Synthesis and SAR of potent and selective tetrahydropyrazinoisoquinolinone 5-HT2C receptor agonists

Article abstract of DOI:10.1016/j.bmcl.2013.04.061

The 5-HT_2C receptor has been implicated as a critical regulator of appetite. Small molecule activation of the 5-HT_2C receptor has been shown to affect food intake and regulate body weight gain in rodent models and more recently in hu

Full text of DOI:10.1016/j.bmcl.2013.04.061

Bioorganic & Medicinal Chemistry Letters 23 (2013) 3914–3919
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Synthesis and SAR of potent and selective
tetrahydropyrazinoisoquinolinone 5-HT2C receptor agonists
⇑
Guohua Zhao , Chet Kwon, Sharon N. Bisaha, Philip D. Stein, Karen A. Rossi, Xueying Cao, Thao Ung,
Ginger Wu, Chen-Pin Hung, Sarah E. Malmstrom, Ge Zhang, Qinling Qu, Jinping Gan, William J. Keim,
Mary Jane Cullen, Kenneth W. Rohrbach, James Devenny, Mary Ann Pelleymounter, Keith J. Miller,
Jeffrey A. Robl
Bristol-Myers Squibb Research and Development, PO Box 5400, Princeton, NJ 08543-5400, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
The 5-HT2C receptor has been implicated as a critical regulator of appetite. Small molecule activation of
the 5-HT2C receptor has been shown to affect food intake and regulate body weight gain in rodent models
and more recently in human clinical trials. Therefore, 5-HT2C is a well validated target for anti-obesity
therapy. The synthesis and structure–activity relationships of a series of novel tetrahydropyrazinoiso-
quinolinone 5-HT2C receptor agonists are presented. Several members of this series were identified as
potent 5-HT2C receptor agonists with high functional selectivity against the 5-HT2A and 5-HT2B receptors
and reduced food intake in an acute rat feeding model upon oral dosing.
Received 13 March 2013
Revised 17 April 2013
Accepted 22 April 2013
Available online 29 April 2013
Keywords:
Tetrahydropyrazinoisoquinolinone
Serotonin
Ó 2013 Elsevier Ltd. All rights reserved.
5-HT2C receptor agonist
Obesity
5-HT2A receptor
5-HT2B receptor
Food intake reduction
Serotonin (5-HT) plays an integral role in a broad range of
cardiovascular, metabolic, and central nervous system pharmaco-
logical pathways.1 Of the 14 known 5-HT receptor subtypes, the
5-HT2C receptor in particular has been implicated as a critical reg-
ulator of appetite. 5-HT2C knockout mice are obese, hyperphagic,
hyperinsulinemic and are insensitive to the action of 5-HT2C ago-
nists.2 The 5-HT2C receptor is a well validated target for anti-obes-
ity therapy since activation of this receptor has been shown to
affect food intake and regulate body weight gain in rodent models3
and human clinical trials.4 A challenge of this target has been to
identify a potent 5-HT2C receptor agonist with high selectivity ver-
sus other 5-HT receptors, primarily the 5-HT2A and 5-HT2B recep-
tors, to avoid side effects such as hallucinogensis and valvular
hypertrophy disease.5 The latter finding led to the withdrawal of
the non-selective 5-HT2C agonists fenfluramine and the (S)-enan-
tiomer dexfenfluramine from the market.6 Recently, the modestly
selective 5-HT2C receptor agonist lorcaserin (1)7 (Fig. 1) has re-
ceived FDA approval for treatment of obesity on the basis of show-
ing statistically significant weight loss in several phase 3 trials.8
Numerous 5-HT2C receptor agonists have been reported in the
literature over the last two decades.9 A common structural motif
in 5-HT2C receptor agonist design has been the presence of a basic
amine attached to a phenyl ring through a linkage (replicating the
indole and primary amine pharmacophore of 5-HT), in which
length and orientation of the groups ultimately contribute to the
potency and selectivity of the compound. Our efforts were focused
on identification of novel structural motifs with potent 5-HT2C
agonism and high selectivity over both the 5-HT2A and 5-HT2B
receptors (preferably with no functional activity at 5-HT2B). Previ-
ously, we had reported that unsubstituted tetrahydropyrazinoiso-
indolone (2) was a moderately potent 5-HT2C agonist (5-HT2C
,
Ki = 630 nM, EC50 = 1500 nM) with ninefold functional selectivity
for the 5-HT2C receptor over the 5-HT2B receptor, and 14-fold over
the 5-HT2A receptor; moreover, substitution on the aryl ring,
NH
Cl
1
Lorcaserin ( )
5-HT2C EC50 = 9 nM (IA = 1)
5-HT2B EC50 = 943 nM (IA = 1)
5-HT2A EC50 = 168 nM (0.75)
⇑
Corresponding author. Tel.: +1 609 818 5553.
