- Email: [email protected]
Portal vascular responsiveness to sympathetic stimulation and nitric oxide in cirrhotic rats
Copyright © EuropeanAssociation for the Study of the Liver 1996
Journal of Hepatology 1996," 25.' 90-97 Printed in Denmark. All rights reserved Munksgaard. Copenhagen
Journal of Hepatology
ISSN 0168-8278
Portal vascular responsiveness to sympathetic stimulation and nitric oxide in cirrhotic rats Robert T. Mathie 1, Vera Ralevic2 and Geoffrey Bumstock2 IDepartment of Surgery, Royal Postgraduate Medical School, Hammersmith Hospital, and 2Department of Anatomy & Developmental Biology, University College, London, UK
Aims~Methods: The modulatory role of nitric oxide in portal vasoconstrictor responses was investigated in the isolated perfused liver of cirrhotic rats (induced by carbon tetrachloride/phenobarbitone; n=6). Age-matched (n=5) and phenobarbitone-treated rats (n=5) served as controls. Results: At a constant flow rate of 5 ml/min there was no difference in basal perfusion pressure between the groups. Responses to electrical field stimulation of perivascular nerves caused frequency-dependent increases in perfusion pressure that were not significantly different between the groups. In contrast, dose-dependent vasoconstrictor responses to bolus injections of noradrenaline were up to two-fold greater than those observed in controls (p<0.05). Vasoconstrictor responses to bolus injections of methoxamine (a selective ~l" adrenoceptor agonist) or adenosine 5'-triphosphate (ATP, a cotransmitter with noradrenaline in sympathetic nerves) were dose-dependent and similar between the groups. Infusion of the nitric oxide synthesis inhibitor NG-nitro-L-arginine methyl ester (L-NAME; 30 p~M) had no effect on basal tone or on responses to electrical field stimu-
lation or injected agents. A step-wise increase in flow to 10, 15 and 20 ml/min produced a similar increase in perfusion pressure within each group. At increased flow, there was a decrease in responsiveness to noradrenaline (5 nmol) in preparations from all groups. In the presence of the K ÷ channel inhibitor glibenclamide (5 ~tM), the effect of noradrenaline in the cirrhotic group at flow rates of 5, 10 and 15 ml/min was maintained to a significantly greater extent than in either control group, suggesting that ATP-sensitive K ÷ channels in the portal venous bed may be activated in cirrhosis. Conclusions: We conclude that portal vasoconstriction associated with noradrenaline, but not with sympathetic nerve stimulation, methoxamine or ATP, is enhanced in cirrhosis. Nitric oxide does not appear to play a modulatory role in these responses.
HEHAEMODYNAMICSof advanced cirrhosis are char-
splanchnic venous network. The syndrome is often associated with increased plasma noradrenaline (NA) caused by increased sympathetic nervous activity (25). Mesenteric vasodilatation is believed to be primarily responsible for the low peripheral vascular resistance of decompensated liver disease, but the precise mechanism is unknown. In addition to impaired reactivity to perivascular sympathetic stimulation, a number of mechanisms have been proposed, including increased blood levels of vasodilators such as glucagon, prostacyclin or adenosine (6,7), or diminished responsiveness to a variety of vasoconstrictor agents (8,9). There is evidence that increased nitric
T acterized by raised cardiac output, decreased peripheral vascular resistance and arterial blood pres-
sure, and a hyperdynamic pre-hepatic splanchnic circulation which contributes to the portal hypertension that originates from a primary hepatic lesion (1). Portal blood flow to the liver is diminished because of the development of an extrahepatic collateral Received 13 September; revised 14 November, accepted 30 November 1995
Correspondence: Dr. R. T. Mathie, Department of Surgery, Royal Postgraduate Medical School, London W12 0NN, UK. Telephone: 44 181 743 2030 (ext. 2267). Facsimile: 44 181 740 3179. 90
Key words: Adenosine 5'-triphosphate; Cirrhosis; Glibenclamide; Methoxamine; Nitric oxide (NO); NG-nitro-L-arginine methyl ester (L-NAME); Noradrenaline; Portal system; Sympathetic nerves.
Portal vascular responsiveness in cirrhosis oxide (NO) production may be involved, though a lack of consensus prevails (9). The recognition that the resistance of the intrahepatic portal vascular bed in cirrhosis may be amenable to pharmacological modulation (10,11) focuses attention on the liver not only as a primary therapeutic target but as a potentially important contributor to the peripheral haemodynamic alterations. Several groups have demonstrated increased responsiveness to NA in the portal vascular bed of the cirrhotic liver (12-15), and it has been shown by one group (15,16) that NO plays a modulatory (vasodilatory) role in both the cirrhotic and the normal animal. While basal hepatic haemodynamics are probably not normally under significant sympathetic control (17), it is unknown whether there is increased responsiveness to perivascular sympathetic nervous activity in the portal bed of the cirrhotic liver. Vascular shear stress, which is likely to be increased in cirrhosis, may activate ATP-sensitive K ÷ channels in the arterial vasculature (18), and it has been shown that increased activation of K ÷ channels may indeed contribute to peripheral vasodilatation in cirrhosis (19). However, nothing is known about the role of K + channels in the portal vascular bed nor their impact in cirrhosis. The aim of the present study was therefore to investigate further the vasoconstrictive responsiveness of the portal vascular bed in cirrhosis, specifically to electrical field stimulation (EFS), NA, the selective C~l-adrenoceptor agonist methoxamine, and the cotransmitter adenosine 5"-triphosphate (ATP). The liver was isolated from rats made cirrhotic by administration of carbon tetrachloride (CC14)/phenobarbitone. We have examined the haemodynamic changes due to NO inhibition by NG-nitro-L-arginine methyl ester (L-NAME). The effects of L-NAME and the ATP-sensitive K + channel inhibitor glibenclamide on pressure changes with increased perfusate flow, and the effect of increased flow on constrictor responses to NA, were also determined.
Materials and Methods Experimental groups Cirrhosis was induced in seven male Wistar rats (140-160 g), by a modification of the method of Proctor & Chatamra (20) as previously described (21). Animals were allowed free access to drinking water containing 350 mg/1 phenobarbitone for 2 weeks before treatment with CC14 and thereafter. CCI 4 was administered by gavage weekly for 3-4 months, with a starting dose of 40 gl increasing to a maximum of 600 gl. Cirrhosis was judged to have
occurred when abdominal ascites was present as indicated by a sudden increase in body weight and by clinical inspection. A further five rats were given phenobarbitone alone, and a group of five agematched animals served as untreated controls. All procedures were carried out under the terms of a UK Home Office Animal Project Licence.
