Design and synthesis of orally-active and selective azaindane 5HT2c agonist for the treatment of obesity

Article abstract of DOI:10.1016/j.bmcl.2009.10.112

Based on our original pyrazine hit, CP-0809101, novel conformationally-restricted 5HT2c receptor agonists with 2-piperazin-azaindane scaffold were designed. Synthesis and structure-activity relationship (SAR) studies are described with emphasis on optimization of the selectivity against 5HT2a and 5HT2b receptors with excellent 2c potency. Orally-active and selective compounds were identified with dose-responsive in vivo efficacy in our pre-clinical food intake model.

Full text of DOI:10.1016/j.bmcl.2009.10.112

Bioorganic & Medicinal Chemistry Letters 20 (2010) 266–271
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Design and synthesis of orally-active and selective azaindane 5HT2c agonist
for the treatment of obesity
b
b
b
b
Kevin K.-C. Liu a, , Peter Cornelius , Terrell A. Patterson , Yuan Zeng , Stephanie Santucci ,
Elizabeth Tomlinson b, Colleen Gibbons b, Tristan S. Maurer b, Ravi Marala b, Janice Brown b, Jimmy X. Kong b,
Eunsun Lee b, Wendy Werner b, Zane Wenzel b, Chandra Vage b
*
a Pfizer Global Research and Development, Chemistry Department, 10770 Science Center Drive, La Jolla, CA 92120, United States
b Pfizer Global Research and Development, Groton Laboratories, 558 Eastern Point Road, Groton, CT 06340, United States
a r t i c l e i n f o
a b s t r a c t
Article history:
Based on our original pyrazine hit, CP-0809101, novel conformationally-restricted 5HT2c receptor ago-
nists with 2-piperazin-azaindane scaffold were designed. Synthesis and structure–activity relationship
(SAR) studies are described with emphasis on optimization of the selectivity against 5HT2a and 5HT2b
receptors with excellent 2c potency. Orally-active and selective compounds were identified with dose–
responsive in vivo efficacy in our pre-clinical food intake model.
Received 25 September 2009
Revised 23 October 2009
Accepted 27 October 2009
Available online 30 October 2009
Ó 2009 Elsevier Ltd. All rights reserved.
Keywords:
5HT2c
Conformationally-restricted
Dose–responsive manner
The GPCR 5-HT2c subtype of serotonin receptors has received
considerable interest as a therapeutic target for the treatment of
a wide variety of conditions including obesity, anxiety, depression,
obsessive compulsive disorder, schizophrenia, migraine and
erectile dysfunction.1 The 5-HT2c receptor has been implicated in
the regulation of body weight in both rodents and humans. 5-
HT2c knockout mice are hyperphagic and gain excess weight
(predominantly adipose tissue) compared to wild-type controls.2
These mice are resistant to the anorectic effects of mCPP (a non-
selective 5-HT2c receptor agonist, 1) and partially resistant to the
anorectic effects of dexfenfluramine (a 5HT-reuptake inhibitor
and releaser). In humans, these and other serotonergic drugs
inhibit food intake and promote beneficial weight loss. For exam-
ple, the non-selective 5-HT2c agonists, mCPP and PNU-22394
(legacy Pharmacia-Upjohn, 2), have both caused significant weight
loss in short-term clinical trials (14–24 days).3 In addition, dexfen-
fluramine (ReduxÒ, AHP) produced sustained body weight
reduction of 10% after 1 year in a subset of patients. Similarly, a
5-HT2C selective agonist (i.e., BVT933, Biovitrum) was reported
to cause body weight reduction (2.2 kg in 4 wks) in a Phase II
study.4 Recently, another 5HT2c agonist, APD-356 (Lorcaserin (3),
Arena), has been shown to inhibit food intake and is currently in
phase III trials.5
N
NH
Cl
N
Cl
N
N
H3C
PNU22394 (2)
mCPP (1)
APD-356 (Lorcaserin, 3)
To find a safe 5HT2c agonist for the treatment of obesity, a key
hurdle is the selectivity against 5HT2a and 5HT2b receptor agonistic
activity. Since the withdrawal of dexfenfluramine and fenfluramine
from the market due to increased incidence of valvular heart disease
(VHD), a significant body of evidence has accumulated that links
activation of the 5-HT2b receptor with this pathology.6 Activation
of 5-HT2a receptors in humans has been associated with hallucina-
tions such as the effects of lysergic acid diethylamide (LSD).7
Our laboratory objectives were to identify a small molecule 5-
HT2c agonist with binding potency of <10 nM and binding selectiv-
ity of >100-fold over other CNS receptors. The compound will not
be an agonist of 5-HT2b or 5-HT2a receptors in humans, while it
may be an antagonist of either of these receptors. In addition, the
compound should display anti-obesity activity after oral dosing
without untoward side effects in rodent models.
Medicinal chemistry activities started with the HTS hit, com-
pound 4 (CP-0809101), which was found in CNS compound li-
braries. However, compounds in this series suffer metabolism-
based genotoxicity.8 In order to expand chemical diversity and
hopefully avoid the reported metabolism-based toxicity of this ser-
ies, we set out to look for different templates with the crucial
* Corresponding author.
E-mail address: kevin.k.liu@pfizer.com (K.K.-C. Liu).
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2009.10.112