E-mail address: guohua.zhao@bms.com (G. Zhao).
Figure 1. FDA-approved selective 5-HT2C agonist as anti-obesity medication.
0960-894X/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmcl.2013.04.061

G. Zhao et al. / Bioorg. Med. Chem. Lett. 23 (2013) 3914–3919
3915
H
10
11
NH
1
R
H
CO2Me
R
CO2Me
NH2
a
9
8
NH
9
A
7
B
C
4
NH
N
N
3
O
6
7
O
O
3
5
2
H
N
R
5-HT2C Ki = 630 nM
5-HT2B Ki = 4540 nM
5-HT2A Ki = 8450 nM
CO2Me
b
c, d
CO2Me
O
7
R'
Figure 2. Known and proposed 5-HT2C agonists.
H
N
R
H
N
R
R
f, g
NH
NH
NH
NH
+
especially at the 7-position, enhanced 5-HT2C potency as much as
20-fold10 (Fig. 2). We envisioned that the conceptual expansion
of the central five-membered ring of the tetrahydropyrazinoisoin-
dolone core (2) to generate the closely related tetrahydropyrazino-
isoquinolinone core (3) could possibly lead to more potent 5-HT2C
agonists with better selectivity against the 5-HT2A and 5-HT2B
receptors. Our confidence in this hypothesis was bolstered by
overlaying low energy conformations of the (R)-tetrahydropyrazi-
noisoquinolinone core (22) onto the 7-trifluoromethyl tetrahydro-
pyrazinoisoindolone analog (4), the lead compound in the series10
(Fig. 3). This analysis predicted that (1) the R enantiomers of the
tetrahydropyrazinoisoquinolinones would be required for 5-HT2C
activity as was the case for the tetrahydropyrazinoisoindolone ser-
ies and (2) only a small (or no) 7-substutution might be required
for 5-HT2C potency and selectivity against the 5-HT2A and 5-HT2B
receptors. Our strategy involved sequential evaluation of the effect
of substitutions at each position on the tetrahydropyrazinoisoquin-
olinone core to enhance potency and selectivity prior to identifica-
tion of potent and selective agonists produced by optimal additive
effects from di-substitution at different positions.
Cl
α-isomer
H
R
O
9
R'
R'
O
10
R'
R'
Cl
e
N
H
N
H
N
R
O
8
R'
f, g
+
β-isomer
O
O
11
12
Scheme 1. Reagents and conditions: (a) triphosgene, ClCH2CH2Cl, then AlCl3, 0 °C to
reflux (40–50%); (b) NaH, R0BrCHCO2Me, DMF, rt to 40 °C (61–75%); (c) LiBH4, THF,
reflux (89–96%); (d) SOCl2, CH3Cl (69–78%); (e) BnNH2, K2CO3, diglyme (52–67%,
a
:b = 2:1); (f) H2, 10% Pd/C, EtOAc, OD Chiral HPLC separation, OD column, 30%
MeOH/EtOH/70% heptane (39–44%).
tetrahydropyrazinoisoquinolinones with a:b = 2:1 ratio after sepa-
ration, which were respectively, subjected to debenzylation under
standard conditions (H2, Pd/C) and chiral separation to provide
compounds 9–12 in good yield. Absolute configuration was deter-
mined by X-ray crystallography of the (1S)-(+)-10-camphor-sul-
fonamide of the final product.