Isolated hepatic portal vascular bed The rats were killed by asphyxiation in C O 2. The portal vein was cannulated with a metal cannula (internal diameter: 1.0 mm), taking care to ensure placement of the tip just below the bifurcation and to secure the ligature above the gastroduodenal vein. The isolated liver was mounted on a stainless steel grid (7 cmx5 cm) in a humid chamber. The preparation was perfused at a constant rate of 5.0 ml/min with Krebs' solution containing (mM): NaC1 133, KC1 4.7, NaHzPO 4 1.35, NaHCO 3 16.3, MgSO 4 0.61, glucose 7.8 and CaC12 2.52, gassed with 95% O 2 and 5% CO 2, and maintained at 37°C. Responses were measured with a pressure transducer (model P23XL; Gould) on a side arm of the perfusion cannula, and recorded on a polygraph (model 79D; Grass); pressure attributable to the cannula alone (4 mmHg) was subtracted from all recordings to provide a measurement of vascular tone. Drugs were administered as 50 ~tl bolus injections. Preparations were allowed to equilibrate for approximately 30 min prior to experimentation. Experimental protocol EFS (90V, 1 ms pulse width, for 30 s) was applied at increasing frequencies (4-32 Hz) to obtain a frequency-response curve. Vasoconstrictor responses of preparations were then tested to increasing doses (50 ~1 bolus injections) of NA, methoxamine and ATE Individual doses were applied at intervals of 2-5 min, depending on the time it took for the tone to return to baseline. Approximately 10 min were allowed between consecutive dose-response curves. Flow/ pressure relationships of the preparations were then studied by incrementally increasing the perfusion rate from 5 to 10, 15 and finally 20 ml/min. At each flow rate, responses to a single dose of NA (5 nmol; approximately the EDs0 dose) were tested as soon as the preparation had stabilized at its new level of tone. Flow was returned to 5 ml/min and the preparation equilibrated with L-NAME (30 gM) for 30 min, after which response curves to EFS, NA, methoxamine and ATP were established, followed by a repeat of the step-wise increases in flow and injections of NA. The preparation was then equilibrated with L-arginine (300 gM) 91
R. T. Mathie et al.
ferences between groups was carried out by one-way analysis of variance; where evidence for a statistically significant effect was detected (p<0.05) a modified t-test (22) was then used to compare cirrhotic data with those of age-matched and phenobarbitonetreated controls. Within each group, the effects of LNAME and L-arginine on vasodilator responses (mmHg) were compared with equivalent data obtained in their absence by Student's paired t-test.
for 20 min (still in the presence of L-NAME) and dose-responses to EFS and NA were re-established. The NA dose-responses also served as an index of the viability of the preparation with time; maintained responsiveness to NA was a prerequisite for continuing each experiment. After washout of all drugs, the preparation was equilibrated with glibenclamide (5 ~tM) for 15 min, and the flow/pressure relationships and NA responses reassessed. Immediately after the end of the perfusion experiment, a segment from each of two liver lobes was placed in 4% paraformaldehyde for subsequent histological examination. The portal bed of two further normal rats was used to obtain frequency-response curves to EFS in the absence and then in the presence of the selective c~]adrenoceptor antagonist prazosin (10 -6 M).
Results Cirrhotic model
Of the seven rats treated with CC14, only six were used in the present study based on the development of fibrosis and regenerative nodules in the liver, as revealed by histological examination. All six rats had mild/moderate ascites on opening the abdomen (5-10 days after the original clinical diagnosis of ascites).
Drugs
Adenosine 5'-triphosphate (disodium salt), methoxamine (hydrochloride), noradrenaline (bitartrate), phenobarbitone (sodium salt), glibenclamide, prazosin, NG-nitro-L-arginine methyl ester hydrochloride and L-arginine were obtained from Sigma, Poole, England. All drugs were made up in distilled water, except for NA which was made up as a stock solution of 10 mM in 0.1 mM ascorbic acid and diluted in distilled water, and glibenclamide which was made up as a stock solution of 10 mM in DMSO.
Basal portal vascular tone
Basal tone did not differ significantly between the three experimental groups (controls: 0.4_+0.24 mmHg; phenobarbitone: 0.8_+0.48 mmHg; cirrhosis: 1.5+0.72 mmHg; see also Table 2). Electrical field stimulation
EFS (4-32 Hz) elicited frequency-dependent vasoconstrictor responses which were similar in magnitude in all groups (Fig. 1a). The addition of L-NAME Fig. lb) or L-arginine had no significant effect on the frequency-response curve in any group. In two nor-
Data analysis
All results are expressed as the mean_+one standard error of the mean (s.e.m). Statistical analysis of difE f f e c t o f E F S in c i r r h o s i s Basal
E f f e c t o f E F S in c i r r h o s i s L-NAME
6
6
:E ES
I ~S
v
~4
~4
o- 3
o- 3
~2
2
c_
f-
(D 0
4
8
12
16
20
Frequency (Hz)
24
28
32
36
0
b
4
8
12
16
20
24
28
32
36
Frequency (Hz)
Fig. 1. Frequency-dependent constrictor responses to electrical field stimulation (EFS; 4-32 Hz, 90V, 1 ms, for 30 s duration) of portal vascular bed preparations from age-matched control (-s-), phenobarbitone-treated (-u-) and cirrhotic (-u-) rats in (a) the absence and (b) the presence of NC-nitro-L-arginine methyl ester (L-NAME). "Basal" denotes responses in the absence of drugs in the perfusate. There were no significant differences between the groups in either (a) or (b). 92
Portal vascular responsiveness in cirrhosis Effect of n o r a d r e n a l i n e in cirrhosis Basal
Effect of n o r a d r e n a l i n e in cirrhosis L-NAME
14
14
!
Z
E12
12
g
10
~10 co co
Q_
Q_
© ~L
©
8
*
6
4
li
4 (1) C~
2
x~
2
o
© I
0 -10
I
h
~
I
-9
,
-8
,
I
I
i
-7
0 -10
-6
L
i
I
L
i
I
-9
NA (log mol)
I
:
I
I
-8
I
I
-7
~
I
-6
NA (log tool)
Fig. 2. Dose-response curves to noradrenaIine (NA) of portal vascular bed preparations from age-matched control (-d-), phenobarbitone-treated (-u-) and cirrhotic (-,-) rats in (a) the absence and (b) the presence of N~-nitro-L-arginine methyl ester (L-NAME). "Basal" denotes responses in the absence of drugs in the perfusate. Differences between cirrhotic and agematched or phenobarbitone-treated controls are denoted by * (p<0.05). Effect of ATP in cirrhosis Basal
Effect of ATP in cirrhosis L-NAME
12
12
~z El0 ~
8
~ 8
~
6
o_
co
©
~ c
6
©
4
Q- 4 c
~2 O3 X:Z
xz
©
© v
-10
-9
-8
-7
-6
ATP (log tool)
-5
o -10
.9
-8
-7
-6
-5
ATP (log mol)
Fig. 3. Dose-response curves to adenosine 5'-triphosphate (ATP) of portal vascular bed preparations from age-matched control (~t-), phenobarbitone-treated (-u-) and cirrhotic (-u-) rats in (a) the absence and (b) the presence of NC-nitro-L-ar ginine methyl ester (L-NAME). "Basal" denotes responses in the absence of drugs in the perfusate. There were no significant differences between the groups in either (a) or (b). In the age-matched group only, L-NAME caused a significant reduction in response to -5. 75 log tool and -5.25 log mol ATP (paired t-test; p
mal livers, prazosin (10-6M) completely abolished vasoconstrictor responses at all stimulation frequencies, showing that these were mediated by NA at post-junctional o~l-adrenoceptors, and therefore involved sympathetic nerves. Noradrenaline
Noradrenaline elicited dose-dependent vasoconstrictor responses in the portal bed, which were significantly greater in the preparations from cirrhotic rats
compared with either controls or phenobarbitonetreated animals at doses o f - 7 . 2 5 log tool (p<0.05), 26.75 log mol (p=0.01) and -6.25 log mol (p<0.01) (Fig. 2a). These responses were not affected by the addition of L-NAME to the perfusate, though the significant difference between cirrhotic and control groups was extended down to NA doses o f - 8 . 2 5 log mol (p<0.05) a n d - 7 . 7 5 log mol (p<0.05) (Fig. 2b). L-arginine had no significant effect on responses in any group. 93
R. T. Mathie et al.
Methoxamine Bolus injections of methoxamine elicited dosedependent vasoconstrictor responses that were similar between the groups and were unaffected by LNAME (Table 1). Adenosine 5"-triphosphate Injections of ATP elicited similar dose-dependent constrictions in cirrhotic, control and phenobarbitone-treated groups (Fig. 3a). L-NAME had no significant effect on responses to ATE except in the control group where an attenuation in response at the top
TABLE 1 Effect of bolus injections of methoxamine on portal vascular tone in three groups of rats under basal perfusion conditions and during infusion of L-NAME (30 ~M). Data are mean_+SEM Methoxamine (nmol)
Increase in perfusion pressure (mmHg) Basal
L-NAME
Controls
0.5 5 50 500 5000
0 0.6_+0.24 2.0-+0.35 4.1_+0.83 4.7+1.45
0 0 1.6_+0.24 3.1_+0.59 3.3_+0.93
Phenobarb.
