Structure-activity relationship study of 1,4-dihydropyridine derivatives blocking N-type calcium channels

Article abstract of DOI:10.1016/j.bmcl.2005.11.021

Cilnidipine is a 1,4-dihydropyridine derived L/N-type calcium channel dual blocker possessing neuroprotective and analgesic effects which are related to its N-type calcium channel inhibitory activity. In order to find specific N-type calcium channel blockers with the least effects on cardiovascular system, we performed structure-activity relationship study on APJ2708, which is a derivative of cilnidipine, and found a promising N-type calcium channel blocker 21b possessing analgesic effect in vivo with a 1600-fold lower activity against L-type calcium channels than that of cilnidipine.

Full text of DOI:10.1016/j.bmcl.2005.11.021

Bioorganic & Medicinal Chemistry Letters 16 (2006) 798–802
Structure–activity relationship study of 1,4-dihydropyridine
derivatives blocking N-type calcium channels
Takashi Yamamoto, Seiji Niwa, Seiji Ohno, Tomoyuki Onishi, Hiroyuki Matsueda,
Hajime Koganei, Hisayuki Uneyama, Shin-ichi Fujita, Tomoko Takeda, Morikazu Kito,
Yukitsugu Ono, Yuki Saitou, Akira Takahara, Seinosuke Iwata and Masataka Shoji*
Pharmaceutical Research Laboratory, Ajinomoto company Inc., 1-1, Suzuki-cho, Kawasaki-ku, Kawasaki-shi, Japan
Received 15 August 2005; revised 5 November 2005; accepted 8 November 2005
Available online 23 November 2005
Abstract—Cilnidipine is a 1,4-dihydropyridine derived L/N-type calcium channel dual blocker possessing neuroprotective and anal-
gesic effects which are related to its N-type calcium channel inhibitory activity. In order to find specific N-type calcium channel
blockers with the least effects on cardiovascular system, we performed structure–activity relationship study on APJ2708, which is
a derivative of cilnidipine, and found a promising N-type calcium channel blocker 21b possessing analgesic effect in vivo with a
1600-fold lower activity against L-type calcium channels than that of cilnidipine.
Ó 2005 Elsevier Ltd. All rights reserved.
Voltage-dependent calcium channels (VDCCs), which
have been divided into five subtypes (L, CaV1; P/Q,
CaV2.1; N, CaV2.2; R, CaV2.3; T, CaV3) based on
their pharmacological and biophysical properties,
mediate a range of cytoplasmic responses, including
muscle contraction, release of neurotransmitters, calci-
um-dependent gene transcription, and the regulation
of neuronal excitability.1 Among them, N-type
calcium channels are extensively distributed on the
sympathetic nerve endings and related to the neuro-
protection and neuropathic pain.2 Several reports have
suggested that a blockade or lack of N-type calcium
channels can suppress the pathological processes of
ischemic brain injury and pain in animal models.3
sympathetic overactivity in hypertensive patients.6
Moreover, N-type calcium channel-blocking profile
of cilnidipine may contribute to its neuroprotective
action in the animal focal brain ischemia model and
its intrathecal analgesic effect in rat formalin-induced
pain model.7,8
NO2
O
O
R
O
Ph
N
H
OMe
Cilnidipine ( 1); R =
O
APJ2708 (2) ; R = OH
Cilnidipine is a 1,4-dihydropyridine derived long-
acting calcium channel blocker which inhibits both
L-type and N-type calcium channels,4 and is currently
used for the treatment of essential hypertension in
Japan.5 Its inhibitory effect for N-type calcium chan-
nel can be clinically observed in the reduction of white
coat effect, cold pressor stress-induced platelet aggre-
gation, urinary catecholamine excretion, and cardiac
Previously, we reported that APJ2708 (2), which is a
carboxylic acid derivative of cilnidipine, has almost the
same inhibitory activity against N-type calcium channels
with far lower activity against L-type channels than that
of cilnidipine.7b As a consequence of this L-type activity,
APJ2708 showed a weak hemodynamic effect and need-
ed more than 100-fold intravenously administered dose
for decreasing the same amount of blood pressure as
cilnidipine in vivo. We assumed that the carboxylic acid
moiety of APJ2708 was the key structure for this selec-
tivity and began our study by optimizing APJ2708 with
Keywords: N-type calcium channels blocker; Analgesic; 1,4-Dihydro-
pyridine; Cilnidipine; Structure–activity relationship study.
