HAL Id: hal-03275373
https://hal.univ-angers.fr/hal-03275373
Submitted on 1 Jul 2021
HAL is a multi-disciplinary open access
archive for the deposit and dissemination of sci-
entic research documents, whether they are pub-
lished or not. The documents may come from
teaching and research institutions in France or
abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est
destinée au dépôt et à la diusion de documents
scientiques de niveau recherche, publiés ou non,
émanant des établissements d’enseignement et de
recherche français ou étrangers, des laboratoires
publics ou privés.
Celiprolol induces β3-adrenoceptors-dependent
relaxation in isolated porcine coronary arteries
Mohammed Abdelkrim, Lionel Martignat, Marc Gogny, Jean-Claude
Desfontis, Jacques Noireaud, Mohamed Yassine Mallem
To cite this version:
Mohammed Abdelkrim, Lionel Martignat, Marc Gogny, Jean-Claude Desfontis, Jacques Noireaud, et
al.. Celiprolol induces β3-adrenoceptors-dependent relaxation in isolated porcine coronary arteries.
Canadian Journal of Physiology and Pharmacology, NRC Research Press, 2013, 91 (10), pp.791-796.
�10.1139/cjpp-2013-0091�. �hal-03275373�
ARTICLE
Celiprolol induces 
3
-adrenoceptors-dependent relaxation in isolated
porcine coronary arteries
Mohammed Amine Abdelkrim, Lionel Martignat, Marc Gogny, Jean-Claude Desfontis, Jacques Noireaud, and Mohamed Yassine Mallem
Abstract: In porcine coronary arteries (PCAs), celiprolol, a selective 
1
-adrenoceptors antagonist, induces vasodilatation by an
endothelium- and nitric oxide (NO)-dependent pathway. However, the mechanisms of that vascular effect have not been
precisely established. 
3
-Adrenoceptors have been shown to be involved in the relaxation per se of various vascular beds,
including coronary vessels. Thus, we evaluated (i) the presence of 
3
-adrenoceptors in the PCA and (ii) their role in celiprolol-
induced vasodilatation. PCA rings were placed in organ baths and preconstricted with KCl. All experiments were performed in
the presence of nadolol (a 
1
/
2
-adrenoceptor antagonist). Cumulative concentration–response curves to SR 58611A and
ICI 215001 (2 
3
-adrenoceptor agonists) and to celiprolol were constructed. We also used semiquantitative reverse transcription –
polymerase chain reaction, which clearly showed the presence of 
3
-adrenoceptor transcripts. SR 58611A, ICI 215001, and
celiprolol induced concentration-dependent relaxations in PCA rings. SR 58611A-induced relaxation was almost abolished after
removal of endothelium or pretreatment with L-NAME (a NO synthase inhibitor). The vasorelaxations induced by SR 58611A and
celiprolol were inhibited in the presence of SR 59230A and L-748337 (2 selective 
3
-adrenoceptor antagonists). We showed (i) that
PCAs possess functional 
3
-adrenoceptors mediating endothelium- and NO-dependent relaxation, and (ii) that celiprolol exerts
a 
3
-adrenoceptor agonistic activity in this vascular bed.
Key words: celiprolol, 
3
-adrenoceptors, vasorelaxation, endothelium, nitric oxide, porcine coronary arteries.
Résumé : Dans l’artère coronaire porcine (ACP), le céliprolol, un antagoniste sélectif des récepteurs adrénergiques (RAs) 
1
,
induit une vasodilatation impliquant une voie de signalisation endothéliale et dépendante du monoxyde d’azote (NO). Les
mécanismes de cet effet vasculaire ne sont pas précisément établis. Les RAs 
3
sont impliqués dans la relaxation per se de
différents lits vasculaires incluant les vaisseaux coronariens. Nous avons donc évalué (i) la présence de RAs 
3
dans l’ACP et
(ii) leur rôle dans la vasodilatation induite par le céliprolol. Des anneaux d’ACP ont été placés dans des cuves a
`
organes isolés,
précontractés avec du KCl et en présence de nadolol (antagoniste des RAs 
1
/
2
). Des courbes cumulatives concentrations
réponses au SR 58611A, au ICI 215001 (agonistes du RA 
3
) et au céliprolol ont été construites. La méthode de réaction en chaîne
par polymérisation de transcription inverse a montré la présence de transcrits du RA 
3
. Le SR 58611A, l’ICI 215001 et le céliprolol
induisent une relaxation concentration-dépendante dans les anneaux d’ACP. La relaxation induite par le SR 58611A est presque
totalement abolie dans les anneaux désendothélialisés ou après prétraitement avec du L-NAME (inhibiteur de la NO synthase).