0.5 5 50 500 5000
0 0.1-+0.13 1.6_+0.63 3.9-+0.43 3.7+0.33
0.1_+0.13 1.5_+0.79 2.1_+0.52 3.5+0.20 3.8_+0.25
Cirrhosis
0.5 5 50 500 5000
0.4_+0.33 0.6_+0.30 2.1_+0.52 4.3_+0.95 5.2_+1.33
0.5_+0.34 0.7_+0.31 1.6_+0.44 2.8_+0.46 4.3_+0.72
TABLE 2 Effect of flow rate on portal vascular tone in three groups of rats under basal conditions, and during infusion of L-NAME (30 p.M) or glibenclamide (5 laM). Data are mean_+SEM Flow rate
Perfusion pressure (mmHg)
(ml/min)
Basal
Controls
5 10 15 20
0.4_+0.24 2.4_+0.40 5.1_+0.68 6.2_+0.86
0.4_+0.24 2.0_+0.35 4.2_+0.64 5.5_+0.89
0.4_+0.24 2.1_+0.24 4.6_+0.68 5.8_+1.08
Phenobarb.
5 10 15 20
0.8_+0.48 3.5_+0.87 6.5_+1.19 9.0_+2.16
0.8+0.48 3.5_+1.19 8.3_+1.11 10.8_+1.80
0.8_+0.48 3.1_+0.72 6.3_+1.20 9.0~_2.13
Cirrhosis
5 I0 15 20
1.5_+0.72 3.7_+0.80 6.7_+1.28 8.3_+1.71
94
L-NAME
TABLE 3 Effect of bolus injection of noradrenaline (5 nmol) on portal vascular tone in three groups of rats under basal conditions and during infusion of L-NAME (30 I.tM) or glibenclarnide (5 gM). Data are mean_+SEM Flow rate
Increase in perfusion pressure (mmHg)
(ml/min)
Basal
L-NAME
Glibenclamide
Controls
5 10 15 20
4.3_+0.63 3.6_-+-0.55 3.0_-+0.41 2.4_+0.38
3.8+0.43 2.6+0.55 2.4_-!-0.47 2.4_+0.47
2.4_+0.24 2.0+0 1.8_+0.14 1.5_+0.20
Phenobarb.
5 10 15 20
3.3+0.25 3.3_+0.60 2.~_-0.35 1.8+0.43
4.3_+0.25 3.2_+0.17 2.3_+0.48 2.1_+0.31
2.8_+0.32 2.4_+0.24 1.9_+0.13 1.9_+0.13
Cirrhosis
5 10 15 20
5.4_+0.80 3.9_+0.40 3.6_+0.48 3.2_+0.72
6.2_+1.07 4.7_+0.73 4.1 +0.64 3.5_+0.79
5.3_+0.66 4.0_+0.57 3.5_+0.69 3.1_+1.03
two doses of ATP was noted (paired t-test; p<0.05) (Fig. 3b).
Flow rate Stepwise increases in flow rate (up to 20 ml/min) produce an increase in perfusion pressure that was not statistically significantly different between the groups, and was independent of the type of agent in the perfusate (Table 2). Noradrenaline with flow There was a progressive decrease in responsiveness to NA with increased flow in all groups for each of the three phases of the experiment (basal, L-NAME and glibenclamide) (Table 3). Responses to NA appeared to be greater in the cirrhotic group at all flow rates under basal conditions, but this did not reach statistical significance. There was no significant effect of L-NAME on responses to NA. In the presence of glibenclamide, responses of both the control groups were significantly reduced compared to those in the cirrhotic group, which were maintained.
Glibenclamide
1.5_+0.72 1.5_+0.72 4 . 8 _ + 1 . 1 5 4.9-+1.08 7 . 6 - + 1 . 8 4 8.8_+1.52 9.4_+2.19 11.4_+2.39
Discussion Our results show that, compared to controls, at a portal flow rate of 5 ml/min, the portal vascular bed of cirrhotic rats is significantly more responsive to NA, whereas responsiveness to EFS, methoxamine and ATP is unchanged. Inhibition of NO synthesis did not augment these responses, in contrast to the potentiation observed in many other vessels and preparations, including the mesenteric bed of normal, cirrhotic and portal hypertensive rats (23,24). Enhancement of responsiveness to NA in the por-
Portal vascular responsiveness in cirrhosis
tal bed of the cirrhotic rat liver has been noted previously (12-15). Explanations for this effect include: an increase in the size and contractility of the intramural smooth muscle due to the sustained rise in intraluminal pressure associated with cirrhosis; the presence of activated myofibroblast-like (Ito) cells surrounding thin-walled venous channels and sinusoids in the cirrhotic liver (14). The lack of potentiation of responses to the c~l-adrenoceptor agonist methoxamine in the present study suggests that another possible explanation may be selective changes in adrenoceptor subtypes in cirrhosis. This hyper-reactivity to NA in the cirrhotic portal bed is in contrast to the hyporesponsiveness to vasoconstrictor agents known to occur in other regions of the peripheral vasculature in cirrhosis (8,9). Our results suggest that there is no increase in hepatic sympathetic nervous function in experimental cirrhosis. While this conclusion is consistent with that of Abergel et al. (25), it is surprising in view of the enhancement in post-junctional responses to NA (though not those to methoxamine) in our cirrhotic experiments. Only a very modest increase in responsiveness to sympathetic nerve stimulation in cirrhotic liver was noted by Zimmermann et al. (13). One explanation for the discrepancy between nerve stimulation and NA in cirrhosis may be the responsiveness of the activated Ito cells, whose contractile apparatus may react to intraluminally applied NA but not to neural stimulation. Our results with the oq-adrenoceptor antagonist prazosin also appear to rule out the possibility of an agent in addition to NA acting as a neurotransmitter in the portal vascular bed of the normal rat liver. However, cotransmission of a modulatory agent with NA from sympathetic nerves in cirrhosis cannot be excluded. Cotransmission of ATP with NA is a recognised feature of sympathetic nerve activity that has been noted in both the isolated portal vein (26) and hepatic artery (27). The implication of our results is that the raised portal vascular resistance of cirrhosis in vivo may be partly attributed to the elevated plasma levels of NA, and not to increased hepatic portal sympathetic nervous activity per se. Induction by endotoxin of NO synthase in vascular smooth muscle may contribute to the peripheral vasodilatation of decompensated liver disease (28). Much of the evidence for increased NO synthase induction has been indirect: inhibition of NO synthesis partly reversed the hyporeactivity to vasoconstriction produced by methoxamine (29,30), and reversed peripheral vasodilatation in cirrhotic rats (31,32). These findings, however, have been both corroborated and unsupported by clinical investigations (33-
35), and some of the more recent experimental work has been unable to confirm a role for NO in portal hypertension (36,37). There is now increasing evidence that changes in K ÷ or Ca 2+ channels may contribute importantly to the peripheral vasodilatation of cirrhosis (19,38). In this study we examined evidence for changes in the role of NO in the portal bed of cirrhotic rats. Our results with NA differ from those reported by Mittal et al. (15,16), in that these authors found evidence for basal release of NO which modulated the NA response similarly in cirrhotic and control portal bed preparations. We found no evidence of basal NO release in the portal bed of either control or cirrhotic animals since L-NAME (30 gM) produced no augmentation of responses to NA or nerve stimulation. It is unlikely that higher concentrations of L-NAME would alter the results, since there was not even a trend towards enhancement of responses to NA or nerve stimulation in the presence of L-NAME. Furthermore, it has been demonstrated that the IC50 for L-NAME is 2.5 gM against constitutive NO synthase and 20-25 ~tM against inducible NO synthase (39), confirming that a concentration of 30 gM will produce a very substantial inhibition