulting residue was dissolved in CH Cl (30 mL) and cooled in an ice bath. DDQ (2.2 g, 1 equiv) was
2
2
added and after 10 min, the ice bath was removed. The mixture was stirred at rt for 2 h and then most of
the solvent was evaporated. Ethyl acetate (100 mL) was added and the resulting organic solution washed
2
7

with water (2x50 mL). Celite (10 g) was added and the solvent evaporated. Flash chromatography using
silica gel with an EtOAc/hexanes gradient provided purified 7.
General Procedure I: Preparation of 8 or 9 from 7. To a solution of 7 (10 mmol, 1 equiv) and 4-N-boc-
aminopiperidine (2.4 g, 1.2 equiv; for intermediates 8) or 4-N-boc-methylaminopiperidine (2.5 g, 1.2 equiv;
for intermediate 9) in NMP (20 mL) was added potassium carbonate (4.1 g, 3 equiv). The mixture was
stirred at rt for 15 min and then at 80 °C for 15 h. Ethyl acetate (100 mL) was added, followed by brine (40
mL) and water (20 mL). The organic layer was washed with brine/water (3/1, 2x30 mL). Celite (10 g) was
added to the organic layer and the solvent evaporated. Flash chromatography using silica gel with an
EtOAc/hexanes gradient provided the intermediate 9-aryl-8-(2-chlorophenyl)-6-(4-N-boc-aminopiperidin-
1
-yl)-9H-purine (for intermediates 8) or 8-(2-chlorophenyl)-9-(4-chlorophenyl)-6-(4-N-boc-
methylaminopiperidin-1-yl)-9H-purine (for intermediate 9). To a mixture of the intermediate (1 mmol, 1
equiv) and ethanol (2 mL) was added 6 N HCl (1 mL). The mixture was stirred at rt for 10 min and then at
5
0 °C for 3 h. The mixture was cooled in an ice bath and chloroform (6 mL) was added, followed by brine
(2 mL). 6 N NaOH (1.1 mL) was added slowly and after 5 min, the bath was removed. After 10 min, the
layers were separated and the aqueous layer was saturated with sodium chloride. The aqueous layer was
extracted with chloroform (2x2 mL). The combined organic layers was dried (sodium sulfate for 20 min),
filtered and evaporated. Toluene (1 mL) was added and the solvent evaporated to provide 8 or 9.
1
-[8-(2-Chlorophenyl)-9-(4-chlorophenyl)-9H-purin-6-yl]-N-phenylpiperidin-4-amine (11). The title
compound was prepared by the general procedure F to provide 129 mg (100%) of a solid, mp 103 °C (dec).
1
H NMR (300 MHz, CDCl ) δ 8.39 (s, 1H), 7.46-7.55 (m, 1H), 7.30-7.44 (m, 5H), 7.12-7.23 (m, 4H), 6.71
3
(t, J = 7.4 Hz, 1H), 6.65 (d, J = 7.7 Hz, 2H), 5.40 (br s, 2H), 3.59-3.80 (m, 1H), 3.56 (br s, 1H), 3.44 (t, J =
1
3
1
1
1
1.9 Hz, 2H), 2.26 (d, J = 10.5 Hz, 2H), 1.43-1.69 (m, 3H). C NMR (75 MHz, CDCl ) δ 154.0, 153.2,
3
52.3, 146.8, 145.7, 134.3, 134.3, 133.1, 132.5, 131.4, 130.1, 129.7, 129.4, 128.0, 127.0, 119.8, 117.6,
13.4, 50.4, 44.4, 32.8. MS (m/z) 515.6 (M+1). HPLC = 99% at 17.32 min.
2
8

N-{1-[8-(2-Chlorophenyl)-9-(4-chlorophenyl)-9H-purin-6-yl]piperidin-4-yl}pyridin-2-amine
(12).
The title compound was prepared by the general procedure G to provide 73 mg (71%) of an off-white solid.