K. K.-C. Liu et al. / Bioorg. Med. Chem. Lett. 20 (2010) 266–271
267
Table 1
Template hopping examples
Cl
NH
NH
NH
N
N
N
N
Cl
N
N
NH
N
N
O
N
N
6
N
5
7
8
S
O
N
S
616 (69%) 1
S
530 (74%) 127
660 (45%) 156
1200 (53%) 4
NH
N
NH
11
NH
Me
O
NH
Cl
N
O
N
N
N
N
N
Br
N
N
N
N
Bn
10
N
N
N
9
12
S
N
Antagonist >500
Antagonist 710
5000 (53%) 23
Cl
1360 (79%) 207
NH
NH
NH
N
N
N
N
N
S
N
NH
O
N
N
N
16
N
15
S
13
N
14
NH
Antagonist 410
187 (60%) 14
Antagonist 24
NH
990 (31%) 450
NH
NH
NH
N
N
N
N
S
Cl
N
N
S
N
N
19
17
N
N
N
20
18
N
O
N
Ph
164 (67%) 22
Cl
400 (81%) 113
Antagonist 540
1250 (55%) 378
Cl
NH
N
N
NH
NH
N
O
S
N
N
O
S
N
N
23
HN
N
N
21
22
1020 (91%) 755
2600 (43%) 28
Antagonist 126
Data are displaced as 5HT2c functional EC50 or IC50 (nM), (% agonism), and Ki (nM).11
N
n
O
N
N
N
N
O
N
conformationally-restricted carbocycles
N
4
carbocyclic
analogs
the original hit
(CP-0809101)
n = 0 or 1
Scheme 1.