Having established the absolute configuration of the active
enantiomer to be R, we developed an enantioselective synthesis
for the analogs at other positions, which began with commercially
Herein we report the synthesis and structure–activity relation-
ship of a series of 5-HT2C receptor agonists by the optimization of
the tetrahydropyrazinoisoquinolinone core. All the tetrahydropy-
razinoisoquinolinone analogs were synthesized by two general
synthetic routes. The first route, based on the work of Singh,11
was developed for the synthesis of the analogs at C4 (Scheme 1)
available D-Boc-phenylalanines 13 (Scheme 2). Coupling of acids
beginning with readily available D-phenylalanine methyl esters 5,
13 with optically pure N-benzyl glycine or alanine ethyl ester
yielded amides 14. Boc-deprotection and subsequent thermal
cyclization gave diones 15. Carbamates 16 were formed by lithium
aluminium hydride reduction of diones 15 followed by treatment
with methyl chloroformate. P2O5/POCl3 catalyzed cyclization and
debenzylation under standard conditions (H2, Pd/C or CH3CHClO-
COCl/MeOH) provided the final products 17 in good yield. Where
which were acylated with triphosgene, followed by treatment with
alumnium trichloride, to give (R)-methyl tetrahydroisoquinolinone
carboxylic acid esters 6. Treatment with NaH concurrently pro-
moted racemization and alkylation with substituted ethyl bro-
moacetates to generate racemic diesters 7. Lithium borohydride
reduction of diesters 7 and subsequent reaction with thionyl chlo-
ride afforded dichlorides 8. Treatment with benzyl amine in the
presence of potassium carbonate yielded the corresponding benzyl
the
D-Boc-phenylalanines were not commercially available, as
was the case with some alkyl and cyano analogs in Table 2, the syn-
thesis was carried out from intermediate bromides 18. Alkylation
of bromides 18 under Negishi, Suzuki, or Stille conditions or copper
catalyzed cyanidation with CuCN, followed by debenzylation under
standard conditions (H2, Pd/C or CH3CHClOCOCl/MeOH), afforded
the corresponding alkyl and cyano analogs 17, respectively. Ana-
logue 21 was synthesized from intermediate methyl ether 19 gen-
erated using the procedure as described above. O-demethylation of
ether 19 with boron tribromide, followed by propargylation of the
resulting phenol, formed ether 20, which was cyclized to give the
final product 21 after debenzylation under acidic conditions. N-
Methyl analog 23 was generated under reductive amination condi-
tions from analog 22 (produced by either of the above two routes).
The primary goal of our research was to design a 5-HT2C agonist
with a Ki and EC50 of less than 25 nM which exhibits greater than
100-fold functional selectivity over the 5-HT2B and 5-HT2A recep-
tors. The preliminary binding and functional assay results for the
unsubstituted tetrahydropyrazinoisoquinolinone core were very
encouraging. As predicted by modeling, the 5-HT2C binding affinity
(24 nM Ki) of the R-enantiomer tetrahydropyrazinoisoquinolinone
NH
H
H
N
NH
N
O
CF3
O
22
4
Green
Blue
Figure 3. Low energy conformation overlay of known and proposed 5-HT2C
agonists.

3916
G. Zhao et al. / Bioorg. Med. Chem. Lett. 23 (2013) 3914–3919
O
tional agonist (EC50 = 36 nM) exhibiting functional selectivity for
O
H
H
R
R
H
R
CO2H
Bn
R'
Bn
R'
the 5-HT2C receptor over the 5-HT2B and 5-HT2A receptors (20-fold
functionally selective over the 5-HT2A receptor with no 5-HT2B
functional activity). The in vitro profile of compound 22 provided
a good starting point for our research.
In hopes of increasing the 5-HT2C potency and 5-HT2A selectiv-
ity of this chemotype, the SAR for methyl substituents was ex-
plored. Methylation of the saturated B and C rings at N2, C3, C11,
and C11a (23, 26, 27, 30, and 25) dramatically reduced the binding
affinity and functional activity at all three serotonin receptor sub-
b, c
N
a
N
NH
NH
EtO2C
*
*
HN
Boc
Boc
O
13
14
15
H
R
Bn
R'
d, e
N
f, g
N
O
*
H
N
MeO
16
R
NH
types as did substitution at C4 with an
a oriented methyl (28). In
*
R'
H
N
Br
h, g
contrast, substitution at C4 with a b-methyl group (29) yielded a
very potent 5-HT2C compound, albeit lacking the functional selec-
tivity versus the 5-HT2A and 5-HT2B receptors exhibited by 22.
