Synthesis and Binding Studies of Trishomocubanes: Novel Ligands for ? Binding Sites

Article abstract of DOI:10.1071/CH99010

Five new analogues of the ?₂-selective trishomocubane N-(3'-fluorophenyl)methyl-4-azahexacyclo[5.4.1.0^{2,6_.03,10_.05,9_.08,11}]dodecan-3-ol (2) have been synthesized and assessed for their affinities at b

Full text of DOI:10.1071/CH99010

C S I R O P U B L I S H I N G
Australian Journal
of Chemistry
Volume 52, 1999
©
CSIRO Australia 1999
A journal for the publication of original research
in all branches of chemistry and chemical technology
w w w. p u b l i s h . c s i ro . a u / j o u rn a l s / a j c
All enquiries and manuscripts should be directed to
The Managing Editor
Australian Journal of Chemistry
CSIRO PUBLISHING
PO Box 1139 (150 Oxford St)
Collingwood
Vic. 3066
Australia
Telephone: 61 3 9662 7630
Facsimile: 61 3 9662 7611
Email: john.zdysiewicz@publish.csiro.au
Published by CSIRO PUBLISHING
for CSIRO Australia and
the Australian Academy of Science

Aust. J. Chem., 1999, 52, 653–656
.
Synthesis and Binding Studies of Trishomocubanes:
Novel Ligands for � Binding Sites
A
B
A
Xiang Liu, Michael Kassiou and MacDonald J. Christie
A
Department of Pharmacology, The University of Sydney, N.S.W. 2006.
Department of PET and Nuclear Medicine, Royal Prince Alfred Hospital,
B
Camperdown, N.S.W. 2050. Author to whom correspondence should be addressed.
0
Five new analogues of the �2-selective trishomocubane N -(3 -�uorophenyl)methyl-4-azahexacyclo-
[
2
,6 3,10 5,9 8,11
5.4.1.0 .0 .0 .0 ]dodecan-3-ol (2) have been synthesized and assessed for their a� nities at
both �1 and �2 sites. The structural requirements which in�uence � binding of functionalized
trishomocubanes were investigated with a view to develop more potent � ligands, with particular
emphasis on the �2 subtype. All synthesized compounds displayed moderate a� nity to � binding
sites. Binding at the �1 site ranged from 107 to 1100 nM while binding at the �2 site ranged from 135
to 250 nM. The subtype selectivity was in�uenced by the nature and position of substitution on the
aromatic ring. Fluorine substitution in the meta position, as in (2), is favoured over ortho, or para
or meta and para di�uoro substitution for �2 selectivity. Compounds (3)–(5) not only demonstrated
lower a� nity at the �2 site but a reversal of subtype selectivity with preferential binding at the �1
site (�1/�2: 0�8 for (3); 0�4 for (4); 0�8 for (6)). Introduction of CF3 and NO2 groups into the meta
position in compounds (6) and (7) retained �2 selectivity although to a lesser extent when compared
to (2) (�1/�2: 7�6 for (2); 2�0 for (6); 4�5 for (7)).
Introduction
uated in in vitro binding experiments for their neuro-
receptor selectivity and a� nity. Most of the members
in this series unexpectedly displayed moderate to high
a� nity for � binding sites with no cross reactivity
to other screened receptors and subtypes, including
dopamine, cholinergic, opioid, N -methyl-D-aspartate
(NMDA), phencyclidine and serotonin. Particularly,
Following two decades of intensive study, sigma (�)
binding sites have been de�ned as a class of high
a� nity and saturable binding sites present in both
1
,2
central nervous system and peripheral tissues. There
are at least two identi�ed subtypes, namely �1 and
3
,4
�
2 sites.
Sigma binding sites have been implicated
�
the aza hexacyclo compound (1) showed the highest
in many pharmacological functions, such as motor
5
,6
7,8
a� nity and subtype selectivity for the �1 site (Ki =
9�4 nM, �1/�2 = 19), whereas compound (2)† displayed
the highest a� nity for the �2 site (Ki = 19�6 nM,
e�ects,
sterol biosynthesis,
neuroprotection and
9
tumour biology. Recently, �2 binding sites have been
described as biomarkers of cellular proliferation in
14,15
1
0
�2/�1 = 8).
breast cancer. However, the functional importance
of � sites still remains elusive due to the lack of
selective � ligands as speci�c tools. A number of high
a� nity and selective �1 ligands have been synthesized,
whereas the �2 ligands to date have either low a� nity
or poor subtype selectivity or cross reactivity for other
1
1
binding sites.