268
K. K.-C. Liu et al. / Bioorg. Med. Chem. Lett. 20 (2010) 266–271
4
5
8
N
N
N
N
3
6
7
NH
O
NH
O
N
N
2
Cl
NH
Cl
25
Flipr-EC50(Max%) ki
2c: 10 nM (97%) 2c:1.5 nM
2a: 70 nM (93%) 2a:0.7 nM
2b: 397 nM (61%) NA
26
Flipr-EC50(Max%) ki
2c: 120 nM (95%)
2c:10 nM
2a: 578 nM (76%) 2a: 2 nM
2b: 998 nM (20%) NA
LE = 0.44
24
ki
Flipr-EC50(Max%)
2c: 25 nM (93%)
2a: 2202 nM (60%)
2b: antagonist
2c:45 nM
2a:21 nM
NA
N
N
N
N
N
N
NH
O
NH
NH
O
O
Cl
Cl
Cl
28
29
27
Flipr-EC50(Max%)
2c: 0.1 nM (93%)
2a: 236 nM (75%)
2b: 187 nM (50%)
ki
Flipr-EC50(Max%) ki
ki
Flipr-EC50(Max%)
2c: 22 nM (97%)
2a: 1311 nM (46%)
2b: 748 nM (21%)
2c:0.9 nM
2c: 0.1 nM (92%) 2c:0.7 nM
2a: 87 nM (87%) 2a:1 nM
2b: 300 nM (58%) NA
2c:20 nM
2a:41 nM
NA
2a:3.8 nM
NA
Scheme 2.
piperazine pharmacophore as exemplified in compound 1 and 4
with or without the aryl side chain. Table 1 outlines the templates
that we used in our template-hopping exercise which were unfor-
tunately confounded by 5HT2c binding/agonistic activity, selectiv-
ity against 5HT2a/2b agonism, selectivity against other CNS
receptors, structural alerts (thiophene etc.)9 or poor ligand effi-
cacy.10 After analysis the SAR data from this exercise we took the
initiative to make the carbocyclic analogs as shown in Scheme 1
to (1) restrict the conformation; (2) introduce bulkiness to the mol-
ecule; make compounds less flat, to mitigate the risk of genotoxi-
city and for better selectivity; and, (3) avoid structural alerts.9
We made cyclohexyl analogs first to test our hypotheses. Much
to our delight, compound 24, the parent compound without the
hydrophobic aryl side chain showed decent 5HT2c agonist activity
with Ki = 45 nM and EC50 = 25 nM as a full agonist. In addition, this
compound also has good selectivity against 5HT2a and 5HT2b
agonism; it is a weak 5HT2a partial agonist and it has no 5HT2b
agonistic activity (Scheme 2). With these exciting biological data
and excellent ligand efficacy10 of compound 24 (LE = 0.44), the
hydrophobic aryl side chain was introduced to the core structure
without hesitation. The 2-chlorophenoxyl group was introduced
to the C-8 position and enantiomers were separated by chiral HPLC
for biological profiling.12 The S enantiomer is more potent than its
corresponding R enantiomer as shown in Scheme 2, (25 vs 26).
Compound 25, with the S configuration at C-8 is a single-digit
nM 5HT2c binder with EC50 = 10 nM as a full agonist. However, it
is also a potent 5HT2a agonist. To explore the phenoxyl side chain
SAR, the corresponding cyclopentyl analogs, the azaindanes, were
made for comparison. Interestingly, the racemic azaindane, com-
pound 27 is an extremely potent 5HT2c full agonist with sub-nM
Ki and EC50. Unfortunately, the more potent S enantiomer, com-
pound 28, in 5HT2c is again a potent 5HT2a agonist (Scheme 2,
N
N
N
N
NH
NH
N
N
O
O
NH
F
F
Azaindane
F
F
28
29
Flipr-EC50(Max%)
Flipr-EC50(Max%)
ki
ki
2c: 1.4 nM (91%)
2a: 1500 nM (13%)
2b: 2700 nM (18%)
2c: 1 nM (96%)
2a: 1428 nM (18%)
2b: antagonist
2c:25 nM
2a:15 nM
2b:153 nM
2c:7 nM
2a:16 nM
2b:83 nM
Scheme 3. Azaindanes.

K. K.-C. Liu et al. / Bioorg. Med. Chem. Lett. 20 (2010) 266–271
269
vehicle
compound 30 (30mg/kg)
25.0
22.5
20.0
17.5
15.0
12.5
10.0
7.5
compound 31 (30mg/kg)
N
N
N
N
O
N
O
N
dose at t=0
Cl
F
5.0
2.5
dark Cycle
31
30
2c Ki: 0.6 nM
0.0
2c Ki:3nM
0
120 240 360 480 600 720 840
time (minutes)
vehicle
25.0
22.5
20.0
17.5
15.0
12.5
10.0
7.5
3mg/kg compound 28
10mg/kg compound 28
N
N
O
N
Cl
dose at t=0
5.0
2.5
0.0
28
2c Ki:0.7 nM
dark Cycle
0
120 240 360 480 600 720 840
time (minutes)
Scheme 4. Spontaneous food intake model.
28 vs 29). At this stage, efforts were focused on azaindanes to re-
duce 5HT2a/2b agonistic activity in this series while keeping the
excellent 5HT2c agonistic activity.
inhibition at moderate doses, shown with compounds 30 and 31.
Compound 28 also demonstrated anorectic activity in a dose–re-
sponse manner in this in vivo efficacy model.
SAR studies of the azaindane template resulted in the identifica-
tion of the benzyloxy group as a preferred replacement for the
phenoxy group. We were delighted to find that this replacement
reduced the 5HT2a/2b agonism dramatically as shown in Scheme
3. For example, compound 28 with the S-2,5-difluorobenzyloxy
side chain is a potent 5HT2c full agonist with almost no 5HT2a
and 5HT2b agonistic activity in vitro. Furthermore, and to our fa-
vor, the corresponding 2-(R)-methyl piperazine, compound 29, is
an even more potent 5HT2c agonist and a 5HT2b antagonist. The
lack of 5HT2a and 5HT2b agonism in these compounds was further
confirmed in our in vivo rat head twitch behavior13 and ex vivo rat
stomach fundus14 assays.
To understand the concentration-effect relationship for these
compounds, inhibition of food intake was plotted against Flipr
EC50 or Ki as shown in Scheme 5. Interestingly, much better corre-
lation was observed between food intake inhibition and the ratio
between free brain concentration and Ki instead of EC50; the per-
centage of food intake inhibition increases as the ratio of free brain
concentration/Ki increase (Scheme 5). This PK-PD correlation al-
lowed us to prioritize compounds based on PK data instead of more
labor-intensive and costly in vivo food intake animal studies.
Compound 32 was identified for further studies because of it
excellent in vitro 5HT2c potency and very weak in vitro 5HT2a
and 2b agonistic activity. In the event, when compound 32 was
tested in our in vivo or ex vivo models for 5HT2a and 2b activity
as mentioned earlier, the compound’s attributes were consistent
with those of a 5HT2a and 2b antagonist. More attributes are listed
To assess weight loss potential, these 5HT2c agonists were
dosed orally in our rodent spontaneous food intake model.11 As
shown in Scheme 4, these compounds showed robust food intake
Scheme 5. PK-PD model. Better correlation was observed with Ki.

