Synthesis and structure-activity relationship of 2-(aminoalkyl)-2,3,3a,8-tetrahydrodibenzo[c,f]isoxazolo[2,3-a]azepine derivatives: A novel series of 5-HT2A/2C receptor antagonists. Part 2

Article abstract of DOI:10.1016/S0960-894X(01)00722-3

Following the programme started at Janssen Research Foundation searching for 5-HT_2A/2C antagonists, we now report on the synthesis of a series of substituted 2-(Dimethylaminomethyl)-2,3,3a,8-tetrahydrodibenzo[c,f]isoxazolo[2,3-a] azepine derivatives. The 5-HT_2A, 5-HT_2C and H₁ receptor affinities as well as the mCPP antagonistic activity of the compounds synthesised is described.

Full text of DOI:10.1016/S0960-894X(01)00722-3

Bioorganic & Medicinal Chemistry Letters 12 (2002) 249–253
Synthesis and Structure–Activity Relationship of 2-(Aminoalkyl)-
2,3,3a,8-tetrahydrodibenzo[c,f ]isoxazolo[2,3-a]azepine Derivatives:
A Novel Series of 5-HT2A/2C Receptor Antagonists. Part 2
J. Ignacio Andres,a,* Jesus Alcazar,a Jose M. Alonso,a Adolfo Dıaz,a Javier Fernandez,a
Pilar Gil,a Laura Iturrino,a Encarna Matesanz,a Theo F. Meert,b
Anton Megensc and Victor K. Sipidod
aJanssen-Cilag, Basic Research Centre, Jarama s/n, 45007 Toledo, Spain
bCNS Department, Janssen Research Foundation, B2340 Beerse, Belgium
cGeneral in vivo Pharmacology, Janssen Research Foundation, B2340 Beerse, Belgium
dMedicinal Chemistry Department, Janssen Research Foundation, B2340 Beerse, Belgium
Received 27 July 2001; revised 18 October 2001; accepted 30 October 2001
Abstract—Following the programme started at Janssen Research Foundation searching for 5-HT2A/2C antagonists, we now report
on the synthesis of a series of substituted 2-(Dimethylaminomethyl)-2,3,3a,8-tetrahydrodibenzo[c,f]isoxazolo[2,3-a]azepine deriva-
tives. The 5-HT2A, 5-HT2C and H1 receptor affinities as well as the mCPP antagonistic activity of the compounds synthesised is
described. # 2002 Elsevier Science Ltd. All rights reserved.
Introduction
for both 5-HT2A and 5-HT2C receptors,10a but was
devoid of in vivo activity in our mCPP challenge test. In
our previous paper, we described a series of 2-(am-
inoalkyl)-2,3,3a,8-tetrahydrodibenzo[c,f]isoxazolo[2,3-a]-
azepine derivatives as novel 5-HT2A/2C antagonists, that
also showed high affinity for histamine-H1 receptors.11
One of those compounds, the dimethylaminomethyl
derivative 2, was a potent mCPP antagonist as shown in
our in vivo mCPP challenge test.5,11 We decided to
explore the influence of the presence of substituents at
different positions of the tetracyclic system, both for the
5-HT2A/2C receptor affinities as well as for the mCPP
antagonistic activity. We also wanted to study the
influence of those substituents on the antagonistic effect
on the H1 receptor, which may contribute to potential
sedative effects of the compounds.12 This paper
describes the synthesis and structure–activity relation-
ship of differently substituted 2-(dimethylaminomethyl)-
2,3,3a,8-tetrahydrodibenzo[c,f]isoxazolo[2,3-a]azepine
derivatives (3) (Fig. 1).10b
The neurotransmitter serotonin (5-HT) induces a vari-
ety of effects that are mediated by specific receptors,
which can be subdivided in seven families, 5-HT1ꢀ7. T he
5-HT2A receptor is widely distributed, both in periph-
eral tissues and in the CNS, while the 5-HT2C receptor
has been found only in the CNS.1 The 5-HT agonist m-
chlorophenylpiperazine (mCPP) is often used as a probe
in clinical studies to challenge brain 5-HTfunctions in
humans.2 In animals, mCPP also induces symptoms of
anxiety in various animal models.3ꢀ7 As this mCPP-
induced anxiety seems to be mediated via the 5-HT2C
receptor, it has been hypothesised that 5-HT2C antag-
onists might be potential drugs for the treatment of
anxiety and depression.8,9
In recent years, we started a programme at Janssen
Research Foundation searching for potent, centrally
active 5-HT2C receptor antagonists as potential anxio-
lytic/antidepressant agents. As a result of our synthesis
program, R95292 (1) was found to display high affinity
Chemistry
The synthesis of the target compounds was carried out
by cycloaddition of N-allyldimethylamine to differently
*Corresponding author. Fax: +34-925-232603; e-mail: jandres@jaces.