The functional response of 5-HT2C, 5-HT2B and 5-HT2A receptors
to substitution of the A ring at C7, C8, C9 and C10 for a range of
substituents was investigated (Table 2). The finding that significant
attenuation in activity against all three receptors ensued upon
introduction of a cyano group at C7, C8, and C9 (37, 43, and 50)
suggested polar groups cannot be tolerated on the aryl ring. There-
fore the SAR focus shifted to non-polar functionalities: either elec-
tron-donating groups such as alkyl (Me, Et, and cyc-Pr) and alkoxy
(OMe, OEt, and OPr) or electron-withdrawing groups such as halo-
alkyl (CF3) and halogen (Cl). These subsituents also served to ex-
plore the effect of subtle steric interactions on functional activity
and selectivity. Though changes in binding affinity did not neces-
sary correlate with changes in functional activity, the goal re-
Bn
O
17
N
*
R
O
18
OMe
O
O
H
N
H
H
N
Bn
Bn
N
NH
N
i, j
k, l
N
O
O
O
21
19
20
H
H
m
NH
N
N
N
O
O
mained identification of a substituted analog with high 5-HT2C
no 5-HT2B and weak 5-HT2A functional activity.
,
23
22
Scheme 2. Reagents and conditions: (a) BnNHCꢀHR0CO2Et (R = H or Me), EDC, HOBT
or DMAP, CH2Cl2 (63–98%); (b) 4 N HCl, dioxane, aq satd Na2CO3 solution
basification; (c) ClCH2CH2Cl, 60 °C (91–95% for two steps); (d) LAH, THF, 0 °C to
reflux (92–96%); (e) MeOCOCl, pyridine (56–78%); (f) P2O5, POCl3, 100 °C (75–94%);
(g) H2, 10% Pd/C, MeOH (83–87%) or for aryl halides, CH3CHClOCOCl, ClCH2CH2Cl,
0 °C to reflux, then MeOH reflux (55–64%); (h) R2Zn, Pd(dppf)Cl2, THF, 65 °C, or R-
B(OH)2, Pd(PPh3)4, K2CO3, DME, H2O, reflux or R-Sn(n-Bu)3, Pd(PPh3)4, toluene,
reflux or CuCN, CuI, 1,3-dimethyl-2-imidazolidinone (50–90% for all the transfor-
mations); (i) BBr3, CH2Cl2, ꢁ78 °C to rt; (j) propargyl bromide, Bu4Br, NaOH, (98% for
two steps); (k) PhNEt2, 200 °C (96%); (l) H2, 10% Pd/C, concd HCl, MeOH (92%); (m)
CH2O, NaBH(OAc)3, CH2Cl2, 0 °C to rt (88%).
While substitution at C7 (31–36) often resulted in significant
increases in binding affinity versus 5-HT2B, these modifications
invariably ablated all functional activity for this receptor. C7 sub-
stitution also typically resulted in an increase in functional activity
for the 5-HT2C receptor, but this was accompanied by a commensu-
rately greater enhancement for 5-HT2A. In case of the ethyl analog
32, the net result was to cause the original 20-fold functional selec-
tivity between 5-HT2C and 5-HT2A of 22 to collapse to near unity.
The C7 methyl derivative 31 exhibited functional selectivity be-
tween that of the H (22) and Et (33) analogs. The consequences
of C8 substitution (38–42) were particularly unfavorable due to
the appearance of potent partial 5-HT2B agonist activity for all
but 40 (CF3). In addition, relative to 22, the functional selectivity
22 was 64-fold greater than that of the S-enantiomer 24. More
intriguingly, compound 22 was an extremely potent 5-HT2C func-
Table 1
Serotonin (5-HT) binding and functional activity of mono-substituted saturated ring (B and C ring) analogs
11a
11
B
R
1
NH
2
A
C
4
N
3
R
O
b
b
b
Compounda
R11a
5-HT2C (nM)
5-HT2B (nM)
Function EC50 (IA)
5-HT2A (nM)
Function EC50 (IA)
#
R
Binding Ki
Function EC50 (IA)
Binding Ki
Binding Ki
22
24
25
23
26
27
28
29
30
R-H
S-H
R-Me
R-H
R-H
R-H
R-H
R-H
R-H
H
H
H
2-Me
3-
3-b-Me
4- -Me
4-b-Me
11-b-Me
24
36 (1.0)
nd
213
>10000
nd
nd
>10000
>2500
>10000
>2500
37 (0.3)
271 (0.3)
326
nd
707 (1.0)
nd
nd
>10000
>2500
>10000
>2500
123 (1.4)
>10000
1780
3453
585
2000
399
2777
58
5710
3445
488
1340
681
1637
39
402
1358 (1.0)
193 (0.8)
1744 (0.7)
330 (1.0)
33 (1.0)
3 (1.0)
6933
1460
2960
1740
1081
81
a
-Me
a
319
56 (0.9)
3020
a
All compounds were enatiomerically pure and exhibited satisfactory analytical data.