Les vasorelaxations induites par le SR 58611A et le céliprolol sont inhibées en présence de SR 59230A et de L-748337 (antagonistes
du RA 
3
). Ainsi, les ACPs possèdent des RAs 
3
, induisant une relaxation endothéliale NO dépendante et le céliprolol y exerce un
effet agoniste via les RAs 
3
.
Mots-clés : céliprolol, récepteurs adrénergiques 
3
, vasorelaxation, endothélium, monoxyde d’azote, artères coronaries porcines.
Introduction
Celiprolol, a third generation -adrenoceptor blocker, is a selec-
tive 
1
-adrenoceptor blocker with ancillary properties that include
partial 
2
-adrenoceptor agonism and the ability to relax vascular
smooth muscle directly (Toda 2003). The vasodilatory property of
celiprolol involving its ability to interact with 
2
-adrenoceptor has
been described in human, rat, and guinea-pig blood vessels (Dhein
et al. 1992; Kakoki et al. 1999; Sauvaget et al. 2010). However, many
experimental works have shown that celiprolol-induced vasodilata-
tion is ascribable to nitric oxide (NO) release from endothelium aside
from that mediated through 
2
-adrenoceptor stimulation. Thus, in
rat thoracic aorta, celiprolol causes vasorelaxation also through
5-hydroxytryptamine (5-HT) receptor activation (Kakoki et al. 1999).
In the arterioles of the rat cremaster muscle, propranolol only par-
tially attenuates the celiprolol-mediated relaxation (Perrone and
Barrett 1991). Furthermore, in porcine and dog coronary arteries,
celiprolol produces a vasodilatation by a mechanism that is insensi-
tive to -adrenoceptor blockade (Noda et al. 2001; Asanuma et al.
2003).

3
-Adrenoceptors have been shown to be involved in the relax-
ation per se of various vascular beds (Trochu et al. 1999; Mallem et al.
2004). Furthermore, vasodilatation induced by nebivolol, another
third generation -adrenoceptor blocker, involves 
3
-adrenoceptors
(for references see Rozec et al. 2006). However, no study has specifi-
cally explored the participation of 
3
-adrenoceptors in the vasodila-
tory effect of celiprolol. Only Noda et al. (2001) addressed that
possibility but failed to find any evidence that 
3
-adrenoceptors were
involved in the celiprolol-induced effect using cyanopindolol, a non-
Received 9 March 2013. Accepted 15 May 2013.
M.A. Abdelkrim, M. Gogny, J.-C. Desfontis, J. Noireaud,* and M.Y. Mallem. L’Université Nantes Angers Le Mans (LUNAM) - Oniris, UPSP 5304 de physiopathologie animale et de
pharmacologie fonctionnelle, Atlanpole-La Chantrerie, B.P. 40706, Nantes F-44307, France.
L. Martignat. L’Université Nantes Angers Le Mans (LUNAM) - Oniris, UPSP SSBR, Sécurité Sanitaire en Biotechnologies de la Reproduction, Atlanpole-La Chantrerie, B.P. 40706, Nantes
F-44307, France.
Corresponding author: M.Y. Mallem (e-mail: yassine.mallem@oniris-nantes.fr).
*Present address: Institut National de la Santé et de la Recherche Médicale (INSERM), UMR 1063, IBS-IRIS, Université d’Angers, 4 rue Larrey CHU, F-49933 Angers cedex 9, France.
A correction was made to the e-First version of this paper on 5 September 2013 prior to final issue publication. The current online and print versions are identical and both contain the correction.
Pagination not final/Pagination non finale
1
Can. J. Physiol. Pharmacol. 91: 1–6 (2013) dx.doi.org/10.1139/cjpp-2013-0091
Published at www.nrcresearchpress.com/cjpp on 22 May 2013.