jnj.com
0960-894X/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(01)00722-3

´
J. I. Andres et al. / Bioorg. Med. Chem. Lett. 12 (2002) 249–253
250
Figure 1.
substituted 11H-dibenzo[b,e]azepine-5-oxides (morph-
anthridine-5-oxides, 9a–9t). The synthesis of the needed
tricyclic precursors, substituted 6,11-dihydro-5H-diben-
zo[b,e]azepines (5,6-dihydromorphanthridines), was
achieved mainly by two general methods, A or B,
depending on the availability of the required starting
materials.
desired target compounds 3a–g. Ring closure of N-for-
myl-2-(30-methylbenzyl)aniline (6a) yielded a mixture of
the two possible regioisomers 7a and 7b, which was used
as such in further synthesis steps and separated into its
single isomers 3a and 3b at the final step.
Method B (Scheme 2)
As we described in our previous article in this series we
had developed a new and efficient method for the
synthesis of 5,6-dihydromorphanthridine.11 That
method has been applied as well to the synthesis of dif-
ferently substituted 5,6-dihydromorphanthridines. This
new method has proved to be a better alternative than
the general methods previously described in litera-
ture,16ꢀ20 whenever the needed starting benzaldehydes
and 2-aminobenzylalcohols were easily available. Con-
densation of 2-aminobenzylalcohols 10 with benzalde-
hydes 11 smoothly gave the unreported benzoxazine
derivatives 12 in excellent yields. Reductive cleavage of
these compounds by means of NaBH4 in ethanol affor-
ded the corresponding N-benzyl open analogues 13,
which were cyclised to the desired 5,6-dihydro-
morphanthridines 8h–t in the presence of sulphuric acid
in good yields and after only three steps from the avail-
able starting materials. The next two reactions were
analogous to those described above in Method A,
yielding final compounds 3h–t. Cyclisation of 13k and
Method A (Scheme 1)
At first we essentially followed one of the methods
described in literature for the synthesis of morphan-
thridines.13 The starting substituted o-aminobenzophe-
nones, 4a and 4c–g, are commercially available or were
prepared by conventional methods described in litera-
ture, such as Friedel–Crafts type reaction of substituted
anilines with substituted benzonitriles.14 Wolf–Kishner
reduction of compounds 4 afforded the corresponding
substituted 2-benzylanilines 5a and 5c–g. Formylation
of the amino groups with formic acid, followed by ring
closure by the Bischler–Napieralski reaction afforded
the morphanthridines 7a–g (CAUTION: skin irritants).
Catalytic hydrogenation of compounds 7a–g yielded the
desired 5,6-dihydromorphanthridines 8a–g, which were
oxidised using 3-phenyl-2-(phenylsulfonyl)oxaziridine
(Davis’ reagent)15 to their corresponding N-oxides 9a–g.
Finally cycloaddition with N-allyldimethylamine, heat-
ing in a sealed tube using toluene as solvent, gave the
ꢁ
ꢁ
.