*
Corresponding author. Tel.: +81 44 210 5826; fax:+81 44 210 5876;
e-mail: masataka_shouji@ajinomoto.com
0960-894X/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.11.021

T. Yamamoto et al. / Bioorg. Med. Chem. Lett. 16 (2006) 798–802
799
the aim of discovering N-type calcium channel blockers
with low activity against L-type channels which have
lesser influence on the cardiovascular systems.
alcohol in the presence of EDC followed by the
removal of 2-cyanoethyl ester.
Dipyridyl-methanone 13 was reacted with triethyl phos-
phonoacetate under Wittig condition to give 3,3-dipyr-
idyl acrylate 14, which led to 3,3-dipyridyl propionate
15 by reduction using sodium borohydride in the pres-
ence of nickel chloride. Following reduction by LAH
yielded 3,3-dipyridyl propanol 16, which was condensed
with 8b to give 9j.
The reaction pathways used to synthesize the designed
26 compounds are described in Scheme 1.
All the 1,4-dihydropyridine ring was constructed by
Hantzsch 1,4-dihydropyridine synthesis reaction.9 The
b-ketoesters (4a, 4b, and 4c) and enamines (5a and
5b) were prepared from diketene 3 and the corre-
sponding alcohol under basic conditions. The 1,4-
dihydropyridine dicarboxylic acid cinnamyl ester
2-cyanoethyl ester 6a–m was obtained using the three
component coupling reaction of b-ketoester, enamine,
and the corresponding benzaldehyde. The selective
deprotection of 2-cyanoethyl ester group of 6a–m
was performed by the treatment with 1 M sodium
hydroxide solution to give 7a–m, since 2-cyanoethyl
ester was removed easily via b-elimination reaction
in the basic condition,10 while cinnamyl ester was
stable for hours. The 1,4-dihydropyridine dicarboxylic
acid mono ester derivatives 9a–g and 9i were prepared
by a similar method. The 1,4-dihydropyridine dicar-
boxylic acid benzyl ester 2-cyanoethyl ester 8a was ob-
tained using 4c and 5b as starting materials. Catalytic
hydrogenation of 8a yielded the 1,4-dihydropyridine
dicarboxylic acid mono 2-cyanoethyl ester 8b. The
1,4-dihydropyridine dicarboxylic acid mono ester 9h
was prepared via condensation with the corresponding
The b-ketoamide 4d was obtained by similar a condition
as that described for b-ketoesters. The 1,4-dihydropyri-
dine dicarboxylic acid 2-cyanoethyl ester cinnamyl
amide 11 was synthesized via two-step reactions, in
which 4d and benzaldehyde were condensed under
Knoevenagel condition followed by 1,4-dihydropyridine
ring construction using 5b. The treatment of 1 N sodium
hydroxide solution yielded the desired 1,4-dihydropyri-
dine dicarboxylic acid mono cinnamyl amide 12.
The synthesis of 2-trifluoromethyl 1,4-dihydropyridine
derivative 21b required a different method since the
dehydration of aminal 19 did not proceed under similar
conditions as those used for 2-methyl 1,4-dihydropyri-
dine derivatives. Compound 19 was isolated and
subsequent phosphorus oxychloride/pyridine adsorbed
silica gel11 treatment was applied for the dehydration.
Following hydrogenation, etherification and removal of
2-cyanoethyl ester yielded the desired compound 21b.