of NO synthesis via either isoform. The discrepancies may, however, relate to the supranormal portal flow rate used by Mittal et al. (40 ml/min), which would have induced a greater shear stress (and thus NO release) than in the current study. Although at the lower end of the normal range in vivo, the basal flow rate of 5 ml/min used in our in vitro experiments resulted in portal pressure recordings similar to those achieved in other perfused rat liver preparations at higher flow rates (40). Our own data showed only a modest increase in portal pressure at increased flow rates, findings also consistent with the highly distensible nature of the portal vascular bed (41). That our preparation was not compromised pharmacologically at 5 ml/min is indicated by the optimal vasoconstriction to bolus injections of NA observed at that flow rate: at higher flow rates there was a progressive decrease in responses (vide infra). L-NAME had no effect on perfusion pressure even at a flow rate of 20 ml/min, suggesting that NO release was independent of flow in our particular experimental system. We do not know whether shear-evoked NO release would be apparent at still higher flows such as the 40 ml/min used by Mittal et al. The important point of agreement between our studies and those of Mittal et al., however, is that portal NO release is not greater in cirrhotic compared with normal liver. This indicates, therefore, that the portal bed 95
R. T. Mathie et al.
is unlikely to participate in any NO-mediated peripheral vasodilatation. In contrast to the augmentation of ATP-mediated vasoconstriction by NO inhibition in the perfused rabbit liver (42), constrictor responses to ATP in the present study were not significantly increased by LNAME in either normal or cirrhotic livers. The magnitude of basal responses to ATP in both series was similar, as was the concentration of L-NAME used to inhibit NO synthesis. Other than species differences and that the hepatic artery was not perfused in the current experiments, therefore, an explanation for the different findings in the two experiments is not obvious. The effect of ATP in the portal bed of cirrhotic animals has not previously been examined. Assessment of the perfusion pressure at different flow rates in our study revealed a similar increase in pressure with step-wise increases in flow between the groups, in contrast to the steeper slope describing the pressure-flow relationship in cirrhosis noted by Zimm o n & Kessler (43). The perfusion pressure at different flow rates was not modulated by L-NAME or glibenclamide. Examination of the constrictor responses of preparations to a single, approximately EDs0, dose of NA at different flow rates revealed that responses were progressively smaller at higher flow rates in all groups. It appears, however, that increased pressure/flow or inhibition of NO at different flow rates has little effect on the hyper-responsiveness to NA in cirrhosis compared with controls. The greater constrictor response to NA of cirrhotic preparations compared with controls in the presence of glibenclamide may be a consequence of activation of K + channels in the portal venous bed in cirrhosis via an unknown mechanism; evidence for the activation of K ÷ channels in the peripheral arterial circulation of cirrhotic rats has been presented previously (19). On the other hand, the maintained responses to NA in the presence of glibenclamide in cirrhotic preparations may simply be a reflection of a sustained hyperresponsiveness to NA in this group. In general, there was good viability of the preparations over the course of the experiment, as evidenced by maintained responsiveness to NA with time, and there is no reason to suppose that the long-term viability of the control preparations was less than that of the cirrhotics. We conclude that portal vasoconstriction associated with NA, but not sympathetic nerve stimulation, methoxamine or ATE is enhanced in experimental cirrhosis. NO does not appear to play a modulatory role. Whether these changes are representative of circumstances in cirrhosis in humans remains to be determined. 96
Acknowledgements The support of the British Heart Foundation and the Hammersmith & Queen Charlotte's Special Health Authority is gratefully acknowledged. We are grateful to Mrs M. Millbourn and Mr J. Godwin, Biological Services Unit, Royal Postgraduate Medical School, for technical help.
References 1. Bosch J, Pizcueta P, Feu E Fernfindez M, Garcfa-Pag~n JC. Pathophysiology of portal hypertension. Gastroenterol Clin North Am 1992; 21: 1-14. 2. Ring-Larsen H, Hesse B, Henriksen JH, Christensen HJ. Sympathetic nervous activity and renal and systemic hemodynamics in cirrhosis: plasma norepinephrine concentration, hepatic extraction, and renal release. Hepatology 1982; 2: 304-10. 3. Bichet DG, Van Putten VJ, Schrier RW. Potential role of increased sympathetic activity in impaired sodium and water excretion in cirrhosis. New Engl J Med 1982; 307: 1552-7. 4. Esler M, Dudley F, Jennings G, Debinski H, Lambert G, Jones P, Crotty B, Colman J, Willett I. Increased sympathetic nervous activity and the effects of its inhibition with clonidine in alcoholic cirrhosis. Ann Intern Med 1992; 116: 44655. 5. Floras JS, Legault L, Morali GA, Hara K, Blendis LM. Increased sympathetic outflow in cirrhosis and ascites: direct evidence from intraneural recordings. Ann Intern Med 1991; 114: 373-80. 6. Groszmann RJ. Hyperdynamic circulation of liver disease 40 years later: pathophysiology and clinical consequences. Hepatology 1994; 20: 1359-63. 7. Stark ME, Szurszewski JH. Role of nitric oxide in gastrointestinal and hepatic function and disease. Gastroenterology 1992; 103: 1928-49. 8. Ryan J, Sudhir K, Jennings G, Esler M, Dudley E Impaired reactivity of the peripheral vasculature to pressor agents in alcoholic cirrhosis. Gastroenterology 1993; 105:1167-72. 9. Bomzon A, Blendis LM. The nitric oxide hypothesis and the hyperdynamic circulation in cirrhosis. Hepatology 1994; 20: 1343-50. 10. Bhathal PS, Grossman HJ. Reduction of the increased portal vascular resistance of the isolated perfused cirrhotic rat liver by vasodilators. J Hepatol 1985; 1: 325-37. 11. Reichen J. Liver function and pharmacological considerations in pathogenesis and treatment of portal hypertension. Hepatology 1990; 11: 1066-78. 12. Ballet F, Chretien Y, Rey C, Poupon R. Differential response of normal and cirrhotic liver to vasoactive agents. A study in the isolated perfused rat liver. J Pharmacol Exp Ther 1988; 244: 283-9. 13. Zimmermann T, Gardemann A, Machnik G, Dargel R, Jungermann K. Metabolic and hemodynamic responses of bivascularly perfused rat liver to nerve stimulation, noradrenaline, acetylcholine and glucagon in thioacetamide-induced micronodular cirrhosis. Hepatology 1992; 15: 464-70. 14. Grossman HJ, Grossman VL, Bhathal PS. Enhanced vasoconstrictor response of the isolated perfused cirrhotic rat liver to humoral vasoconstrictor substances found in portal venous blood. J Gastroenterol Hepatol 1992; 7: 283-7.