Scheme 1. Reagents and conditions: (i) BCl3, AlCl3, xylene/CH2Cl2, 120 C, overnight, 35–52%; (ii) NH2NH2 H2O, HOH2CCH2OH, KOH, 200 C,
12 h, 80–97%; (iii) HCOOH, reflux, 2 h, 75–85%; (iv) PPA, POCl3, xylene, 100 ꢁC, overnight, 57–66%; (v) H2, Pd/C (10%), CH3OH/THF, rt, 53–
90%; (vi) 3-phenyl-2-(phenylsulfonyl)oxaziridine, CH2Cl2, rt, 2 h, 82–93%; (vii) N-allyldimethylamine, toluene, 100 ꢁC, sealed tube, overnight, 46–
65%.

´
J. I. Andres et al. / Bioorg. Med. Chem. Lett. 12 (2002) 249–253
251
Scheme 2. Reagents and conditions: (i) ClCH2COOH (cat), CH3OH or (CH3)2CHOH, rt, overnight, 95–100%; (ii) NaBH4, CH3CH2OH, reflux, 1 h,
90–95%; (iii) H2SO4, CH2Cl2, ꢀ20 ꢁC, then rt, overnight, 65–86%; (iv) Davis’ reagent, CH2Cl2, rt, 1–2 h, 90–98%; (v) N-allyldimethylamine, tolu-
ene, 100 ꢁC, sealed tube, 16 h, 30–65%.
receptors expressed in CHO cells, using [3H]pyrilamine
as radioligand.21 The affinities of the compounds for
other serotonergic as well as dopaminergic and adre-
nergic receptors were also measured. The experiments to
measure the in vivo activity of the compounds in our
mCPP challenge test were performed in male Wistar
rats, weighing between 200 and 220 g, following the
method described by Meert and co-workers.5 The test
Scheme 3. Reagents and conditions: (i) PhSn(CH3)3, Pd(PPh3)4, tolu-
ene, reflux, 24 h, 42%.
compounds were administered subcutaneously or orally
or both, depending on the availability and solubility of
the compounds as well as on the preliminary activity
13m yielded both regioisomers 8k, 8l and 8m, 8n,
respectively. Both isomeric mixtures were used in fur-
ther steps and separated by preparative HPLC into their
single products 3k, 3l and 3m, 3n, respectively, after the
last reaction of this synthetic pathway.
found. Table 1 shows the affinities for the 5-HT2A, 5-
HT2C and H1 receptors of the synthesised compounds as
well as their activity in our mCPP challenge test.
As can be deduced from the data shown in Table 1, the
introduction of substituents at different positions of the
core tetracyclic structure 2 had remarkable influence on
the affinity for both 5-HT2 as well as for the H1 recep-
tors. Introduction of a halogen atom at 4 position of the
tetracycle resulted in complete loss of affinity for recep-
tors (3h). But the presence of a methyl or a tri-
fluoromethyl group at that same position gave
compounds with comparable receptor binding profile as
that of the unsubstituted lead 2. Even more, the 4-tri-
fluoromethyl derivative 3i showed the highest selectivity
for 5-HT2C receptor. On the other hand, the 5-tri-
fluoromethyl substituted compound completely lost the
affinity for all receptors, while the 5-fluoro derivative 3k
showed a quite similar receptor profile to 2. Introduc-
tion of a strong electron donor group such as methoxy
at that same position resulted in a significant decrease of
affinity for the 5-HT2A receptor. The effect of a fluorine
atom at the 7-, 9- and 10-positions on the binding affi-
nity is also noteworthy. While the 7-fluoro derivative 3l
showed a significant increase only in its H1 affinity, the
9-fluoro analogue 3q improved its affinity for the 5-
HT2C receptor as well, and the 10-fluoro compound 3c
showed increased affinity for the three receptors. Sur-
prisingly the 7-chloro analogue 3p was completely
devoid of in vitro affinity, while the 7-trifluoromethyl
The 11-phenyl derivative 3u was synthesised by Stille
coupling of the 11-bromo analogue 3f with phenyl-
trimethyltin in the presence of tetrakis(triphenylphos-
phine)palladium (0) (Scheme 3).
The cycloaddition of the substituted morphanthridine-
5-oxides with N-allyldimethylamine afforded in all cases
the cis isomers, and only traces of the corresponding
trans isomers were occasionally detected. The determi-
nation of the relative configuration of the chiral centres
was carried out by means of NOE difference experi-
ments on the protons of the isoxazolidine ring.