O
R2
O
R3
O
O
NH2
O
a
4a + 5b
or
4b + 5a
b
O
c
O
O
Ph
R1
R1
O
N
H
3
CN
6a-m: R3 =
4a: R1 =
4b: R1 =
5a: R1 =
5b: R1 =
O
O
O
Ph
Ph
Ph
d
7a-m: R3 = H
CN
CN
O
4c: R1 = OBn
4d: R1 = HN
Cl
O
Cl
Cl
O
O
O
O
O
O
O
N
H
Ph
g, h
j
i
e
NC
R4O
N
Ph
k
4c + 5b
4d
Cl
R3
O
R3
HO
H
N
H
N
N
H
H
10
8a: R3 = OBn
8b: R3 = OH
CN
9h, 9j
11: R4 =
12: R4 = H
f
N
N
N
N
l
O
O
O
O
m
o
n
O
CF3
CF3
OMe
OBn
CO2Et
N
CO2Et
OH
N
N
N
17
18
13
14
15
16
Cl
O
Cl
O
Cl
O
Ph
O
O
O
O
p
q
s
5b + 18
NC
NC
R6
OBn
O
R5
O
O
Ph
N
H
CF3
OH
N
CF3
N
H
CF3
CN
H
20a: R5 = OBn
20b: R5 = OH
19
21a: R6 =
21b: R6 = H
r
t
Scheme 1. Reagents: (a) alcohol or amine, Et3N, toluene; (b) AcONH4, i-PrOH; (c) R2CHO, i-PrOH; (d) 1 M NaOH, MeOH; (e) 3-
chlorobenzaldehyde, i-PrOH; (f) H2, Pd/C, EtOH; (g) alcohol, EDC, Et3N, CH2Cl2; (h) 1 M NaOH, MeOH; (i) 3-chlorobenzaldehyde, i-PrOH; (j)
5b, i-PrOH; (k) 1 M NaOH, MeOH; (l) triethyl phosphonoacetate, NaH, EtOH; (m) NaBH4, NiCl2Æ6H2O, EtOH; (n) LAH, THF; (o) BnOH,
toluene; (p) 3-chlorobenzaldehyde, i-PrOH; (q) POCl3, pyridine, silica gel, 1,2-dichloroethane; (r) H2, Pd/C, EtOH; (s) 3,3-diphenylpropanol, EDC,
DMAP, Et3N, CH2Cl2; (t) 1 M NaOH, MeOH.

800
T. Yamamoto et al. / Bioorg. Med. Chem. Lett. 16 (2006) 798–802
The lead compound, APJ2708 (2), showed two times
lower activity against N-type calcium channels
(IC50 = 3.5 lM; IMR-32 assay12) than that of cilnidipine
(IC50 = 1.6 lM), while its L-type activity was more than
40-fold lower (IC50 = 0.046 lM; Magnus method13). We
ascribed this significant decrease in L-type activity to the
presence of carboxyl group in 2. Therefore, we started
the optimizations by changing the substituent on the
benzene ring bound to 4-position of the 1,4-dihydropyri-
dine ring (Table 1).
that there was no influence of electronic effect of the sub-
stituent on the benzene ring. However, we chose 7h as
the best since it had more potent activity against N-type
calcium channels and it may have more potential as a
second lead compound.
The optimization of cinnamyl moiety of 7h was also
investigated (Table 2). The activity of cinnamyl amide
derivative 12 was tested but its activity against N-type
channels was decreased (IC50 = 4.5 lM). Changing
the benzene ring to the cyclohexyl ring (9a) resulted
in increased N-type calcium channel activity
(IC50 = 0.83 lM), but its activity against L-type channel
was also high (IC50 = 0.0037 lM). Introduction of an
electron-withdrawing chlorine atom at 4-position (9b)
increased L-type activity (IC50 = 0.0058 lM) with a
First, the modification of nitro group at 3-position of
benzene ring was investigated. The derivatives possess-
ing 2- or 4-nitro group had a significantly lower activity
against N-type calcium channels (7a; IC50 = 37 lM, 7b;
21 lM) than that of cilnidipine. The derivative with a
phenyl group (7c) resulted in increased activity against
N-type channel (IC50 = 2.0 lM). However, the deriva-
tives with a heterocyclic ring (7k, 7l, and 7m) showed
lower activities for N-type channels (IC50 = 23, 2.9,
and 11 lM, respectively) than that of 7c. Therefore,
the optimization was next focused on 3-position of the
benzene ring. The introduction of an electron-donating
methyl group (7f) showed a slightly improved activity
for N-type (IC50 = 1.8 lM) with 70-fold lower activity
for L-type (IC50 = 0.076 lM). As for the electron-with-
drawing substituents, methoxycarbonyl (7d) and car-
boxylic acid (7e) derivatives resulted in lower activities
for N-type channels (IC50 = 4.6 and 37 lM, respective-
ly). On the other hand, the derivatives with halogen
atom revealed to have improved activity against N-type