Can. J. Physiol. Pharmacol. Downloaded from www.nrcresearchpress.com by Couperin on 09/06/13
For personal use only.
specific 
3
-adrenoceptor antagonist in porcine coronary arteries
(PCAs). Hence, the aim of the present study was to investigate
whether PCAs possess functional 
3
-adrenoceptors and to evaluate
whether celiprolol-induced vasorelaxation involves activation of 
3
-
adrenoceptors in this vascular bed.
Materials and methods
Biological materials
Porcine hearts were collected at a local abattoir (Montfort sur
Meu, France). From the freshly excised heart, the left anterior
descending coronary arteries were isolated and placed in ice-cold
Tyrode solution of the following composition (mmol/L): NaCl, 137;
KCl, 5.4; CaCl
2
, 2.5; Na
2
HPO
4
, 1.2; MgCl
2
, 1.2; HEPES, 20; and
D-glucose, 15 (pH adjusted to 7.4 with 5 mol/L NaOH). After collec-
tion, PCAs were maintained in ice-cold Tyrode solution for trans-
port to the laboratory (45 min).
Semiquantitative reverse transcription – polymerase chain
reaction (RT-Q-PCR) for mRNA amplification
Total RNAs were extracted from porcine adipose tissues (taken
as a positive control for 
2
- and 
3
-adrenoceptor expressions) and
PCAs rings, using the NucleoSpin RNA II kit (Macherey Nagel
EURL, Hoerd, France). First-strand cDNA was synthesized,1hat
37 °C, using M-MLV reverse transcriptase (Promega, Charbon-
nières, France) and random hexamer primers (Promega). RT-
Q-PCR was performed on 2 L of cDNA, with 5× Hot Firepol
EvaGreen qPCR Mix (Solis BioDyne, Tartu, Estonia), on an ABI
Prism 7300 real-time sequence detection system (Applied Biosy-
tems, Courtaboeuf, France). Cycling parameters were initiation at
50 °C for 2 min and denaturation at 95 °C for 15 min, followed
by 40 cycles of denaturation 95 °C for 20 s and annealing at
68 °C for 30 s. The following pig-specific primer pairs were used:
hypoxanthine-guanine phosphoribosyltransferase (forward 5=-TGG-
TCAAGCAGCATAATCCAAAGA-3=, reverse 5=-AGTCAAGGGCATAGC-
CTACCACAA-3=), 
2
-adrenoceptor (ENSSSCT00000015773, forward
5=-CCCGATGATAGCACTGATTCACAG-3=, reverse 5=-GGAAAGGGGCGC-
TTAGAAAGTAGA-3=), and 
3
-adrenoceptor (ENSSSCT00000017229, for-
ward 5=-GCTCATCATGGGAACCTTCACTCT-3=, reverse 5=-CTAGCCAGT-
TAAGGGCGAGGAAAG-3=). All experiments were normalized with a
unique reference sample. Relative RT-PCR data were processed ac-
cording to a standard method generating reference hypoxanthine-
guanine phosphoribosyltransferase gene normalized expression
values.
Contraction–relaxation studies
In the laboratory, arteries were cleaned of fat and connective
tissues and cut into rings (3–4 mm long). In some experiments, the
rings were denuded of endothelium by gently rubbing the intimal
surface with a pair of small, fine forceps. Rings were suspended on
stainless-steel wires ina5mLorgan bath containing Krebs solu-
tion of the following composition (mmol/L): NaCl, 118.3; KCl, 4.7;
MgSO
4
, 1.2; KH
2
PO
4
, 1.2; NaHCO
3
, 20; EDTA, 0.016; D-glucose, 11.1;
and CaCl
2
, 2.5. The temperature of the bath was maintained at
37 ± 0.5 °C, and the Krebs solution was continuously oxygenated
with a 95% O
2
,5%CO
2
gas mixture. PCA rings were progressively
stretched to a resting tension of 2 g. Isometric tension was re-
corded using a force displacement transducer (EMKA Technolo-
gies, Paris, France) and displayed on a computer (Acqknowledge
version 4.1, MP150 BIOPAC system, Cerom, France). Aftera1h
equilibration period, with the Krebs solution being changed every
15 min, the rings were contracted with 80 mmol/L KCl. Thereafter,
the rings were partially contracted with KCl (20–30 mmol/L, con-
centration adjusted to produce a similar tone level of 50% of the
maximal response), and the presence of a functional endothelium
was evaluated by the response (≥60% relaxation) to a single con-
centration of bradykinin (1 mol/L). In endothelium-denuded
rings, endothelium removal was confirmed by the absence of
bradykinin-induced relaxation.