Radioligand binding studies were conducted to determine the binding affinities (Ki values) of compounds for the human recombinant 5-HT2A, 5-HT2B, and 5-HT2C
b
receptors. Functional screenings were carried out in HEK293E cells expressing the human 5-HT2A, 5-HT2B, or 5-HT2C receptor. The reported intrinsic activity (IA) is relative to
receptor activation by serotonin at 3 M (defined as 1). nd = value not determined. See reference 10 for complete description of all in vitro and in vivo assays.
l

G. Zhao et al. / Bioorg. Med. Chem. Lett. 23 (2013) 3914–3919
3917
Table 2
Serotonin (5-HT) binding and functional activity of mono-substituted compounds
10
H
N
R
9
8
NH
7
O
b
b
b
Compounda
5-HT2C (nM)
5-HT2B (nM)
Function EC50 (IA)
5-HT2A (nM)
Function EC50 (IA)
#
R7
R8
R9
R10
Binding Ki
Function EC50 (IA)
Binding Ki
Binding Ki
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
21
Me
Et
cycPr
CF3
Cl
OMe
CN
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
Me
Me
Me
H
H
H
H
H
H
H
H
Me
Et
CF3
Cl
OMe
CN
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
Me
Et
Pr
CF3
Cl
OMe
CN
H
H
H
H
H
H
Me
Et
H
Me
-(CH2)3O-
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
9
8
80
33
5 (0.9)
13
2
6
47
>10000
>10000
>10000
>10000
>10000
>10000
>10000
12 (0.2)
28 (0.4)
>10000
9 (0.6)
110
24
73
64 (0.7)
22 (1.0)
109 (0.3)
107 (0.3)
381 (1.0)
193 (0.4)
>10000
53 (0.9)
37 (1.0)
106 (1.4)
97 (1.3)
47 (1.5)
>2500
>10000
>10000
>10000
>10000
>10000
>10000
>2500
14 (1.0)
64 (0.9)
16 (1.5)
22 (0.9)
12 (1.2)
320 (1.0)
8 (0.8)
84
57
8
120
1655
2710
322
111
108
85
203
1661
92
33
39
30
37
1760
143
56
139
510
32
47
340
54
4 (0.8)
18 (1.0)
10 (0.8)
3 (1.0)
115
nd
56 (0.4)
>2500
85
600 (1.0)
55 (0.9)
55 (1.0)
108 (1.0)
583 (1.3)
171 (1.0)
241 (1.0)
610 (0.9)
28 (1.3)
48 (1.3)
3337 (1.0)
12 (1.0)
25 (1.0)
70 (0.8)
14 (1.0)
92 (1.0)
6 (1.0)
4000
260
231
275
795
396
1716
6027
1010
1424
>10000
511
1034
1448
326
456
146
231
337
105
227
340
778
127
196
5816
118
930
5513
78
63
129
21
40
17
>10000
>10000
>10000
>10000
>10000
>10000
>2500
>10000
>10000
nd
>10000
>10000
>10000
>10000
2366 (0.4)
>10000
827 (0.7)
>10000
77
H
H
H
H
456
234
255
5931
114
269
1428
78
19
78
24
98
Me
Et
CF3
OMe
OEt
OPr
H
H
OMe
OMe
206 (0.8)
>3000 (0.2)
nd
1030 (1.0)
>10000
>10000
1046 (0.1)
>10000
148 (0.5)
921(0.9)
1379 (0.9)
52
235
26
77 (1.2)
16 (1.0)
217
21
H
H
a
See Table 1 legend.
See Table 1 legend.
b
of these analogs (5-HT2C vs 5-HT2A) was diminished despite often
robust increases in 5-HT2C functional activity (see analogs 39 and
42). In contrast, introduction of substituents at C9 (44–50) ap-
peared to abolish functional responses versus the 5-HT2A and 5-
HT2B receptors, although functional activity for the 5-HT2C receptor
seemed to vary based upon both electronic and steric factors. The
methyl and ethyl analogs 44 and 45 were identified as particularly
to explore multiple di-substitution patterns in hopes of maximiz-
ing 5-HT2C functional activity while while maximizing 5-HT2A
and 5-HT2B selectivity.
Representative di-substituted analogs obtained by combining
preferred functionalities at favored positions, that is, 7-methyl, 9-
methyl and ethyl, and 10-methoxy (57–60) are listed in Table 2.