calcium channels (IC50 = 1.3 lM; 7g, 1.0 lM; 7h,
0.79 lM; 7i and 1.7 lM; 7j, respectively) than that of
7c. Among them, 7h was most promising since it had
high selectivity for N-type channels. The selectivity of
3-methyl derivative 7f was also high, which suggested
Table 2. In vitro results of dihydropyridine derivatives
Cl
O
O
HO
R
N
H
Compound
R
N-type IMR-32 L-type Magnus
IC50 (lM)
IC50 (lM)
12
4.5
0.85
HN
Ph
O
O
9a
9b
9c
9d
9e
9f
0.83
1.5
0.0037
0.0058
0.0068
0.32
Cl
O
O
0.89
6.2
Table 1. In vitro results of dihydropyridine derivatives
O
R
O
N
HO
O
Ph
O
Ph
N
H
0.68
28
0.0028
0.050
0.014
0.20
Compound
R
N-type IMR-32 L-type Magnus
IC50 (lM)
H
N
IC50 (lM)
O
O
Ph
O
Cilnidipine (1)
2
1.6
3.5
37
0.0011
0.046
0.031
1.6
3-NO2-Ph
2-NO2-Ph
4-NO2-Ph
Ph
7a
7b
7c
7d
7e
7f
9g
9h
9i
1.0
21
2.0
4.6
37
0.0090
4.4
Ph
Ph
3-CO2Me-Ph
3-CO2H-Ph
3-Me-Ph
3-F-Ph
3.9
O
O
>10
Ph
1.8
1.3
1.0
0.79
1.7
24
0.076
0.020
0.020
0.010
0.013
0.19
7g
7h
7i
1.1
0.36
Ph
N
3-Cl-Ph
3-Br-Ph
3-I-Ph
7j
9j
10
1.6
7k
7l
3-Pyridyl
3-Thienyl
3-Furyl
N
O
2.9
11
0.044
0.13
7m
The inhibition against N-type (calcium influx using IMR-32 cells) and
L-type (magnus method) calcium channels.
The inhibition against N-type (calcium influx using IMR-32 cells) and
L-type (Magnus method) calcium channels.

T. Yamamoto et al. / Bioorg. Med. Chem. Lett. 16 (2006) 798–802
801
Table 4. Inhibitory activities of compounds in phase II pain responses
following footpad injection of formalin in the rat
slightly lower N-type activity (IC50 = 1.5 lM). The elec-
tron-donating methyl group at 4-position showed
improved potency for N-type calcium channels
(IC50 = 0.89 lM), but its activity against L-type was also
increased (IC50 = 0.0068 lM). The inhibitory activity
for N-type calcium channels of 4-pyridyl derivative 9d
was lower than that of cilnidipine.
Compound
Rat formalin test (30 mg/kg, po)
Inhibition %
n
Cilnidipine
21b
46
46
5
5
Inhibition % is given by the control and test compound.
The introduction of a b-methyl group (9e) or a naphthyl
group (9g) resulted in increased activities against L-type
channels (IC50 = 0.0028 and 0.014 lM, respectively).
The amide (9f) derivatives showed lower activity against
N-type calcium channel. The diphenylmethyl derivative
9i had an almost equivalent potency for N-type
(IC50 = 1.1 lM) with that of 7h with considerably de-
creased activity against L-type calcium channels
(IC50 = 0.36 lM), which was more than 320-fold lower
than that of cilnidipine. However, the N-type activity
of its diphenylmethylidene (9h) and bipyridylmethyl
(9j) derivatives for N-type channels was decreased.
romethyl-1,4-dihydropyridine-3,5-dicarboxylic acid 5-
(3,3-diphenylpropyl) ester 21b, which had an almost
equipotent activity against N-type calcium channels
but 1600-fold lower activity for L-type than that of ciln-
idipine, showed equivalent oral analgesic activity to ciln-
idipine. This result reemphasized the importance of
inhibitory activity against N-type channels for the treat-
ment of neuropathic pain. Further SAR studies are
under progress and will be reported elsewhere.
Finally, we introduced the trifluoromethyl moiety at the
6-position of the 1,4-dihydropyridine ring and found
that the derivative 21b had a 1600-fold lower activity
against L-type channels (IC50 = 1.8 lM) with an almost
equivalent activity for N-type (IC50 = 1.7 lM) to cilnid-
ipine (Table 3). This considerable decrease of L-type
activity might be ascribed to the inductive effect of triflu-
oromethyl moiety which increases the acidity of carbox-
ylic acid at 3-position of the 1,4-dihydropyridine ring.