Our interest in exploring the bioactivity of polycyclic
molecules, particularly trishomocubanes, was initiated
by the observation that several D3 trishomocubane
derivatives demonstrated promising anticataleptic activ-
ities in vivo presumably through an e�ect on dopamine
Herein we report the synthesis and binding studies
of new analogues of (2). This work aims at developing
more potent and selective �2 ligands.
1
2,13
14
neurotransmission.
In our earlier study,
20
trishomocubane analogues were synthesized and eval-
�
0
0
2,6 3,10 5,9 8,11
N -(4 -Phenylbutyl)-4-azahexacyclo[5.4.1.0 .0
.0 .0
]dodecan-3-ol.
2,6 3,10 5,9 8,11
†
N -(3 -Fluorophenyl)methyl-4-azahexacyclo[5.4.1.0 .0
.0 .0
]dodecan-3-ol.
Manuscript received 23 January 1999
᭧ CSIRO 1999
10.1071/CH99010
0004-9425/99/070653$ 05.00

6
54
X. Liu, M. Kassiou and M. J. Christie
Results and Discussion
The results are presented in Table 1 as inhibition
constants (K ). Compounds (3) and (4) with �uorine
i
The general synthetic route used for the prepa-
ration of functionalized trishomocubane analogues is
outlined in Scheme 1. The Cookson diketone (8)
was reacted with ethylene glycol in the presence of
catalytic amount of p-toluenesulfonic acid to yield the
monoketal (9). Compound (9) was used as the general
starting material for all new compounds. When the
monoketal (9) was heated with appropriately substi-
substitution in the ortho and para position respectively
resulted in weaker �2 binding and a reversal of subtype
selectivity when compared to (2) with preferential
binding at the �1 site (�1/�2: 7�6 for (2); 0�8 for (3);
0
�4 for (4)). Similarly the m,p-di�uoro-substituted
compound (5) negated the �2 selectivity a�orded by a
single meta �uorine substitution as seen with compound
(2). Interestingly, all �uorinated compounds (3)–(5)
showed a similar 10-fold decrease in a� nity for the �2
site while their a� nity for the �1 site was retained
at the same order of magnitude when compared with
compound (2).
�
tuted benzylamines in ethanol at 100 C overnight,
it a�orded the intermediate imines which were then
treated with sodium borohydride without isolation to
give the pentacycloundecylamines (10)–(14). Hydrol-
ysis of the ketal function with aqueous HCl yielded
the corresponding azahexacyclododecanes (3)–(7) as
Table 1. In vitro binding results of trishomocubane analogues
at �1 and �2 sites
15
a result of transannular cyclization. All compounds
1
were characterized by H n.m.r. spectroscopy, mass
spectrometry and elemental analysis. Evaluation of
the newly prepared compounds for their a� nity to �1
and �2 binding sites was measured in guinea pig brain
membranes according to the binding procedure given
in the Experimental section.
Com-
pound
Substituents
Y
Ki (nM)
�1 �2
�1/�2
ratio
X
Z
(
(
(
(
(
(
2)
3)
4)
5)
6)
7)
H
F
H
H
H
H
F
H
H
F
CF3
NO2
H
H
F
F
H
H
152
182
107
198
270
1100
20
7�6
0�8
0�4
0�8
2�0
4�5
230
250
239
135
242
To determine the in�uence of functional groups other
than halogens on � binding and subtype selectivity, we
introduced CF3 and NO2 groups in the meta position
on the aromatic ring, and obtained compounds (6)
and (7) respectively. These two compounds retained
�2 selectivity (�1/�2: 7�6 for (2); 2�0 for (6); 4�5
for (7)) even though their a� nity for the �2 binding
site was considerably less than that of (2), 7 times
lower for (6) and 12 times lower for (7) (�2: Ki =
20 nM for (2); Ki = 135 nM for (6); Ki = 242 nM
for (7)) (Table 1). A comparison of the data for �1
binding revealed that, although (7) has lower a� nity
for �2 binding sites when compared with (6), the nitro
compound (7) displayed a better �2 selectivity due to
its markedly lower a� nity for �1 sites. Compound
(2) still remains as the most potent and selective
trishomocubane analogue for binding at the �2 site.