Biological Results and Discussion
The affinities of the compounds for the 5-HT2 receptors
were measured by means of radioligand binding studies
conducted with: (a) human cloned 5-HT2A receptor,
expressed in L929 cells using [125I]R91150 as radio-
ligand21 and (b) human cloned 5-HT2C receptor,
expressed in CHO cells using [3H]mesulergine as radio-
ligand.22 The experiments to measure the affinities for
the H1 receptor were conducted with human cloned H1

´
J. I. Andres et al. / Bioorg. Med. Chem. Lett. 12 (2002) 249–253
252
Table 1. 5-HT2A/2C and H1 affinities and activity in mCPP challenge test of 2-(dimethylaminomethyl)-2,3,3a,8-tetrahydrodibenzo[c,f]isoxazolo[2,3-
a]azepine derivatives
Compd
R1, R2
Synthetic
Method
5HT2A
pIC50
5HT2C
pIC50
H1
pIC50
mCPP (sc)
ED50 (mg/kg)
mCPP (po)
ED50 (mg/kg)
2
H, H
4-CH3
6-CH3
10-F
11-Cl
11-CH3
11-Br
12-Cl
4-F
7.64
7.32
7.98
8.16
8.46
8.51
n.t.
7.91
7.52
8.20
8.33
8.98
8.52
n.t.
8.17
7.93
7.80
8.89
8.63
8.23
n.t.
0.04
n.t.
n.t.
0.63
>2.5
2.5
2.5
0.63
2.5
2.5
>2.5
n.t.
>2.5
>2.5
>2.5
2.5
>2.5
>2.5
n.t.
>2.5
2.5
3a
3b
3c
3d
3e
3f
3g
3h
3i
A
A
A
A
A
A
A
B
B
B
B
B
B
B
B
B
B
B
B
B
n.t.
0.16
0.16
n.t.
7.49
<6
7.99
<6
7.81
<6
n.t.
>2.5
2.5
n.t.
>2.5
2.5
2.5
4-CF3
5-CH3O
5-F
7.17
6.52
7.27
7.50
<6
6.00
n.t.
<6
7.62
7.04
8.94
7.97
7.65
8.12
7.71
7.97
7.93
<6
6.36
n.t.
<6
8.30
8.41
8.94
8.21
7.95
7.32
7.84
8.42
8.64
<6
7.43
n.t.
<6
8.76
8.22
8.59
7.77
8.10
3j
3k
3l
7-F
3m
3n
3o
3p
3q
3r
3s
3t
5-CF3
7-CF3
6-Cl
7-Cl
9-F
9-CH3
11-F
5,11-diF
11-Ph
n.t.
>2.5
n.t.
n.t.
n.t.
>2.5
2.5
0.63
>2.5
0.04
0.04
>2.5
3u
n.t., not tested.
derivative 3n retained some H1 activity. The 9-methyl
substituted compound 3r presented a comparable beha-
viour to that of 3q.
but less potent after oral administration. The 11-chloro
analogue 3d and the 11-methyl derivative 3e showed
slightly less activity than 2 after subcutaneous adminis-
tration, but 3d was as active as 2 when orally adminis-
tered. In contrast, the 11-phenyl substituted compound
3u was not active in our challenge test. The reason why
some compounds which showed quite potent affinity for
the 5-HT2C receptor did not show high activity in our
mCPP in vivo model is not well understood. In this
respect it could be hypothesised that metabolic stability,
absorption and/or distribution of the compounds might
play a vital role. There are also substantial species dif-
ferences for 5-HT2 receptors, which might explain as
well the apparent discrepancies between the high bind-
ing affinity of some compounds and their lack of in vivo
activity.