References and notes
1. (a) Bowersox, S. S.; Valentino, K. L.; Luther, R. R. Drug
News Perspect. 1994, 7, 261; (b) Bean, B. P. Annu. Rev.
Physiol. 1989, 51, 367; (c) Hess, P. Annu. Rev. Neurosci.
1990, 13, 337; (d) Varadi, G.; Mori, Y.; Mikala, G.;
Schwartz, A. Trends Pharmacol. Sci. 1995, 16, 43; (e)
Nooney, J. M.; Lambert, R. C.; Feltz, A. Trends
Pharmacol. Sci. 1997, 18, 363.
2. (a) Hirning, L. D.; Fox, A. P.; McCleskey, E. W.; Olivera,
B.; Thayer, S.; Miller, R.; Tsien, R. Science 1988, 239, 57;
(b) Mori, Y.; Nishida, M.; Shimizu, S.; Ishii, M.; Yoshi-
naga, T.; Ino, M.; Sawada, K.; Niidome, T. Trends
Cardiovasc. Med. 2002, 12, 270.
3. (a) Saegusa, H.; Kurihara, T.; Zong, S.; Kazuno, A.;
Matsuda, Y.; Nonaka, T.; Han, W.; Toriyama, H.;
Tanabe, T. EMBO J. 2001, 20, 2349; (b) Takizawa, S.;
Matsushima, K.; Fujita, H.; Nanri, K.; Ogawa, S.;
Shinohara, Y. J. Cereb. Blood Flow Metab. 1995, 15,
611; (c) Perez-Pinzon, M. A.; Yenari, M. A.; Sun, G. H.;
Kunis, D. M.; Steinberg, G. K. J. Neurol. Sci. 1997, 153,
25.
4. (a) Oike, M.; Inoue, Y.; Kitamura, K.; Kuriyama, H. Circ.
Res. 1990, 67, 993; (b) Takahara, A.; Fujita, S.; Moki, K.;
Ono, Y.; Koganei, H.; Iwayama, S.; Yamamoto, H.
Hypertens. Res. 2003, 26, 743.
5. (a) Uneyama, H.; Uchida, H.; Konda, T.; Yoshimoto, R.
Cardiovasc. Drug Rev. 1999, 17, 341; (b) Rose, G. W.;
Kanno, Y.; Ikebukuro, H.; Kaneko, M.; Kaneko, K.;
Kanno, T.; Ishida, Y.; Suzuki, H. Hypertens. Res. 2001,
24, 377.
Subsequently, the oral analgesic effect of 21b was tested
in vivo using formalin-induced pain model.14 As can be
seen in Table 4, 21b showed equivalent oral activity to
cilnidipine. This result indicated that 21b was an orally
effective N-type calcium channel blocker. Moreover, this
result showed that the activity for N-type calcium chan-
nels is more important than that for L-type to exert the
analgesic efficacy.
In conclusion, the SAR study on 4,5 and 6-position of
series of 1,4-dihydropyridine derivatives starting from
APJ2708 led to the discovery of novel N-type calcium
channel blockers with less effects on cardiovascular sys-
tem than those of cilnidipine and APJ2708. Among
them, promising 4-(3-chloro-phenyl)-2-methyl-6-trifluo-
Table 3. In vitro results of dihydropyridine derivatives
Cl
6. (a) Tomiyama, H.; Kimura, Y.; Kuwabara, Y.; Maruy-
ama, C.; Yoshida, Y.; Kuwata, S.; Kinouchi, T.; Yoshida,
H.; Doba, N. Hypertens. Res. 2001, 24, 679; (b) Takeda,
S.; Ueshiba, H.; Hattori, Y.; Irie, M. Diabetes Res. Clin.
Pract. 1999, 44, 197; (c) Sakata, K.; Shirotani, M.;
Yoshida, H.; Nawada, R.; Obayashi, K.; Togi, K.; Miho,
M. Hypertension 1999, 33, 1447.
O
O
R
Ph
HO
O
Ph
N
H
7. (a) Takahara, A.; Konda, T.; Enomoto, A.; Kondo, N.