Therefore, meta substitution in these compounds
has been found to be essential for enhanced binding
and selectivity at �2 sites. Also, �uorine substitution in
the meta position a�ords markedly higher a� nity and
subtype selectivity for �2 binding when compared with
other halogens including chlorine, bromine and iodine
Scheme 1. (i) Ethylene glycol; (ii) p-toluenesulfonic acid,
re�ux in benzene; (iii) substituted benzylamine, 100 C for 14
h; (iv) sodium borohydride, room temperature for 6 h; (v) 2
M HCl, room temperature overnight.
�
Our earlier work with functionalized trishomocubanes
revealed that a one-carbon alkyl chain between the
trishomocubane moiety and the aromatic ring was
essential for �2 binding. Increasing the alkyl chain to
two carbons resulted in a reversal of subtype selectivity.
In addition, �uorine substitution in the meta position
on the benzene ring enhanced selectivity and a� nity
14
for �2 binding sites. As a result, (2) displayed the
highest a� nity as well as the best selectivity among all
other compounds and was chosen as a potential lead
for development of �2-selective ligands. In order to
develop more potent �2 ligands the one-carbon alkyl
chain was retained and the e�ect of �uorine substitution
around the aromatic ring was investigated, along with
other meta substituents.
15
as we have described earlier. Fluorine substitution
is found to be more favourable when compared to
CF3 and NO2 groups for �2 binding. Compound
All newly synthesized compounds (3)–(7) displayed
moderate a� nities and varying � subtype selectivity.

Trishomocubane Ligands for � Binding Sites
655
tered and the solvent was removed by rotary evaporation.
The crude products were crystallized and recrystallized from
isopropyl alcohol (approx. 20 ml); this resulted in the desired
(
2) still appears as the most promising lead for the
development of more potent and selective �2 binding
ligands. Detailed structure–activity analyses are in
progress to evaluate parameters such as changes in
molecular conformation, steric in�uences and e�ects
of substitution electronegativity on �2 binding.
15
azahexacyclododecanes (3)–(7) (30–40%).
0
N-(2 -Fluorophenyl)methyl-4-azahexacyclo-
2
,6
3,10
5,9
8,11
[
5.4.1.0 .0
.0 .0
]dodecan-3-ol (3)
Treating the monoketal (9) with 2-�uorobenzylamine accord-
Experimental
ing to the method described above a�orded the title compound
�
as a white solid (30%), m.p. 164–165 C (Found: C, 75�9; H,
The reagents, including the Cookson diketone (8), ethy-
lene glycol, p-toluenesulfonic acid, sodium borohydride, 2-
6
�2; N, 4�9. C18H18FNO requires C, 76�3; H, 6�4; N, 4�9%).
1
Mass spectrum: c.i. m/z 284 (99%, M+1). H n.m.r. (300
MHz): � 1�5, 1H, d, J 10�5 Hz, CHCH2CH; 1�8, d, J 10�2
Hz, CHCH2CH; 2�4–3�0, 8H, m, CH; 3�3, 1H, t, J 4�7 Hz,
NCH; 3�5, 1H, d, J 14 Hz, NCH2; 3�8, 1H, d, J 13�7 Hz,
NCH2; 7�0, 1H, m, ArH; 7�1, 1H, m, ArH; 7�2, 1H, m, ArH;
�uorobenzylamine, 4-�uorobenzylamine, 3,4-di�uorobenzyl-
amine, 3-tri�uoromethylbenzylamine, and 3-nitrobenzylamine
hydrochloride salt, were purchased from Aldrich and used
with no further puri�cation. All synthesized compounds were
examined by thin-layer chromatography on silica and alumina.
Melting points were performed in a sealed capillary and are
7
�4, 1H, m, ArH.
1
uncorrected. Unless speci�ed otherwise H n.m.r. spectra were
0
N-(4 -Fluorophenyl)methyl-4-azahexacyclo-
recorded as solutions in CDCl3 at 300 MHz on a Varian instru-
ment. Elemental analyses were performed by the Department
of Chemical Engineering, The University of Sydney. Mass
spectra were recorded on a Finnigan/MAT TSQ46 system,
with chemical ionization (c.i.) being used, by the Department
of Pharmacy, The University of Sydney.