Portal vascular responsiveness in cirrhosis
15. Mittal MK, Lee F-Y, Groszmann RJ. Role of nitric oxide in the intrahepatic circulation of cirrhotic rats. Gastroenterology 1993; 104: A956. 16. Mittal MK, Gupta TK, Lee F-Y, Sieber CC, Groszmann RJ. Nitric oxide modulates hepatic vascular tone in normal rat liver. Am J Physiol 1994; 267: G416-22. 17. Wheatley AM, Stuart ET, Zhao D, Zimmermann A, Gassel H-J, Blumgart LH. Effect of orthotopic transplantation and chemical denervation of the liver on hepatic hemodynamics in the rat. J Hepatol 1993; 19: 442-50. 18. Hassdssian H, Bodin P, Burnstock G. Blockade by glibenclamide of the flow-evoked endothelial release of ATP that contributes to vasodilatation in the pulmonary vascular bed of the rat. Br J Pharmacol 1993; 109: 466-72. 19. Moreau R, Komeichi H, Kirstetter P, Ohsuga M, Cailmail S, Lebrec D. Altered control of vascular tone by adenosine triphosphate-sensitive potassium channels in rats with cirrhosis. Gastroenterology 1994; 106: 1016-23. 20. Proctor E, Chatamra K. High yield micronodular cirrhosis in the rat. Gastroenterology 1982; 83:1183-90. 21. Firth JD, Gove C, Panos MZ, Raine AEG, Williams R, Ledingham JGG. Sodium handling in the isolated perfused kidney of the cirrhotic rat. Clin Sci 1989; 77: 657-61. 22. Altman DG. In: Altman DG. Practical Statistics for Medical Research. London: Chapman & Hall, 1991: 210-2. 23. Ralevic V, Mathie RT, Moore KP, Burnstock G. Vasoconstrictor responsiveness of the rat mesenteric arterial bed in cirrhosis. Br J Pharmacol 1996; in press. 24. Sieber CC, Groszmann RJ. Nitric oxide mediates hyporeactivity to vasopressors in mesenteric vessels of portal hypertensive rats. Gastroenterology 1992; 103: 235-9. 25. Abergel A, Braillon A, Gaudin C, Kleber G, Lebrec D. Persistence of a hyperdynamic circulation in cirrhotic rats following removal of the sympathetic nervous system. Gastroenterology 1992; 102: 656-60. 26. Burnstock G, Crowe R, Kennedy C, T6rrk J. Indirect evidence that purinergic modulation of perivascular adrenergic neurotransmission in the portal vein is a physiological process. Br J Pharmacol 1984; 82: 359-68. 27. Brizzolara AL, Burnstock G. Evidence for noradrenergicpurinergic cotransmission in the hepatic artery of the rabbit. Br J Pharmacol 1990; 99: 835-9. 28. Whittle BJR, Moncada S. Nitric oxide: The elusive mediator of the hyperdynamic circulation of cirrhosis? Hepatology 1992; 16: 1089-92. 29. Sieber CC, Groszmann RJ. In vitro hyporeactivity to methoxamine in portal hypertensive rats: reversal by nitric oxide blockade. Am J Physiol 1992; 262: G996-1001. 30. Lee F-Y, Albillos A, Colombato LA, Groszmann RJ. The role of nitric oxide in the vascular hyporesponsiveness to methoxamine in portal hypertensive rats. Hepatology 1992; 16: 1043--48.
31. Pizcueta E Piqu6 JM, Fern~indez M, Bosch J, Rodrs J, Whittle BJR, Moncada S. Modulation of the hyperdynamic circulation of cirrhotic rats by nitric oxide inhibition. Gastroenterology 1992; 103: 1909-15. 32. C15xia J, Jimrnez W, Ros J, Asbert M, Castro A, Arroyo V, Rivera F, Rodrs J. Pathogenesis of arterial hypotension in cirrhotic rats with ascites: role of endogenous nitric oxide. Hepatology 1992; 15: 343-9. 33. Guarner C, Soriano G, Tomas A, Bulbena O, Novella MT, Balanzo J, Villardell F, Mourelle M, Moncada S. Increased serum nitrite and nitrate levels in patients with cirrhosis: relationship to endotoxemia. Hepatology 1993; 18: 1139-43. 34. Hori N, Takahashi H, Okanoue T, Mori T, Sakamoto S, Nanbu A, Yoshimura M, Kashima K. Nitric oxide production in patients with chronic liver diseases. Hepatology 1993; 18: 101A. 35. Le Moine O, Soupison T, Sogni P, Husson C, Cruziaux A, Marchant A, Moreau R, Hadengue A, Goldman M, Deviate J, Lebrec D. The role of endotoxins and TNFc~ in the hyperkinetic circulation of cirrhosis. J Hepatol 1993; 18 (Suppl 1): $40. 36. Fern~indez M, Garcfa-Pag~in JC, Casadevall M, Bernardich C, Piera C, Whittle BJR, Piqu6 JM, Bosch J, Rodrs J. Evidence against a role for inducible nitric oxide synthase in the hyperdynamic circulation of portal-hypertensive rats. Gastroenterology 1995; 108: 1487-95. 37. Soubrane O, Lacronique V, Miquerol L, Mignon A, Houssin D. The role of inducible NO synthase (iNOS) in the hyperdynamic circulation of portal hypertension. Hepatology 1995; 22: 156A. 38. Cailmail S, Moreau R, Gadano A, Tazi KA, Lebrec D. The role of calcium channel inhibition in in vitro vascular hyporeactivity to endothelin-1 (ET-1) in rats with cirrhosis. Hepatology 1995; 22: 258A. 39. Mitchell JA, Kohlhaas KL, Sorrentino R, Warner TD, Murad F, Vane JR. Induction by endotoxin of nitric oxide synthase in the rat mesentery: lack of effect on action of vasoconstrictors. Br J Pharmacol 1993; 109: 265-70. 40. Grossman HJ, Grossman VL, Bhathal PS. Hemodynamic characteristics of the intrahepatic portal vascular bed over an extended flow range: a study in the isolated perfused rat liver. Hepatology 1995; 21: 162-8. 41. Lautt WW, Legare DJ. Passive autoregulation of portal venous pressure: distensible hepatic resistance. Am J Physiol 1992; 263: G702-8. 42. Browse DJ, Mathie RT, Benjamin IS, Alexander B. The transhepatic action of ATP on the hepatic arterial and portal venous vascular beds of the rabbit: the role of nitric oxide. Br J Pharmacol 1994; 113: 987-93. 43. Zimmon DS, Kessler RE. Effect of portal venous blood flow diversion on portal pressure. J Clin Invest 1980; 65: 138897.