The most interesting results were obtained with the
introduction of substituents at the 11-position of the
tetracyclic structure. The presence of a halogen atom or
a methyl group at that position resulted in compounds
with significant increase in 5-HT2 and H1 receptor
binding affinity. The 11-chloro derivative 3d showed a
slightly better selectivity for 5-HT2C receptor than the
corresponding 11-fluoro analogue 3s, while both com-
pounds showed a 10-fold increase in potency for 5-HT2
receptors when compared to the unsubstituted lead
compound 2. We also introduced a bulkier group in
that position such as phenyl, exemplified by compound
3u, but its in vitro profile was almost the same as the
one showed by 2. As can be deduced from the data of
compound 3t, the introduction of two halogen atoms in
the 5- and 11-positions proved to increase the affinity
for both 5-HT2 receptors, with a slight decrease of the
affinity for the H1 receptor.
Table 2 shows the binding affinity for several other ser-
otonergic, dopaminergic and adrenergic receptors of
compounds 3d, 3e, 3s and 3t which bear substituents in
the 11-position. It is noteworthy that the presence of
halogen atoms in that position enhanced the affinity for
the dopamine receptors, especially in the case of the 11-
chloro derivative 3d, and more moderately when the
halogen is a fluorine atom. However, when a second
fluorine atom was introduced in position 5, the affinity
for dopaminergic receptors dropped again, as can be
deduced from the binding affinities of compound 3t.
These combined data suggest that the activity found in
our mCPP challenge test is indeed due to their affinity
for the 5-HT2A/2C receptors, so that the anxiolytic
Although quite potent in vitro compounds were syn-
thesised, only those derivatives bearing some substituent
in the 11-position were found to show interesting in vivo
activity in our mCPP challenge test. The 5,11-difluoro
substituted compound 3t showed the same activity as
the unsubstituted reference compound 2, both sub-
cutaneously and orally, while the monosubstituted 11-
fluoro analogue 3s was equally active subcutaneously

´
J. I. Andres et al. / Bioorg. Med. Chem. Lett. 12 (2002) 249–253
253
Table 2. Other receptor binding affinities of compounds 3d, 3e, 3s, 3t, 2 and mianserin (pIC50 values)a
Compd
a1A
a2A
a2C
5-HT1A
5-HT1D
5-HT3
D2 (rat)
D2L
D3
D4
Mianserin
2
3d
3e
3s
3t
6.39
6.27
7.84
7.37
7.35
6.79
7.68
6.43
6.32
<6
6.30
<6
7.37
5.73
6.06
<6
5.93
<6
6.40
6.01
6.7
<6
5.80
<6
6.42
6.18
6.87
<6
6.52
<6
5.85
5.92
6.32
<6
<6
<6
5.45
6.71
8.05
<6
<6
n.t.
5.15
6.06
8.06
7.26
7.44
6.35
5.41
6.77
9.26
<6
7.85
<6
5.48
6.88
8.58
<6
7.84
<6
aAll assays were performed with human cloned receptors (D2 also with rat receptor) following standard procedures.
6. Whitton, P.; Curzon, G. Psychopharmacology 1990, 100,
138.
effects should be mediated by one of the 5-HT2 recep-
tors or by both of them.
7. Rodgers, R. J.; Cole, J. C.; Cobain, M. R.; Daly, P.; Doran,
P. J.; Eells, J. R.; Wallis, P. Behav. Pharmacol. 1992, 3, 621.
8. Bromidge, S. M.; Dabbs, S.; Davies, D. T.; Davies, S.;
Duckworth, D. M.; Forbes, I. T.; Gaster, L. M.; Ham, P.;
Jones, G. E.; King, F. D.; Mulholland, K. R.; Saunders, D. V.;
Wyman, P. A.; Blaney, F. E.; Clarke, S. E.; Blackburn, T. P.;
Holland, V.; Kennett, G. A.; Lightowler, S.; Middlemiss,
D. N.; Trail, B.; Riley, G. J.; Wood, M. D. J. Med. Chem.
2000, 43, 1123.
In conclusion, we have found that the introduction of
diverse substituents at different positions on the phenyl
rings of the 2,3,3a,8-tetrahydrodibenzo[c,f]isoxazolo[2,3-
a]azepine structure of 2 afforded several compounds
with potent 5-HT2A/2C as well as H1 receptors affinity.