270
K. K.-C. Liu et al. / Bioorg. Med. Chem. Lett. 20 (2010) 266–271
5HT2c
1.5
5HT2a
4.4
5HT2b
7.7
Binding, Ki (nM)
N
N
Flipr EC50 (nM)
(% max) agonism
3.4 (94) 630 (46) 566 (28)
O
F
NH
H3C
32
gamma-GTP (nM)
(%max)
1.9 (86)
agonist
IP3, EC50
agonism
NA
NA
62
>10 uM
4
Effect of compound 32
on spontaneous feeding
Vehicle
KB (nM) antagonism NA
30
25
20
15
10
5
Compound 32 30mg/kg
Compound 32 60mg/kg
Food intake @ 100mpk, p.o. -74-63% @ 2 hr
Exposure in free brain: 7.4 nM
Compound 32 100mg/kg
Compound 32 200mg/kg
dose at t=0
HLM (ml/min/kg)
RLM (ml/min/kg)
9
9
dark Cycle
0
0
120 240 360 480 600 720 840
Herg IC50
3900 nM
>30 uM
time (minutes)
Mechanism-Based
inhibition
Genotoxicity
Ames negative
Micronucleus negative
Scheme 6. Compound 32.
Table 2
Compound 32 PK data
Species
Compound 32 (0.3 mM) in microsomes
Compound 32 (1 mM) in hepatocytes
In vivo
T1/2 (min)
CLint (mL/min/kg)
Predicted CLh
T1/2 (min)
CLint (mL/min/kg)
Predicted CLh
CLn (mL/min/kg)
Rat
Dog
Monkey
Human
2.2
24
12
44
1500
59
350
20
67
24
36
10
4.7
64
53
400
41
43
60
20
22
4.0
90
34
8.6
—
170
5.1
Distribution.
Mean brain/plasma: 3.4.
CSF/plasma ratios: 0.015.
CSF/free plasma: 0.5.
Consistent with the potent centrally-mediated effects observed in rat pharmacology experiments.
in Scheme 6. Compound 32 also demonstrated anorectic activity in
our food intake model in a dose–responsive manner. It is selective
against other GPCR receptors as well. In addition, compound 32 has
good ADME properties with good human predictions for clearance
and t1/2 (Table 2).
The syntheses of azaindanes are outlined in Scheme 7. The syn-
thesis starts with enamine alkylation of cyclopentanone with acry-
lonitrile to give ketone nitrile compound 33 in 77% yield.15
Compound 33 is then cyclized under acidic condition in the pres-
ence of bromine to give bicyclic pyridinone compound 34 in a
moderate yield which is subsequently converted to chloropyridine
35 in good yield with POCl3. The N-oxide intermediate 36 is gener-
ated by oxidation of compound 35 with m-chloroperbenzoic acid.
Compound 36 is then treated with acetic anhydride at high tem-
perature to give rearranged acetate 37. The racemic acetate 37 is
submitted to chiral separation to give optical pure (R) and (S)-ace-
tates which are hydrolyzed to the corresponding (R) and (S)-alco-
hols (39 and 38). Alkylation or Mitsunobu reaction of compound
38 or 39 generates the corresponding chiral alkyl ethers or phen-
oxy ethers, respectively.16Finally the crucial piperazine moiety is
installed to the template by Pd-coupling under basic condition in
the presence of catalytic amount Davephos ligand.17
In summary, we employed the strategy of conformational
restriction to give us potent 5HT2c agonists and identified the ben-
zyloxy substituent to improve selectivity in a novel azaindane ser-
ies. Furthermore, orally-active compounds were identified in our
rodent food intake model in a dose–responsive manner. Finally,
compound 32 with good pharmacology and ADME was chosen
for further studies, and additional results with this compound will
be reported in due course.