1
H NMR (300 MHz, CDCl ) δ 8.39 (s, 1H), 8.10 (dd, J = 1.1, 5.1 Hz, 1H), 7.47-7.56 (m, 1H), 7.29-7.46
3
(
(
(
m, 6H), 7.14-7.25 (m, 2H), 6.57 (dd, J = 5.5, 6.7 Hz, 1H), 6.40 (d, J = 8.4 Hz, 1H), 5.41 (br s, 2H), 4.38
d, J = 7.9 Hz, 1H), 3.90-4.09 (m, 1H), 3.47 (t, J = 12.0 Hz, 2H), 2.26 (dd, J = 3.1, 12.8 Hz, 2H), 1.44-1.79
1
3
m, 3H). C NMR (75 MHz, CDCl ) δ 157.7, 153.9, 153.2, 152.3, 148.2, 145.6, 137.4, 134.2, 134.2, 133.1,
3
1
32.5, 131.4, 130.0, 129.7, 129.4, 128.0, 126.9, 119.8, 112.9, 107.5, 48.5, 44.4, 32.7. MS (m/z) 516.4
(M+1). HPLC = >99% at 17.43 min.
N-{1-[8-(2-Chlorophenyl)-9-(4-chlorophenyl)-9H-purin-6-yl]piperidin-4-yl}pyrimidin-2-amine (13).
The title compound was prepared by the general procedure G to provide 28 mg (21%) of a solid, mp 221-
1
2
23 °C. H NMR (300 MHz, CDCl ) δ 8.39 (s, 1H), 8.29 (d, J = 4.8 Hz, 2H), 7.46-7.57 (m, 1H), 7.29-7.43
3
(m, 5H), 7.15-7.25 (m, 2H), 6.54 (t, J = 4.8 Hz, 1H), 5.43 (br s, 2H), 5.18 (d, J = 7.9 Hz, 1H), 4.15-4.32
1
3
(
m, 1H), 3.45 (t, J = 11.7 Hz, 2H), 2.17-2.34 (m, 2H), 1.49-1.70 (m, 2H). C NMR (75 MHz, CDCl ) δ
3
1
1
61.7, 158.1, 153.9, 153.2, 152.3, 145.6, 134.2, 133.1, 132.5, 131.4, 130.0, 129.7, 129.4, 128.0, 126.9,
19.7, 110.8, 48.3, 44.4, 32.5. MS (m/z) 517.5 (M+1). HPLC = >99% at 18.19 min.
N-{1-[8-(2-Chlorophenyl)-9-(4-chlorophenyl)-9H-purin-6-yl]piperidin-4-yl}pyridin-4-amine
(14).
The title compound was prepared by the general procedure G to provide 128 mg (98%) of a solid, mp 185-
1
1
7
86 °C. H NMR (300 MHz, CDCl ) δ 8.40 (s, 1H), 8.20 (d, J = 6.2 Hz, 2H), 7.50 (d, J = 6.8 Hz, 1H), 7.29-
3
.44 (m, 5H), 7.20 (d, J = 8.8 Hz, 2H), 6.48 (d, J = 6.3 Hz, 2H), 5.45 (d, J = 8.8 Hz, 2H), 4.25 (d, J = 7.7
1
3
Hz, 1H), 3.62-3.79 (m, 1H), 3.44 (t, J = 12.0 Hz, 2H), 2.15-2.30 (m, 3H), 1.48-1.69 (m, 2H). C NMR (75
MHz, CDCl ) δ 153.9, 153.2, 152.4, 152.2, 150.0, 145.8, 134.4, 134.2, 133.0, 132.5, 131.5, 130.1, 129.6,
3
1
29.4, 128.0, 127.0, 119.8, 107.9, 49.5, 44.2, 32.3. MS (m/z) 516.7 (M+1). HPLC = 99% at 17.90 min.
1
-[8-(2-Chlorophenyl)-9-(4-chlorophenyl)-9H-purin-6-yl]-N-benzyl-piperidin-4-amine (15). The title
1
compound was prepared by the general procedure F to provide 41 mg (39%) of an off-white solid. H NMR
2
9