2
,6
3,10
5,9
8,11
[
5.4.1.0 .0
.0 .0
]dodecan-3-ol (4)
Treating the monoketal (9) with 4-�uorobenzylamine accord-
ing to the method described above a�orded the title compound
�
as a white solid (30%), m.p. 166–167 C (Found: C, 75�9; H,
6
�4; N, 5�0. C18H18FNO requires C, 76�3; H, 6�4; N, 4�9%).
3
3
�
1
[
H](+)-Pentazocine, [Ring-1,3- H] (58 Ci/mmol), and
Mass spectrum: c.i. m/z 284 (99%, M+1). H n.m.r. (300
MHz): � 1�5, 1H, d, J 10�4 Hz, CHCH2CH; 1�8, 1H, d, J
3
3
3
[
H]DTG, [5- H](1,3-di-o-tolylguanidine di-[p-Ring- H]) (35
Ci/mmol), were purchased from Dupont/New England Nuclear
Boston, MA, U.S.A.). (+)-Pentazocine and 1,3-di-o-tolyl-
1
0�4 Hz, CHCH2CH; 2�4–3�0, 8H, m, CH; 3�3, 1H, t, J 4�8
(
Hz, NCH; 3�4, 1H, d, J 13�5 Hz, NCH2; 3�75, 1H, d, J 13�5
guanidine (DTG) were purchased from Research Biochemi-
cals Inc. (Natick, MA, U.S.A.). Haloperidol was purchased
from Sigma Chemical Co. (St. Louis, MO, U.S.A.). The
scintillant used was Emulsi�er-Safe and was purchased from
Packard Instruments B.V.–Chemical Operations, Groningen,
The Netherlands. Whatman GF/B �lters were purchased
from Whatman. Binding assays were performed in polycarbon
P3 tubes purchased from John’s Medical Supplies. Liquid
scintillation spectrometry was carried out by using a Packard
Hz, NCH2; 7�0, 2H, m, ArH; 7�3, 2H, m, ArH.
0
0
N-(3 ,4 -Di�uorophenyl)methyl-4-azahexacyclo-
2
,6
3,10
5,9
8,11
[5.4.1.0 .0
.0 .0
]dodecan-3-ol (5)
Treating the monoketal (9) with 3,4-di�uorobenzylamine
according to the method described above a�orded the title
�
compound as a white solid (30%), m.p. 161–162 C (Found:
C, 71�3; H, 5�6; N, 4�7. C18H17F2NO requires C, 71�7; H,
1
5
�7; N, 4�7%). Mass spectrum: c.i. m/z 302 (99%, M+1).
H
1500 Tri-Carb liquid scintillation analyser (Packard Instrument
Co., Downers Grove, IL, U.S.A.).
n.m.r. (300 MHz): � 1�5, 1H, d, J 10�4 Hz, CHCH2CH; 1�8,
d, J 10�4 Hz, CHCH2CH; 2�4–3�0, 8H, m, CH; 3�3, 1H, t, J
4
�7 Hz, NCH; 3�4, 1H, d, J 14 Hz, NCH2; 3�8, 1H, d, J 13�7
General Method for Preparation of 4-Azahexacyclo-
2
,6
3,10
5,9
8,11
Hz, NCH ; 7�0–7�3, 3H, m, ArH.
2
[
5.4.1.0 .0
A mixture of the Cookson diketone (8) (1 g, 5�7 mmol),
ethylene glycol (0�32 ml, 5�7 mmol) and p-toluenesulfonic acid
.0 .0
]dodecanes
0
N-(3 -Tri�uoromethylphenyl)methyl-4-azahexacyclo-
2
,6
3,10
5,9
8,11
[
5.4.1.0 .0
.0 .0
]dodecan-3-ol (6)
(
0�01 g) were reacted in boiling benzene (5 ml) for 5 h,
Treating the monoketal (9) with 3-tri�uoromethylbenzyl-
allowed to cool and basi�ed with a saturated solution of sodium
bicarbonate (25 ml). The benzene layer was separated and
the aqueous layer extracted with dichloromethane (10 ml� 2).