The 7,9-dimethyl analog 57 exhibited enhanced 5-HT2C functional
potency (EC50 = 14 nM) and >75-fold 5-HT2A functional selectivity
with minimal agonist response (IA = 0.1). This compound was also
functional inactive versus 5-HT2B. This potency and selectivity pro-
file lies between that of the 7-methyl (31) and the 9-methyl (44)
analogs suggesting that an additive effect exists for this combina-
tion of substituents. In contrast, no comparable additive effect
was obtained for the 7-methyl, 9-ethyl analog 58. Additionally,
neither the 7-methyl,10-methoxy analog 59 nor the 9-methyl,10-
methoxy analog 60 improved 5-HT2C potency and selectivity over
the 5-HT2B and 5-HT2A receptors relative to that of their corre-
sponding mono-substituted counterparts, suggesting that adding
bulk from either of the two opposing positions of the benzene ring
is not preferred. Interestingly, internal cyclization of the methyl
and methoxy groups of 60 to generate the 9,10-pyrano counterpart
21 improved 5-HT2C functional potency (EC50 = 16 nM) and 5-HT2A
functional selectivity (86-fold) while maintaining 5-HT2B func-
tional inactivity.
functionally potent and selective 5-HT2C agonists (5-HT2C
,
EC50 = 55 nM; inactive for the 5-HT2B and 5-HT2A receptors); how-
ever, neither compound exhibited enhanced potency for the target
receptor as compared with lead 22.
The SAR response for C10 substituents (51–56) differed in that
steric effects were no longer the primary factor dictating functional
potency. The greatest increases for both 5-HT2C affinity and agonist
responses were observed with the methoxy 54 and ethoxy 55 sub-
stitutions (superior to the methyl (51) and ethyl (52) analogs).
Replacement of ethoxy with a propoxy group (56) resulted in only
a threefold decrease in functional activity (vs 5-HT2C). Alkoxy ana-
logs 54–56 exhibited significantly diminished 5-HT2A and 5-HT2B
functional activity. Among this set, compounds 54 (5-HT2C
,
EC50 = 12 nM; 5-HT2B, functionally inactive; 5-HT2A, 86-fold func-
tional selectivity) and 55 (5-HT2C, EC50 = 25 nM; inactive for both
the 5-HT2B and 5-HT2A receptors) were identified as potent and
selective 5-HT2C agonists.
Given the varied effect of mono-substitution of the A ring on
functional activity (pictorially summarized in Fig. 4), we decided
Abbreviated rat pharmacokinetic studies were conducted to
determine general exposure (Cmax and AUC (0–4 h)) and the

3918
G. Zhao et al. / Bioorg. Med. Chem. Lett. 23 (2013) 3914–3919
Table 4
Improves 5-HT2C functional potency
Retains 5-HT2B functional selectivity
Enhances 5-HT2A selectivity
Acute feeding studies in ratsa
Cmpd
22
Dose (mpk)
Pellets consumed (% change from vehicle)
1 h
2 h
4 h
8 h
12 h
20 h
Slightly attenuates 5-HT2C potency
Retains 5-HT2B functional selectivity
Enhances 5-HT2A functional selectivity
1
3
10
1
ꢁ100*
ꢁ100*
ꢁ100*
+14
ꢁ100*
ꢁ100*
ꢁ100*
ꢁ8
ꢁ67*
ꢁ49*
ꢁ70*
ꢁ97*
0
ꢁ18
ꢁ51*
ꢁ92*
ꢁ6
ꢁ14
ꢁ25
ꢁ73*
ꢁ2
OMe
ꢁ95*
ꢁ100*
+4
Me
Et
10
9
44
NH
A
B
C
3
10
1
ꢁ45
ꢁ24
ꢁ13
ꢁ54*
0
ꢁ12
ꢁ26*
0
ꢁ12
ꢁ17*
ꢁ4
ꢁ13
ꢁ11
ꢁ7
N
ꢁ98*
ꢁ45
ꢁ90*
ꢁ15
Improves 5-HT2C potency
7
Retains 5-HT2B functional selectivity
Decreases 5-HT2A selectivity
O
52
3
ꢁ30
ꢁ39
ꢁ23*
ꢁ66*
ꢁ78*
ꢁ9
ꢁ10*
ꢁ52*
ꢁ32*
ꢁ6
ꢁ10*
ꢁ45*
ꢁ35*
0
ꢁ4
Me
10
10
10
10
ꢁ100*
ꢁ100*
ꢁ38
ꢁ86*
ꢁ100*
ꢁ15
ꢁ29*
ꢁ32*
0
54
55
21
ꢁ56
ꢁ23
ꢁ20
ꢁ10
ꢁ10
ꢁ5
Preferred combinations:
additive effect?