Biol. Pharm. Bull. 2004, 27, 1388; (b) Konda, T.; Takeda,
T.; Takahara, A.; Uneyama, H. J. Hypertens. 2003, 21,
P82 (Abstract).
8. Murakami, M.; Nakagawasai, O.; Fujii, S.; Hosono, M.;
Hozumi, S.; Esashi, A.; Taniguchi, R.; Okamura, T.;
Suzuki, T.; Sasano, H.; Yanagisawa, T.; Tan-no, K.;
Compound
R
N-type IMR-32
IC50 (lM)
L-type Magnus
IC50 (lM)
9i
21b
Me
CF3
1.1
1.7
0.36
1.8
The inhibition against N-type (calcium influx using IMR-32 cells) and
L-type (Magnus method) calcium channels.

802
T. Yamamoto et al. / Bioorg. Med. Chem. Lett. 16 (2006) 798–802
Tadano, T.; Kitamura, K.; Kisara, K. Brain Res. 2000,
868, 123.
of adhering periadventitial fat, and cut into rings of 3 mm
width. Removal of the endothelium was achieved by
gently rubbing the luminal surface. The ring was mounted
in an organ bath filled with warmed (37 °C), oxygenated
(95% O2/5% CO2) Tyrodeꢀs solution (pH 7.4). The ring
was equilibrated under resting tension of 2g for 1 h. Then
the ring was incubated in high K+ solution and then in
Tyrodeꢀs solution for 45 min each. The solution was
replaced with high K+ solution again. After attaining the
maximum contraction reaction, the test compound was
cumulatively added at intervals of 90 min to attain
concentrations of 10À9, 10À8, 10À7, and 10À6 M. The
inhibitory rate of the test compound against the maximum
contraction response was employed as the index of the
antagonistic activity on L-type calcium channels. Impor-
tantly, we have confirmed that the current IMR model is
about 1000 times less sensitive than the Magnus or in vivo
model for detecting each calcium channel-blocking action,
as described in our previous report (Takahara, A.; Fujita,
S.; Moki, K.; Ono, Y.; Koganei, H.; Iwayama, S.;
Yamamoto, H. Hypertens. Res. 2003, 26, 743). Therefore,
the IC50 values of the method against L-type calcium
channel could not be simply compared with the IC50
values of the cell-based IMR-32 method against N-type
calcium channels.
9. (a) Hantzsch, A. Ann. Chem. 1882, 215, 72; (b) Mosher, H.
S. Heterocycl. Compounds 1950, 1, 462; (c) Richard, M.;
Kellogg, T. J.; Bergen, V.; Van Doren, H.; Hedstrand, D.;
Kooi, J.; Kruizinga, W. H.; Troostwijk, C. B. J. Org.
Chem. 1980, 45, 2854.
10. (a) Hirose, Y.; Kariya, K.; Sasaki, I.; Kurono, Y.; Achiwa,
K. Tetrahedron Lett. 1993, 34, 3441; (b) Ferrer, E.;
Eisenhut, M.; Eritja, R. Lett. Pept. Sci. 1999, 6, 209; (c)
Sobolev, A.; Franssen, M. C. R.; Vigante, B.; Cekavicus,
B.; Zhalubovskis, R.; Kooijman, H.; Spek, A. L.; Duburs,
G.; de Groot, A. J. Org. Chem. 2002, 67, 401; (d) Sobolev,
A.; Franssen, M. C. R.; Vigante, B.; Cekavicus, B.;
Makarova, N.; Duburs, G.; de Groot, A. Tetrahedron
Asymmetry 2001, 12, 3251.
11. Katsuyama, I.; Funabiki, K.; Matsui, M.; Muramatsu, H.;
Shibata, K. Tetrahedron Lett. 1996, 37, 4177.
12. The inhibitory activity against N-type calcium channels
was measured by calcium influx into IMR-32 cells as
described Takahara, A.; Fujita, S.; Moki, K.; Ono, Y.;
Koganei, H.; Iwayama, S.; Yamamoto, H. Hypertens. Res.
2003, 26, 743.
13. Inhibitory activity against L-type calcium channels was
estimated from the effects on high K+-induced contraction
in rat thoracic aorta ring as follows: male SD rats (7 weeks
old) were used. The thoracic aorta was removed, cleared
14. Malmberg, A. B.; Yaksh, T. L. J. Neurosci. 1994, 14,
4882.