The combined organic layers were dried over anhydrous sodium
sulfate, �ltered and the solvent was removed by rotary evapo-
ration. Recrystallization of the residue by using diethyl ether
amine according to the method described above a�orded the
�
title compound as a white solid (30%), m.p. 163–164 C (Found:
C, 68�1; H, 5�4; N, 4�2. C19H18F3NO requires C, 68�5; H,
1
5
�4; N, 4�2%). Mass spectrum: c.i. m/z 334 (99%, M+1).
H
n.m.r. (300 MHz): � 1�5, 1H, d, J 10�2 Hz, CHCH2CH; 1�8,
d, J 10�7 Hz, CHCH2CH; 2�4–3�0, 8H, m, CH; 3�3, 1H, t, J
(
30 ml) gave the monoketal (9) as a white solid (86%). The
4
�8 Hz, NCH; 3�5, 1H, d, J 14 Hz, NCH2; 3�8, 1H, d, J 14
monoketal (9) (2 g, 9�2 mmol) was added to appropriately
Hz, NCH2; 7�4–7�6, 4H, m, ArH.
substituted benzylamines (9�2 mmol) in a sealed tube with
�
ethanol (20 ml) and the mixture was heated to 100 C for 14
0
N-(3 -Nitrophenyl)methyl-4-azahexacyclo-
h. The mixture was allowed to cool and treated with excess
sodium borohydride (0�4 g) with stirring at room temperature
for 6 h. Ethanol was removed by rotary evaporation and
water (30 ml) was added to the residue. The mixture was
extracted with dichloromethane (20 ml� 3) and the extracts
were dried over anhydrous sodium sulfate and �ltered; the
solvent was removed by rotary evaporation. The resulting
2,6
3,10
5,9
8,11
[
5.4.1.0 .0
.0 .0
]dodecan-3-ol (7)
3-Nitrobenzylamine hydrochloride (2�5 g) was mixed with
1M NaOH until pH 14 and extracted with dichloromethane to
give the free base, 3-nitrobenzylamine, as a brown oil (1�8
g, 89%). Treating the monoketal (9) with 3-nitrobenzylamine
according to the method described above a�orded the title
2,6 3,10 5,9
�
pentacyclo[6.3.0.0 .0
.0 ]undecylamines (10)–(14) were
compound as a brown solid (30%), m.p. 169–170 C (Found:
hydrolysed with 2 M HCl (40 ml) at room temperature overnight.
The reaction mixture was then basi�ed with 1 M sodium hydrox-
ide to pH 14 and extracted with dichloromethane (20 ml� 3).
The extracts were dried over anhydrous sodium sulfate, �l-
C, 69�7; H, 5�9; N, 9�1. C18H18N2O3 requires C, 69�7; H,
1
5�9; N, 9�0%). Mass spectrum: c.i. m/z 311 (99%, M+1).
H
n.m.r. (300 MHz): � 1�5, 1H, d, J 10�7 Hz, CHCH2CH; 1�8,
d, J 10�4 Hz, CHCH2CH; 2�4–3�0, 8H, m, CH; 3�3, 1H, t, J
�
10
10 � 1
s .
1
Ci = 3�7� 10 Bq = 3�7� 10

6
5
56
X. Liu, M. Kassiou and M. J. Christie
Hz, NCH; 3�5, 1H, d, J 14�6 Hz, NCH2; 3�9, 1H, d, J 14�6
References
Hz, NCH2; 7�5, 1H, t, J 7�8 Hz, ArH; 7�7, 1H, d, J 7�8 Hz,
1
ArH; 8�1, 1H, d, J 8�1 Hz, ArH; 8�2, 1H, s, ArH.
Martin, W. R., Eades, C. G., Thompson, J. A., Happler,
R. E., and Gilbert, P., J. Pharmacol. Exp. Ther., 1976,
1
97, 517.
2
Walker, J. M., Bowen, W. D., Walker, F. O., Matsumoto,
R. R., De Costa, B., and Rice, K. C., Pharmacol. Rev.,
In Vitro Binding Procedures
Receptor source. Membranes of guinea pig whole brain plus
1
990, 42, 355.
cerebellum were prepared as the source of receptors.
3
4
5
Wu, X. Z., Bell, C. E., Spivak, C. E., London, E. D., and
Su, T. P., J. Pharmacol. Exp. Ther., 1991, 257, 351.
Bowen, W. D., Hellewell, S. B., and McGarry, K. A., Eur.