All functional studies presently reported were carried out using
rings pre-equilibrated for 30 min in Krebs solution containing
nadolol (a 
1
- and 
2
-adrenoceptor antagonist, 10 mol/L). After
equilibrating for 30 min, rings were contracted again with KCl
(20–30 mmol/L, see above). Once the contraction reached a pla-
teau, cumulative concentration–response curves (CCRCs) to
SR 58611A (0.03–100 mol/L) and ICI 215001 (0.03–300 mol/L) (2

3
-adrenoceptor agonists) and to celiprolol (0.1–300 mol/L) were
then constructed. The relaxation produced by each concentration
of the drugs was measured after a steady state was reached. Values
were expressed as the percentage change in the maximal tension
of rings after addition of KCl. Some rings were equilibrated for
30 min in Krebs solution containing L-748337 (3 and 10 mol/L) or
SR 59230A (10 mol/L) (2 
3
-adrenoceptor antagonists) or L-NAME
(a nonselective NO synthase inhibitor, 100 mol/L) before plotting
CCRCs to SR 58611A or celiprolol. It should be noted that since
celiprolol was previously reported to induce endothelium- and
NO-dependent relaxation in PCAs (Noda et al. 2001), the effects of
endothelium removal and NO synthase inhibition upon the
celiprolol-induced response have not been determined in the
present work.
Drugs
Nadolol, SR 59230A (3-(2-ethylphenoxy)-1-[[(1S)-1,2,3,4-tetra-
hydronaphthalen-1-yl]amino]-(2S)-2-propanol oxalate salt) and
L-NAME (N-nitro-
L-arginine methyl ester hydrochloride) were
obtained from Sigma-Aldrich (France). SR 58611A ((RS)-N-[(25)-
7-ethoxycarbonylmethoxy-1,2,3,4-tetrahydronaphtalen-2-yl]-(2R)-2
(3-chlorophenyl)-2 hydroethanamine hydrochloride) was a gener-
ous gift from Sanofi Synthelabo Recherche (France). L-748337
(N-[[3-[(2S)-2-hydroxy-3-[[2-[4-[(phenylsulfonyl)amino]phenyl]ethyl]
amino]propoxy]phenyl]methyl]-acetamide) and ICI 215001 ((S)-4-
[2-hydroxy-3-phenoxypropylaminoethoxy]phenoxyacetic acid hy-
drochloride) were provided by Tocris bioscience (Bristol, UK).
Celiprolol hydrochloride was purchased from LGC standards (Mol-
sheim, France). All drugs were prepared as stock solutions in dis-
tilled water, with the exception of nadolol, which was dissolved in
HCl and neutralized with NaOH to pH 7.4, and of L-748337,
ICI 215001, SR 58611A, and SR 59230A, which were dissolved in
dimethylsulphoxide (Sigma-Aldrich). The final concentration of
the solvents in the organ bath was less than 0.1% v/v and was used
as a control for the effect of the active drugs.
Fig. 1. Semiquantitative reverse transcription – polymerase chain
reaction of 
2
- and 
3
-adrenoceptor (
2
-AR and 
3
-AR) mRNA
expression in porcine coronary artery (PCA) rings and adipose tissue
(AT). -Adrenoceptor mRNA levels were expressed as % relative to
hypoxanthine-guanine phosphoribosyltransferase (Hprt) expression.
Pagination not final/Pagination non finale
2 Can. J. Physiol. Pharmacol. Vol. 91, 2013
Published by NRC Research Press
Can. J. Physiol. Pharmacol. Downloaded from www.nrcresearchpress.com by Couperin on 09/06/13
For personal use only.
Statistical analysis
Results were expressed as a mean ± SE of n experiments, where
n is the number of pig hearts (one PCA/pig heart). Different CCRCs
were compared on R software (R Development Core Team 2007),
using the linear mixed-effects model (Pinheiro and Bates 2000;
Thorin et al. 2010; Kabbesh et al. 2012). The linear mixed-effects
model must follow a normal distribution of residuals. Normality
of residuals was systematically verified and validated. A level of
P < 0.05 was considered to be statistically significant. For RT-Q-
PCR experiments, statistical comparisons between groups of por-
cine adipose tissues versus PCA samples were performed using a
nonparametric Mann–Whitney test.