K. K.-C. Liu et al. / Bioorg. Med. Chem. Lett. 20 (2010) 266–271
271
O
O
CN
a
c
b
Cl
N
O
N
H
35
33
34
Cl
Cl
N
HO
HO
f-g
d
e
38
Cl
N
O
Cl
N
AcO
36
37
N
Boc
N
39
R2
N
H
+
i
h
Cl
N
Cl
N
N
N
j
R1O
HO
R1O
NH
40
38
41
R2
Boc
N
R3
R2
+
k
N
H
+
i
OH
Cl
N
N
Cl
N
N
j
O
O
HO
NH
R2
42
43
39
R3
R3
Scheme 7. Synthesis of azaindane. Reagents and conditions: (a) acrylonitrile, pyrrolidine, toluene, dean-stark refluxing at 140 °C (77%); (b) bromine, acetic acid (60%); (c)
POCl3 (76%); (d) m-CPBA, CH2Cl2, 0 °C to rt (80%) ; (e) acetic anhydride, 110 °C (81%); (f) chiral HPLC separation; (g) K2CO3, MeOH, H2O, rt (98%); (h) NaH (1.1 equiv), TBAI, R1X
(X = Br or Cl), DMF; (i) Pd2dba3, Davephos, t-BuONa, toluene; (j) TFA, CH2Cl2; (k) DEAD, PPh3, toluene.
6. Loke, Y. K.; Derry, S.; Pritchard-Copley, A. BMC Clin. Pharmacol. 2002, 2, 6.
7. Nichlos, D. E. Pharmacol. Ther. 2004, 101, 131.
Acknowledgments
8. Kalgutkar, A. S.; Bauman, J. N.; McClure, K. F.; Aubrecht, J.; Cortina, S. R.;
Paralkar, J. Bioorg. Med. Chem. Lett. 2009, 19, 1559.
9. Blagg, J. Ann. Rep. Med. Chem. 2006, 41, 353.
10. Hopkins, A. L.; Groom, C. R.; Alex, A. Drug Discovery Today 2004, 9, 430.
11. For the details of our biological assays and the spontaneous food intake model,
please see: Chen, Hou.; Coffey, Steven Blair.; Lefker, Bruce Allen.; Liu, Kevin, K.-
C. WO 2006103511, 2006.
Thanks to Dr. Matthew Marx for assistance with this
manuscript.
References and notes
12. Single crystal x-ray was used to determine the absolute configuration of the
chiral compounds.
1. Hoyer, D.; Clarke, D. E.; Fozard, J. R. Pharmacol. Rev. 1994, 46, 157; Saxena, P. R.
Pharmacol. Rev. 1994, 66, 339.
13. Tizabi, Y.; Russell, L. T.; Johnson, M.; Darmani, N. A. Prog.
Neuropsychopharmacol. Biol. Psychiatry 2001, 25, 1445.
2. Nonogaki, K.; Strack, A. M.; Dallman, M. F.; Tecott, L. H. Nat. Med. 1998, 4, 1152.
3. Sargent, P. A.; Sharpley, A. L.; Willians, C.; Goodall, E. M.; Cowen, P. J.
Psychopharmacology 1997, 133, 309.
4. Halford, J. C.; Harrold, J. A.; Lawton, C. L.; Blundell, J. E. Curr. Drug Targets 2005,
6, 201.
5. Smith, S.; Prosser, W.; Donahue, D.; Anderson, C.; Shanahan, W. 66th Annu
Meet Sci Sess Am Diabetes Assoc (ADA) (June 9–13, Washington, DC) 2006,
Abst 344-OR.
14. Singh, P.; Kumar, R. J. Enzyme Inhib. 2001, 16, 491.
15. Castelijns, A.; Dielemans, Hubertus J. EP 564030, 1993.
16. Mitsunobu conditions gave us mainly the compounds with reversed
configuration, along with 5–10% of the other isomer.
17. Davephos: CAS number: 213697-53-, 2-(dicyclohexylphosphino)-20-(N,N-
dimethylamino)-biphenyl.