Some of the 11-substituted derivatives were also orally
potent mCPP antagonists. Separation of compounds 3d
and 3t into their corresponding enantiomers as well as
further pharmacological results will be published else-
where.
9. Di Matteo, V.; Di Giovanni, G.; Esposito, E. CNS Drug
Rev. 2000, 6, 195.
10. (a) Fernandez-Gadea, F. J.; Sipido, V. K.; Andres-Gil, J.
I.; Meert, T. F. WO 9614321, 1996; Chem. Abstr. 1996, 125,
142792. (b) Sipido, V. K.; Fernandez-Gadea, F. J.; Andres-
Gil, J. I.; Meert, T. F.; Gil-Lopetegui, P. WO 9614320, 1996;
Chem. Abstr. 1996, 125, 142705.
Acknowledgements
11. Andres, J. I.; Alcazar, J.; Alonso, J. M.; Dıaz, A.;
Fernandez, J.; Gil, P.; Iturrino, L.; Matesanz, E.; Meert,
T. F.; Megens, A.; Sipido, V. K. Bioorg. Med. Chem. Lett.
2002, 12, 243.
12. Hall, H.; Oegren, S. O. Life Sci. 1984, 34, 597.
13. Wardrop, A. W. H.; Sainsbury, G. L.; Harrison, J. M.;
Inch, T. D. J. Chem. Soc., Perkin Trans 1 1976, 1279.
14. Sugasawa, T.; Toyoda, T.; Adachi, M.; Sasakura, K. J.
Am. Chem. Soc. 1978, 100, 4842.
15. Davis, F. A.; Stringer, O. D. J. Org. Chem. 1982, 42, 1774.
16. Protiva, M.; Borovicka, M.; Hach, V.; Votava, Z.; Sram-
kova, J.; Horakova, Z. Experientia 1957, 13, 291.
17. Werner, L. H.; Ricca, S.; Mohacsi, E.; Rossi, A.; Arya,
V. P. J. Med. Chem. 1965, 8, 74.
The authors would like to thank Ms. Valle Ancos,
Alcira Del Cerro, Rosario Dıaz, Celia Perez and Vic-
toria Perez for their experimental work performed. They
would also like to thank Mr. Willy Van Gorp for his help
in management and coordination of all data generated.
References and Notes
1. Hoyer, D.; Clarke, D. E.; Fozard, J. R.; Hartig, P. R.;
Martin, G. R.; Mylecharane, E. J.; Saxena, P. R.; Humphrey,
P. P. A. Pharmacol. Rev. 1994, 46, 157.
2. Sanders-Bush, E. and Mayer, S. E. In Goodman & Gilman’s
The Pharmacological Basis of Therapeutics, 9th ed.; Hardman,
J. G., Limbird, L. E., Molinoff, P. B., Ruddon, R. W., Good-
man A. Gilman, Eds.; McGraw-Hill: New York, 1996; p 249.
3. Gibson, E. L.; Barnfield, A. M. C.; Curzon, G. Neuro-
pharmacology 1994, 33, 457.
4. Kennett, G. A.; Whitton, P.; Shah, K.; Curzon, G. Eur. J.
Pharmacol. 1989, 164, 445.
5. Meert, T. F.; Melis, W.; Aerts, N.; Clincke, G. Behav.
Pharmacol. 1997, 8, 353.
18. Wardrop, A. W. H.; Sainsbury, G. L.; Harrison, J. M.;
Inch, T. D. J. Chem. Soc., Perkin Trans 1 1976, 1279.
19. Nagarajan, K.; David, J.; Bhat, G. A. Indian J. Chem.,
Sect. B 1985, 24B, 840.
20. Albright, J. D.; Venkatesan, A. R.; De los Santos, E. G.
WO 9622282, 1996; Chem. Abstr. 1996, 125, 221884.
21. Leysen, J. E.; Niemegeers, C. J.; Van Neuten, J. M.;
Laduron, P. M. Mol. Pharmacol. 1982, 21, 301.
22. Pazos, A.; Hoyer, D.; Palacios, J. M. Eur. J. Pharmacol.
1985, 106, 539.