Results

3
-Adrenoceptor mRNA expression in PCA rings
RT-Q-PCR results clearly indicated the expression of 
2
-
adrenoceptor mRNA in both adipose tissues and PCAs (Fig. 1).
There were no statistical differences between the 2 tissue types
regarding 
2
-adrenoceptor mRNA expression. More interestingly,
we also found the unambiguous expression of 
3
-adrenoceptor
transcripts in PCAs. The 
3
-adrenoceptor mRNA level in PCAs,
hence, represented 35.8% ± 6.4% (n = 6) of the hypoxanthine-
guanine phosphoribosyltransferase gene basal transcription.
Thus, in PCAs, 
3
-adrenoceptor mRNA expression was about half
of 
2
-adrenoceptor mRNA expression (68.05 ± 5.16, P < 0.005).
Relaxant effects of SR 58611A and ICI 215001
Both 
3
-adrenoceptor agonists, SR 58611A and ICI 215001, in-
duced a concentration-dependent relaxation of PCA rings con-
tracted with KCl (20–30 mmol/L) after pretreatment and in the
presence of 10 mol/L nadolol, with SR 58611A having a greater
efficacy than ICI 215001 (Figs. 2A and 2B). The maximal responses
obtained for the highest concentration of SR 58611A and
ICI 215001 used were 94.25% ± 2.37% (n = 12) and 73.70% ± 3.63%
(n = 11), respectively, (P < 0.001).
After pretreatment and in the presence of 10 mol/L nadolol,
CCRCs to SR 58611A were concentration-dependently shifted to
the right by pretreatment with 3 and 10 mol/L L-748337 (P < 0.01),
and the SR 58611A-induced relaxation was significantly antago-
nized by pretreatment with 10 mol/L SR 59230A (P < 0.01) (Fig. 3).
The relaxant effect of SR 58611A was almost abolished after
removal of the endothelium or pretreatment with 100 mol/L
Fig. 2. (A) The effects of SR 58611A (a preferential 
3
-adrenoceptor agonist) and ICI 215001 (a partial 
3
-adrenoceptor agonist) in the presence
of nadolol (a 
1
- and 
2
-adrenoceptor antagonist) in intact porcine coronary artery rings. Cumulative concentration–response curves to
SR 58611A (0.03–100 mol/L) and ICI 215001 (0.03–300 mol/L) were constructed after 30 min of pretreatment with and in the presence of
nadolol (10 mol/L). Each point is the mean of n experiments obtained from n pig hearts, and vertical lines show the standard error of the
mean. When no error bar is shown, the error is smaller than the symbol. * indicates a significant difference (at P < 0.001) between ICI 215001
and SR 58611A. (B) A typical recording of relaxant effects of SR 58611A (3–100 mol/L) in an intact porcine coronary artery ring in the presence
of nadolol (10 mol/L) and preconstriction with KCl (20–30 mmol/L). SR 58611A induced a concentration-dependent relaxation (characterized
by its slow kinetics and long duration) at concentrations up to 100 mol/L.
Pagination not final/Pagination non finale
Abdelkrim et al. 3
Published by NRC Research Press
Can. J. Physiol. Pharmacol. Downloaded from www.nrcresearchpress.com by Couperin on 09/06/13
For personal use only.
L-NAME (Fig. 4). For the highest concentration used, the maximal
responses obtained following pretreatment with L-NAME or re-
moval of endothelium were 22.86% ± 8.33% (n = 8) and 20.95% ±
3.54% (n = 6), respectively, which are significantly different
(P < 0.0002) from the control.
Relaxant effect of celiprolol
After pretreatment and in the presence of nadolol, celiprolol
(0.1–300 mol/L) induced a concentration-dependent relaxation of
PCA rings. The maximal response obtained for the highest con-
centration used was 72.38% ± 5.76% (n = 8). This effect was signifi-
cantly antagonized by L-748337 (3 mol/L) and SR 59230A
(10 mol/L) with a rightward shift of the CCRC (P < 0.0001), com-
pare with the control (Fig. 5).
Discussion
In the present study, we show for the first time, by both molec-
ular and functional approaches, the presence of 
3
-adrenoceptors
in PCA. Furthermore, our study provides evidence of the involve-
ment of 
3
-adrenoceptor in the celiprolol-induced vasorelaxation
in this vascular bed.