97
Journal of Hepatology 1996," 25.' 90-97 Printed in Denmark. All rights reserved Munksgaard. Copenhagen
Journal of Hepatology
ISSN 0168-8278
Portal vascular responsiveness to sympathetic stimulation and nitric oxide in cirrhotic rats Robert T. Mathie 1, Vera Ralevic2 and Geoffrey Bumstock2 IDepartment of Surgery, Royal Postgraduate Medical School, Hammersmith Hospital, and 2Department of Anatomy & Developmental Biology, University College, London, UK
Aims~Methods: The modulatory role of nitric oxide in portal vasoconstrictor responses was investigated in the isolated perfused liver of cirrhotic rats (induced by carbon tetrachloride/phenobarbitone; n=6). Age-matched (n=5) and phenobarbitone-treated rats (n=5) served as controls. Results: At a constant flow rate of 5 ml/min there was no difference in basal perfusion pressure between the groups. Responses to electrical field stimulation of perivascular nerves caused frequency-dependent increases in perfusion pressure that were not significantly different between the groups. In contrast, dose-dependent vasoconstrictor responses to bolus injections of noradrenaline were up to two-fold greater than those observed in controls (p<0.05). Vasoconstrictor responses to bolus injections of methoxamine (a selective ~l" adrenoceptor agonist) or adenosine 5'-triphosphate (ATP, a cotransmitter with noradrenaline in sympathetic nerves) were dose-dependent and similar between the groups. Infusion of the nitric oxide synthesis inhibitor NG-nitro-L-arginine methyl ester (L-NAME; 30 p~M) had no effect on basal tone or on responses to electrical field stimu-
lation or injected agents. A step-wise increase in flow to 10, 15 and 20 ml/min produced a similar increase in perfusion pressure within each group. At increased flow, there was a decrease in responsiveness to noradrenaline (5 nmol) in preparations from all groups. In the presence of the K ÷ channel inhibitor glibenclamide (5 ~tM), the effect of noradrenaline in the cirrhotic group at flow rates of 5, 10 and 15 ml/min was maintained to a significantly greater extent than in either control group, suggesting that ATP-sensitive K ÷ channels in the portal venous bed may be activated in cirrhosis. Conclusions: We conclude that portal vasoconstriction associated with noradrenaline, but not with sympathetic nerve stimulation, methoxamine or ATP, is enhanced in cirrhosis. Nitric oxide does not appear to play a modulatory role in these responses.
HEHAEMODYNAMICSof advanced cirrhosis are char-
splanchnic venous network. The syndrome is often associated with increased plasma noradrenaline (NA) caused by increased sympathetic nervous activity (25). Mesenteric vasodilatation is believed to be primarily responsible for the low peripheral vascular resistance of decompensated liver disease, but the precise mechanism is unknown. In addition to impaired reactivity to perivascular sympathetic stimulation, a number of mechanisms have been proposed, including increased blood levels of vasodilators such as glucagon, prostacyclin or adenosine (6,7), or diminished responsiveness to a variety of vasoconstrictor agents (8,9). There is evidence that increased nitric
T acterized by raised cardiac output, decreased peripheral vascular resistance and arterial blood pres-
sure, and a hyperdynamic pre-hepatic splanchnic circulation which contributes to the portal hypertension that originates from a primary hepatic lesion (1). Portal blood flow to the liver is diminished because of the development of an extrahepatic collateral Received 13 September; revised 14 November, accepted 30 November 1995
Correspondence: Dr. R. T. Mathie, Department of Surgery, Royal Postgraduate Medical School, London W12 0NN, UK. Telephone: 44 181 743 2030 (ext. 2267). Facsimile: 44 181 740 3179. 90
Key words: Adenosine 5'-triphosphate; Cirrhosis; Glibenclamide; Methoxamine; Nitric oxide (NO); NG-nitro-L-arginine methyl ester (L-NAME); Noradrenaline; Portal system; Sympathetic nerves.
Portal vascular responsiveness in cirrhosis oxide (NO) production may be involved, though a lack of consensus prevails (9). The recognition that the resistance of the intrahepatic portal vascular bed in cirrhosis may be amenable to pharmacological modulation (10,11) focuses attention on the liver not only as a primary therapeutic target but as a potentially important contributor to the peripheral haemodynamic alterations. Several groups have demonstrated increased responsiveness to NA in the portal vascular bed of the cirrhotic liver (12-15), and it has been shown by one group (15,16) that NO plays a modulatory (vasodilatory) role in both the cirrhotic and the normal animal. While basal hepatic haemodynamics are probably not normally under significant sympathetic control (17), it is unknown whether there is increased responsiveness to perivascular sympathetic nervous activity in the portal bed of the cirrhotic liver. Vascular shear stress, which is likely to be increased in cirrhosis, may activate ATP-sensitive K ÷ channels in the arterial vasculature (18), and it has been shown that increased activation of K ÷ channels may indeed contribute to peripheral vasodilatation in cirrhosis (19). However, nothing is known about the role of K + channels in the portal vascular bed nor their impact in cirrhosis. The aim of the present study was therefore to investigate further the vasoconstrictive responsiveness of the portal vascular bed in cirrhosis, specifically to electrical field stimulation (EFS), NA, the selective C~l-adrenoceptor agonist methoxamine, and the cotransmitter adenosine 5"-triphosphate (ATP). The liver was isolated from rats made cirrhotic by administration of carbon tetrachloride (CC14)/phenobarbitone. We have examined the haemodynamic changes due to NO inhibition by NG-nitro-L-arginine methyl ester (L-NAME). The effects of L-NAME and the ATP-sensitive K + channel inhibitor glibenclamide on pressure changes with increased perfusate flow, and the effect of increased flow on constrictor responses to NA, were also determined.
Materials and Methods Experimental groups Cirrhosis was induced in seven male Wistar rats (140-160 g), by a modification of the method of Proctor & Chatamra (20) as previously described (21). Animals were allowed free access to drinking water containing 350 mg/1 phenobarbitone for 2 weeks before treatment with CC14 and thereafter. CCI 4 was administered by gavage weekly for 3-4 months, with a starting dose of 40 gl increasing to a maximum of 600 gl. Cirrhosis was judged to have
occurred when abdominal ascites was present as indicated by a sudden increase in body weight and by clinical inspection. A further five rats were given phenobarbitone alone, and a group of five agematched animals served as untreated controls. All procedures were carried out under the terms of a UK Home Office Animal Project Licence.
Isolated hepatic portal vascular bed The rats were killed by asphyxiation in C O 2. The portal vein was cannulated with a metal cannula (internal diameter: 1.0 mm), taking care to ensure placement of the tip just below the bifurcation and to secure the ligature above the gastroduodenal vein. The isolated liver was mounted on a stainless steel grid (7 cmx5 cm) in a humid chamber. The preparation was perfused at a constant rate of 5.0 ml/min with Krebs' solution containing (mM): NaC1 133, KC1 4.7, NaHzPO 4 1.35, NaHCO 3 16.3, MgSO 4 0.61, glucose 7.8 and CaC12 2.52, gassed with 95% O 2 and 5% CO 2, and maintained at 37°C. Responses were measured with a pressure transducer (model P23XL; Gould) on a side arm of the perfusion cannula, and recorded on a polygraph (model 79D; Grass); pressure attributable to the cannula alone (4 mmHg) was subtracted from all recordings to provide a measurement of vascular tone. Drugs were administered as 50 ~tl bolus injections. Preparations were allowed to equilibrate for approximately 30 min prior to experimentation. Experimental protocol EFS (90V, 1 ms pulse width, for 30 s) was applied at increasing frequencies (4-32 Hz) to obtain a frequency-response curve. Vasoconstrictor responses of preparations were then tested to increasing doses (50 ~1 bolus injections) of NA, methoxamine and ATE Individual doses were applied at intervals of 2-5 min, depending on the time it took for the tone to return to baseline. Approximately 10 min were allowed between consecutive dose-response curves. Flow/ pressure relationships of the preparations were then studied by incrementally increasing the perfusion rate from 5 to 10, 15 and finally 20 ml/min. At each flow rate, responses to a single dose of NA (5 nmol; approximately the EDs0 dose) were tested as soon as the preparation had stabilized at its new level of tone. Flow was returned to 5 ml/min and the preparation equilibrated with L-NAME (30 gM) for 30 min, after which response curves to EFS, NA, methoxamine and ATP were established, followed by a repeat of the step-wise increases in flow and injections of NA. The preparation was then equilibrated with L-arginine (300 gM) 91
R. T. Mathie et al.
ferences between groups was carried out by one-way analysis of variance; where evidence for a statistically significant effect was detected (p<0.05) a modified t-test (22) was then used to compare cirrhotic data with those of age-matched and phenobarbitonetreated controls. Within each group, the effects of LNAME and L-arginine on vasodilator responses (mmHg) were compared with equivalent data obtained in their absence by Student's paired t-test.