J. Pharmacol., 1989, 163, 309.
Matsumoto, R. R., Hemstreet, M. K., Lai, N., Thurkauf,
A., De Costa, B. R., Rice, K. C., Hellewell, S. B., Bowen,
W. D., and Walker, J. M., Pharmacol., Biochem. Behav.,
3
�
1 Binding assay. [ H](+)-Pentazocine was used to label
16
�
1 sites. Non-speci�c binding was determined by using 10 �M
of the high a� nity � binding drug haloperidol. Brie�y, each
assay tube contained radioligand at a �nal concentration of
3
2
1
nM [ H](+)-pentazocine, tissue suspension (approximately
8 mg fresh weight per tube), various concentrations of test
compounds (1 nM–100 �M) and the assay bu�er in a �nal
volume of 1 ml. After incubation at 37 C for 150 min, the
reaction was terminated by rapid �ltration through a Brandel
1
990, 36, 151.
�
6
7
Walker, J. M., Bowen, W. D., Patrick, S. L., Williams, W.
E., Mascarella, S. W., Bai, X., and Carroll, F. I. A., Eur.
J. Pharmacol., 1993, 231, 61.
Hanner, M., Moebius, F. F., Flandorfer, A., Knaus, H. G.,
Striessnig, J., Kempner, E., and Glossmann, H., Proc. Natl
Acad. Sci. U.S.A., 1996, 93, 8072.
Moebius, F. F., Reiter, R. J., Hanner, M., and Glossmann,
H., Br. J. Pharmacol., 1997, 121, 1.
Vilner, B. J., John, C. S., and Bowen, W. D., Cancer Res.,
2
4-well cell harvester over Whatman GF/B glass �bre �lters
that were presoaked in a solution of 0�5% polyethylenimine
at room temperature for at least 2 h prior to use. Follow-
ing addition of scintillation cocktail, samples were allowed
to equilibrate overnight. The amount of bound radioactivity
was determined by liquid scintillation spectrometry with a
Packard liquid scintillation analyser. Each concentration of
test compounds was tested in quadruplicate. Binding data
were �tted by using KaleidaGraph, Version 3.0, Abelbeck
8
9
1
995, 55, 408.
10
Mach, R. H., Smith, C. R., Al-Nabulsi, I., Whirrett, B. R.,
Childers, S, R., and Wheeler, K. T., Cancer Res., 1997,
3
software. [ H](+)-Pentazocine saturation binding data were
best described by a one-site model.
5
7, 156.
�
2 Binding assay. The selectivity for �2 sites of the test
11
12
13
14
15
Perregaard, J., Moltzen, E. K., Meier, E., and Sanchez, C.,
J. Med. Chem., 1995, 38, 1998.
Oliver, D. W., Dekker, T. G., Snyckers, F. O., and Fourie,
T. G., J. Med. Chem., 1991, 34, 851.
Oliver, D. W., Dekker, T. G., and Snyckers, F. O., Eur. J.
Med. Chem., 1991, 26, 375.
Nguyen, V. H., Kassiou, M., Johnston, G. A. R., and
Christie, M. J., Eur. J. Pharmacol., 1996, 311, 233.
Kassiou, M., Nguyen, V. H., Knott, R., Christie, M. J.,
and Hambley, T. W., Bioorg. Med. Chem. Lett., 1996, 6,
3
compounds was measured by labelling with [ H]DTG. The
3
[
two-site model.
H]DTG saturation binding data were best described by a
14
Binding was therefore carried out in the
17
presence of (+)-pentazocine to mask �1 sites. Similarly, each
assay tube contained 10 nM [ H]DTG, tissue suspension, various
concentrations of test compounds and the assay bu�er in a
3
�
�
nal volume of 1 ml. Following incubation at 25 C for 90 min
3
in the [ H]DTG assay, the workup procedure was as described
in the �1 assay.
5
95.
Hudkins, R. L., and De Haven-Hudkins, D. L., Life Sci.,
991, 49, 1229.
Weber, E., Sonders, M., Quarum, M., McLean, S., Pou, S.,
and Keana, J. F. W., Proc. Natl Acad. Sci. U.S.A., 1986,
83, 8784.
1
1
6
7
1
Acknowledgment
This work was supported by the National Health
and Medical Research Council (Grant No. 970652).