7,9-
*
p <0.05.
a
Effects on feeding in an acute (20 h) rat operant model upon treatment with
7,10-
9,10-
selected compounds. Male Sprague–Dawley rats (n = 6) were dosed orally with test
compound or vehicle (14% propylene glycol, 1% tween, 85% water) 60 min prior to
the onset of the dark cycle. Pellets consumed for compound treated animals were
compared to that of vehicle treated animals in order to determine percent reduction
of food intake over a 20 h period. Data are shown for the 1, 2, 4, 8, 12 and 20 h
timepoints. See Ref.10 for a detailed description of all in vivo studies.
Figure 4. A SAR summary of mono-substitutions on aryl ring and additive effect
strategy by combinations.
relative plasma and brain exposures (brain/plasma ratio) (Table 3)
for several tetrahydropyrazinoisoquinolinones exhibiting potent
selective 5-HT2C agonist activity. Following oral administration
typically at 10 mg/kg, brain exposures relative to plasma (brain/
plasma ratio, 3.5–9.0) was invariably greater than unity although
the plasma exposures and brain/plasma ratios varied widely. The
lead 22 and alkyl analogs 40, 44, 45, and 52 all achieved acceptable
plasma exposures, however exposures of the alkoxy analogs 54, 55,
and 21 were comparatively diminished, likely as a result of re-
duced metabolic stability. Data for two closely related pairs (9-
methyl and ethyl, 44 and 45, 10-methoxy and ethoxy, 54 and 55)
suggests that increasing lipophilicity in this series increases CNS
penetration but reduces exposure.
Based on their potency, selectivity and pharmacokinetic expo-
sure, several of the most promising compounds were assessed as
anorectic agents in a 20 h rat operant feeding model. Compounds
were dosed 60 min prior to the onset of the dark cycle (the most
active time of feeding). The number of food pellets consumed rel-
ative to vehicle-treated animals was assessed at several time
points (1, 2, 4, 8, 12, and 20 h) to determine percent reduction in
feeding (Table 4). Alkoxy analogs 55 and 21 failed to display a sig-
nificant reduction in food intake at any time point, probably due to
poor pharmacokinetic exposure. The parent 22 and alkyl analogs
44 and 52 caused a dose dependent reduction in food intake with
a statistically significant reduction at 10 mpk. The bulk of the
reduction occurring at the earlier time points especially for the 1
and 3 mpk doses. Alkoxy analog 54 also reduced food intake at
10 mpk.
In conclusion, our studies reveal that the tetrahydropyrazino-
isoquinolinone core 22 is a potent 5-HT2C agonist with functional
selectivity against the 5-HT2B and 5-HT2A receptors. Systematic
substituent variation at different positions identified a set of
mono-substituted analogs as potent and selective 5-HT2C agonists
devoid of 5-HT2B and 5-HT2A activities. An additive effect can be
observed with appropriate substituent combinations: specific sub-
stituents attached at positions 7, 9 or 9, 10 improve 5-HT2C potency
and 5-HT2A selectivity while retaining 5-HT2B functional selectiv-
ity. In a rat acute feeding model, several compounds were found
to significantly reduce food intake in a dose-dependent manner.
Acknowledgment
We thank Dr. William N. Washburn for his suggestions during
the manuscript preparation.
References and notes
1. Jones, B. J.; Blackburn, T. P. Pharmacol. Biochem. Behav. 2002, 71, 555.
2. (a) Tecott, L. H.; Sun, L. M.; Akana, S. F.; Strack, A. M.; Lowenstein, D. H.;
Dallman, M. F.; Julius, D. Nature 1995, 374, 542; (b) Vickers, S. P.; Clifton, P. G.;
Dourish, C. T.; Tecott, L. H. Psychopharmacology 1999, 143, 309.
3. (a) Vickers, S. P.; Benwell, K. R.; Porter, R. H.; Bickerdike, M. J.; Kennett, G. A.;
Dourish, C. T. Br. J. Pharmacol. 2000, 130, 1305; (b) Largent, B. L.; Robichaud, A.