for 20 min (still in the presence of L-NAME) and dose-responses to EFS and NA were re-established. The NA dose-responses also served as an index of the viability of the preparation with time; maintained responsiveness to NA was a prerequisite for continuing each experiment. After washout of all drugs, the preparation was equilibrated with glibenclamide (5 ~tM) for 15 min, and the flow/pressure relationships and NA responses reassessed. Immediately after the end of the perfusion experiment, a segment from each of two liver lobes was placed in 4% paraformaldehyde for subsequent histological examination. The portal bed of two further normal rats was used to obtain frequency-response curves to EFS in the absence and then in the presence of the selective c~]adrenoceptor antagonist prazosin (10 -6 M).
Results Cirrhotic model
Of the seven rats treated with CC14, only six were used in the present study based on the development of fibrosis and regenerative nodules in the liver, as revealed by histological examination. All six rats had mild/moderate ascites on opening the abdomen (5-10 days after the original clinical diagnosis of ascites).
Drugs
Adenosine 5'-triphosphate (disodium salt), methoxamine (hydrochloride), noradrenaline (bitartrate), phenobarbitone (sodium salt), glibenclamide, prazosin, NG-nitro-L-arginine methyl ester hydrochloride and L-arginine were obtained from Sigma, Poole, England. All drugs were made up in distilled water, except for NA which was made up as a stock solution of 10 mM in 0.1 mM ascorbic acid and diluted in distilled water, and glibenclamide which was made up as a stock solution of 10 mM in DMSO.
Basal portal vascular tone
Basal tone did not differ significantly between the three experimental groups (controls: 0.4_+0.24 mmHg; phenobarbitone: 0.8_+0.48 mmHg; cirrhosis: 1.5+0.72 mmHg; see also Table 2). Electrical field stimulation
EFS (4-32 Hz) elicited frequency-dependent vasoconstrictor responses which were similar in magnitude in all groups (Fig. 1a). The addition of L-NAME Fig. lb) or L-arginine had no significant effect on the frequency-response curve in any group. In two nor-
Data analysis
All results are expressed as the mean_+one standard error of the mean (s.e.m). Statistical analysis of difE f f e c t o f E F S in c i r r h o s i s Basal
E f f e c t o f E F S in c i r r h o s i s L-NAME
6
6
:E ES
I ~S
v
~4
~4
o- 3
o- 3
~2
2
c_
f-
(D 0
4
8
12
16
20
Frequency (Hz)
24
28
32
36
0
b
4
8
12
16
20
24
28
32
36
Frequency (Hz)
Fig. 1. Frequency-dependent constrictor responses to electrical field stimulation (EFS; 4-32 Hz, 90V, 1 ms, for 30 s duration) of portal vascular bed preparations from age-matched control (-s-), phenobarbitone-treated (-u-) and cirrhotic (-u-) rats in (a) the absence and (b) the presence of NC-nitro-L-arginine methyl ester (L-NAME). "Basal" denotes responses in the absence of drugs in the perfusate. There were no significant differences between the groups in either (a) or (b). 92
Portal vascular responsiveness in cirrhosis Effect of n o r a d r e n a l i n e in cirrhosis Basal
Effect of n o r a d r e n a l i n e in cirrhosis L-NAME
14
14
!
Z
E12
12
g
10
~10 co co
Q_
Q_
© ~L
©
8
*
6
4
li
4 (1) C~
2
x~
2
o
© I
0 -10
I
h
~
I
-9
,
-8
,
I
I
i
-7
0 -10
-6
L
i
I
L
i
I
-9
NA (log mol)
I
:
I
I
-8
I
I
-7
~
I
-6
NA (log tool)
Fig. 2. Dose-response curves to noradrenaIine (NA) of portal vascular bed preparations from age-matched control (-d-), phenobarbitone-treated (-u-) and cirrhotic (-,-) rats in (a) the absence and (b) the presence of N~-nitro-L-arginine methyl ester (L-NAME). "Basal" denotes responses in the absence of drugs in the perfusate. Differences between cirrhotic and agematched or phenobarbitone-treated controls are denoted by * (p<0.05). Effect of ATP in cirrhosis Basal
Effect of ATP in cirrhosis L-NAME
12
12
~z El0 ~
8
~ 8
~
6
o_
co
©
~ c
6
©
4
Q- 4 c
~2 O3 X:Z
xz
©
© v
-10
-9
-8
-7
-6
ATP (log tool)
-5
o -10
.9
-8
-7
-6
-5
ATP (log mol)
Fig. 3. Dose-response curves to adenosine 5'-triphosphate (ATP) of portal vascular bed preparations from age-matched control (~t-), phenobarbitone-treated (-u-) and cirrhotic (-u-) rats in (a) the absence and (b) the presence of NC-nitro-L-ar ginine methyl ester (L-NAME). "Basal" denotes responses in the absence of drugs in the perfusate. There were no significant differences between the groups in either (a) or (b). In the age-matched group only, L-NAME caused a significant reduction in response to -5. 75 log tool and -5.25 log mol ATP (paired t-test; p
mal livers, prazosin (10-6M) completely abolished vasoconstrictor responses at all stimulation frequencies, showing that these were mediated by NA at post-junctional o~l-adrenoceptors, and therefore involved sympathetic nerves. Noradrenaline
Noradrenaline elicited dose-dependent vasoconstrictor responses in the portal bed, which were significantly greater in the preparations from cirrhotic rats
compared with either controls or phenobarbitonetreated animals at doses o f - 7 . 2 5 log tool (p<0.05), 26.75 log mol (p=0.01) and -6.25 log mol (p<0.01) (Fig. 2a). These responses were not affected by the addition of L-NAME to the perfusate, though the significant difference between cirrhotic and control groups was extended down to NA doses o f - 8 . 2 5 log mol (p<0.05) a n d - 7 . 7 5 log mol (p<0.05) (Fig. 2b). L-arginine had no significant effect on responses in any group. 93
R. T. Mathie et al.
Methoxamine Bolus injections of methoxamine elicited dosedependent vasoconstrictor responses that were similar between the groups and were unaffected by LNAME (Table 1). Adenosine 5"-triphosphate Injections of ATP elicited similar dose-dependent constrictions in cirrhotic, control and phenobarbitone-treated groups (Fig. 3a). L-NAME had no significant effect on responses to ATE except in the control group where an attenuation in response at the top
TABLE 1 Effect of bolus injections of methoxamine on portal vascular tone in three groups of rats under basal perfusion conditions and during infusion of L-NAME (30 ~M). Data are mean_+SEM Methoxamine (nmol)
Increase in perfusion pressure (mmHg) Basal
L-NAME
Controls
0.5 5 50 500 5000
0 0.6_+0.24 2.0-+0.35 4.1_+0.83 4.7+1.45
0 0 1.6_+0.24 3.1_+0.59 3.3_+0.93
Phenobarb.