The predominantly expressed -adrenoceptor in PCA is the 
1
-
adrenoceptor (Nishimura et al. 1987; Toda and Okamura 1990). In
the present study, by using RT-Q-PCR analysis, we clearly demon-
strated that both the 
2
- and 
3
-adrenoceptors are expressed
in PCA. A quantitative determination of mRNA levels for the
different -adrenoceptor subtypes has not been systematically
performed before in the coronary arteries of the pig or of other
animal species. To the best of our knowledge, only one study
reported 
3
-adrenoceptor transcripts in endothelial cells from hu-
man coronary microarteries (Moniotte et al. 2001), suggesting that

3
-adrenoceptors may also play a role in small coronary blood
vessels. Although we have quantified only mRNA expression and
not the receptor protein, the presence of 
3
-adrenoceptor on
endothelial cells of PCA is supported by the combination of RT-
Q-PCR analysis and functional data. They indicate that the 
3
-
adrenoceptor agonist SR 58611A relaxed PCAs through the
endothelium-, NO-dependent pathway, as very weak relaxation
was observed in endothelium-denuded PCA or in the presence of
L-NAME.
In our study, pharmacological evidence for the presence of func-
tional 
3
-adrenoceptor in PCA was reinforced by using 
3
-
adrenoceptor antagonists. The relaxant effect of SR 58611A was
inhibited by pretreatment with 2 selective 
3
-adrenoceptor antago-
nists, SR 59230A and L-748337 (Manara et al. 1995, 1996; Candelore
et al. 1999). L-748337 is actually a selective competitive antagonist of

3
-adrenoceptors in both rat (Mallem et al. 2004) and human (Rozec
et al. 2005) blood vessels but seems to have a higher affinity in vitro
for 
3
-adrenoceptors in humans than in rats (Candelore et al. 1999).
In the same recombinant cell preparations, SR 59230A showed sim-
ilar affinity for human and rat 
3
-adrenoceptors but also exhibited
Fig. 3. The effects of L-748337 and SR 59230A (2 
3
-adrenoceptor
antagonists) upon relaxations induced by SR 58611A (a preferential

3
-adrenoceptor agonist) in the presence of nadolol (a 
1
- and

2
-adrenoceptor antagonist) in intact porcine coronary artery
rings. Cumulative concentration–response curves to SR 58611A
(0.03–100 mol/L) were constructed after 30 min pretreatment with
and in the presence of nadolol (10 mol/L) and in the presence of
L-748337 (3 and 10 mol/L) or SR 59230A (10 mol/L). Each point is
the mean of n experiments obtained from n pig hearts, and vertical
lines show the standard error of the mean. When no error bar is
shown, the error is smaller than the symbol. *, $, and # indicate that
the relaxation effect of L-748337 (10 mol/L), SR 59230A, and
L-748337 (3 mol/L), respectively, was significant different (at
P < 0.01) from that of the SR 58611A control alone and in the
presence of nadolol.
Fig. 4. The effects of L-NAME (a nitric oxide synthase inhibitor) or
endothelium removal upon the relaxations induced by SR 58611A
(a preferential 
3
-adrenoceptor agonist) in the presence of nadolol
(a 
1
- and 
2
-adrenoceptor antagonist) in intact porcine coronary
artery rings. Cumulative concentration–response curves to
SR 58611A (0.03–100 mol/L) were constructed after 30 min of
pretreatment with and in the presence of nadolol (10 mol/L) in
endothelium-intact (control), endothelium-denuded rings, or
endothelium-intact rings pretreated with 100 mol/L L-NAME. Each
point is the mean of n experiments obtained from n pig hearts, and
vertical lines show the standard error of the mean. When no error
bar is shown, the error is smaller than the symbol. * and $ indicate
that the relaxation effect following removal of endothelium and
pretreatment with L-NAME, respectively, was significantly different
(at P < 0.0002) from that of the SR 58611A control.
Pagination not final/Pagination non finale
4 Can. J. Physiol. Pharmacol. Vol. 91, 2013
Published by NRC Research Press
Can. J. Physiol. Pharmacol. Downloaded from www.nrcresearchpress.com by Couperin on 09/06/13
For personal use only.