J.; Miller, K. J. Ann. Rep. Med. Chem. 2002, 37, 1.
4. (a) Walsh, A. E. S.; Smith, K. A.; Oldman, A. D.; Williams, C.; Goodall, E. M.;
Cowen, P. J. Psychopharmacology 1994, 116, 120; (b) Sargent, P. A.; Sharpley, A.
L.; Williams, C.; Goodall, E. M.; Cowen, P. J. Psychopharmacology 1997, 133, 309.
5. Smith, B. M.; Thomsen, W. J.; Grottick, A. J. Expert Opin. Investig. Drugs 2006, 15,
257.
Table 3
Rat pharmacokinetic data
Compound Dose
(mpk)
Cmax
(nM)
AUC 0–4 h
(nM h)
Brain/plasma
ratio
6. (a) Fitzgerald, L. W.; Burn, T. C.; Brown, B. S.; Patterson, J. P.; Corjay, M. H.;
Valentine, P. A.; Sun, J.-H.; Link, J. R.; Abbaszade, I.; Hollis, J. M.; Largent, B. L.;
Hartig, P. R.; Hollis, G. F.; Meunier, P. C.; Robichaud, A. J.; Robertson, D. W. Mol.
Pharmacol. 2000, 57, 75; (b) Rothman, R. B.; Baumann, M. H.; Savage, J. E.;
Rauser, L.; McBride, A.; Hufeisen, S. J.; Roth, B. L. Circulation 2000, 102, 2836.
7. (a) Thomsen, W. J.; Grottick, A. J.; Menzaghi, F.; ReyesSaldana, H.; Espitia, S.;
Yuskin, D.; Whelan, K.; Martin, M.; Morgan, M.; Chen, W.; AlShamma, H.;
Smith, B.; Chalmers, D.; Behan, D. J. Pharmacol. Exp. Ther. 2008, 325, 577; (b)
Smith, S. R.; Prosser, W. A.; Donahue, D. J.; Morgan, M. E.; Anderson, C. M.;
Shanahan, W. R. Obesity 2009, 17, 494; (c) Smith, B. M.; Smith, J. M.; Tsai, J. H.;
Schultz, J. A.; Gilson, C. A.; Estrada, S. A.; Chen, R. R.; Park, D. M.; Prieto, E. B.;
Gallardo, C. S.; Sengupta, D.; Dosa, P. I.; Covel, J. A.; Ren, A.; Webb, R. R.; Beeley,
N. R. A.; Martin, M.; Morgan, M.; Espitia, S.; Saldana, H. R.; Bjenning, C.;
Whelan, K. T.; Grottick, A. J.; Menzaghi, F.; Thomsen, W. J. J. Med. Chem. 2008,
51, 305.
22
40
44
45
52
54
55
21
3
10
10
10
10
10
10
10
1742
890
5954
2176
3618
766
4586
2576
17223
2779
9643
2015
51
4.3
9.0
3.5
8.0
3.9
4.2
6.8
4.6
26
272
582
PK studies were performed in Sprague–Dawley rats using a PO dosing solution of
15% propylene glycol, 1% tween, 84% water. Plasma sampling of compounds was
determined over a 4 h period except where noted. B/P ratio determined at end of
study.
8. http://arna.client.shareholder.com/releasedetail.cfm?ReleaseID=419009.

G. Zhao et al. / Bioorg. Med. Chem. Lett. 23 (2013) 3914–3919
3919
9. (a) Wacker, D. A.; Miller, K. J. Curr. Opin. Drug Disc. Devel. 2008, 11, 438; (b)
Bishop, M. J.; Nilsson, B. M. Expert Opin. Ther. Patents 2003, 13, 1691.
10. Wacker, D. A.; Varnes, J. G.; Malmstrom, S. E.; Cao, X.; Hung, C.-P.; Ung, T.; Wu,
G.; Zhang, G.; Zuvich, E.; Thomas, M. A.; Keim, W. J.; Cullen, M. J.; Rohrbach, K.
W.; Qu, Q.; Narayanan, R.; Rossi, K.; Janovitz, E.; LehmanMcKeeman, L.; Malley,
M. F.; Devenny, J.; Pelleymounter, M. A.; Miller, K. J.; Robl, J. A. J. Med. Chem.
2007, 50, 1365.
11. Singh, H.; Sharma, S.; Iyer, R. N.; Anand, N. Indian J. Chem. 1977, 15B, 70.