0.5 5 50 500 5000
0 0.1-+0.13 1.6_+0.63 3.9-+0.43 3.7+0.33
0.1_+0.13 1.5_+0.79 2.1_+0.52 3.5+0.20 3.8_+0.25
Cirrhosis
0.5 5 50 500 5000
0.4_+0.33 0.6_+0.30 2.1_+0.52 4.3_+0.95 5.2_+1.33
0.5_+0.34 0.7_+0.31 1.6_+0.44 2.8_+0.46 4.3_+0.72
TABLE 2 Effect of flow rate on portal vascular tone in three groups of rats under basal conditions, and during infusion of L-NAME (30 p.M) or glibenclamide (5 laM). Data are mean_+SEM Flow rate
Perfusion pressure (mmHg)
(ml/min)
Basal
Controls
5 10 15 20
0.4_+0.24 2.4_+0.40 5.1_+0.68 6.2_+0.86
0.4_+0.24 2.0_+0.35 4.2_+0.64 5.5_+0.89
0.4_+0.24 2.1_+0.24 4.6_+0.68 5.8_+1.08
Phenobarb.
5 10 15 20
0.8_+0.48 3.5_+0.87 6.5_+1.19 9.0_+2.16
0.8+0.48 3.5_+1.19 8.3_+1.11 10.8_+1.80
0.8_+0.48 3.1_+0.72 6.3_+1.20 9.0~_2.13
Cirrhosis
5 I0 15 20
1.5_+0.72 3.7_+0.80 6.7_+1.28 8.3_+1.71
94
L-NAME
TABLE 3 Effect of bolus injection of noradrenaline (5 nmol) on portal vascular tone in three groups of rats under basal conditions and during infusion of L-NAME (30 I.tM) or glibenclarnide (5 gM). Data are mean_+SEM Flow rate
Increase in perfusion pressure (mmHg)
(ml/min)
Basal
L-NAME
Glibenclamide
Controls
5 10 15 20
4.3_+0.63 3.6_-+-0.55 3.0_-+0.41 2.4_+0.38
3.8+0.43 2.6+0.55 2.4_-!-0.47 2.4_+0.47
2.4_+0.24 2.0+0 1.8_+0.14 1.5_+0.20
Phenobarb.
5 10 15 20
3.3+0.25 3.3_+0.60 2.~_-0.35 1.8+0.43
4.3_+0.25 3.2_+0.17 2.3_+0.48 2.1_+0.31
2.8_+0.32 2.4_+0.24 1.9_+0.13 1.9_+0.13
Cirrhosis
5 10 15 20
5.4_+0.80 3.9_+0.40 3.6_+0.48 3.2_+0.72
6.2_+1.07 4.7_+0.73 4.1 +0.64 3.5_+0.79
5.3_+0.66 4.0_+0.57 3.5_+0.69 3.1_+1.03
two doses of ATP was noted (paired t-test; p<0.05) (Fig. 3b).
Flow rate Stepwise increases in flow rate (up to 20 ml/min) produce an increase in perfusion pressure that was not statistically significantly different between the groups, and was independent of the type of agent in the perfusate (Table 2). Noradrenaline with flow There was a progressive decrease in responsiveness to NA with increased flow in all groups for each of the three phases of the experiment (basal, L-NAME and glibenclamide) (Table 3). Responses to NA appeared to be greater in the cirrhotic group at all flow rates under basal conditions, but this did not reach statistical significance. There was no significant effect of L-NAME on responses to NA. In the presence of glibenclamide, responses of both the control groups were significantly reduced compared to those in the cirrhotic group, which were maintained.
Glibenclamide
1.5_+0.72 1.5_+0.72 4 . 8 _ + 1 . 1 5 4.9-+1.08 7 . 6 - + 1 . 8 4 8.8_+1.52 9.4_+2.19 11.4_+2.39
Discussion Our results show that, compared to controls, at a portal flow rate of 5 ml/min, the portal vascular bed of cirrhotic rats is significantly more responsive to NA, whereas responsiveness to EFS, methoxamine and ATP is unchanged. Inhibition of NO synthesis did not augment these responses, in contrast to the potentiation observed in many other vessels and preparations, including the mesenteric bed of normal, cirrhotic and portal hypertensive rats (23,24). Enhancement of responsiveness to NA in the por-
Portal vascular responsiveness in cirrhosis
tal bed of the cirrhotic rat liver has been noted previously (12-15). Explanations for this effect include: an increase in the size and contractility of the intramural smooth muscle due to the sustained rise in intraluminal pressure associated with cirrhosis; the presence of activated myofibroblast-like (Ito) cells surrounding thin-walled venous channels and sinusoids in the cirrhotic liver (14). The lack of potentiation of responses to the c~l-adrenoceptor agonist methoxamine in the present study suggests that another possible explanation may be selective changes in adrenoceptor subtypes in cirrhosis. This hyper-reactivity to NA in the cirrhotic portal bed is in contrast to the hyporesponsiveness to vasoconstrictor agents known to occur in other regions of the peripheral vasculature in cirrhosis (8,9). Our results suggest that there is no increase in hepatic sympathetic nervous function in experimental cirrhosis. While this conclusion is consistent with that of Abergel et al. (25), it is surprising in view of the enhancement in post-junctional responses to NA (though not those to methoxamine) in our cirrhotic experiments. Only a very modest increase in responsiveness to sympathetic nerve stimulation in cirrhotic liver was noted by Zimmermann et al. (13). One explanation for the discrepancy between nerve stimulation and NA in cirrhosis may be the responsiveness of the activated Ito cells, whose contractile apparatus may react to intraluminally applied NA but not to neural stimulation. Our results with the oq-adrenoceptor antagonist prazosin also appear to rule out the possibility of an agent in addition to NA acting as a neurotransmitter in the portal vascular bed of the normal rat liver. However, cotransmission of a modulatory agent with NA from sympathetic nerves in cirrhosis cannot be excluded. Cotransmission of ATP with NA is a recognised feature of sympathetic nerve activity that has been noted in both the isolated portal vein (26) and hepatic artery (27). The implication of our results is that the raised portal vascular resistance of cirrhosis in vivo may be partly attributed to the elevated plasma levels of NA, and not to increased hepatic portal sympathetic nervous activity per se. Induction by endotoxin of NO synthase in vascular smooth muscle may contribute to the peripheral vasodilatation of decompensated liver disease (28). Much of the evidence for increased NO synthase induction has been indirect: inhibition of NO synthesis partly reversed the hyporeactivity to vasoconstriction produced by methoxamine (29,30), and reversed peripheral vasodilatation in cirrhotic rats (31,32). These findings, however, have been both corroborated and unsupported by clinical investigations (33-
35), and some of the more recent experimental work has been unable to confirm a role for NO in portal hypertension (36,37). There is now increasing evidence that changes in K ÷ or Ca 2+ channels may contribute importantly to the peripheral vasodilatation of cirrhosis (19,38). In this study we examined evidence for changes in the role of NO in the portal bed of cirrhotic rats. Our results with NA differ from those reported by Mittal et al. (15,16), in that these authors found evidence for basal release of NO which modulated the NA response similarly in cirrhotic and control portal bed preparations. We found no evidence of basal NO release in the portal bed of either control or cirrhotic animals since L-NAME (30 gM) produced no augmentation of responses to NA or nerve stimulation. It is unlikely that higher concentrations of L-NAME would alter the results, since there was not even a trend towards enhancement of responses to NA or nerve stimulation in the presence of L-NAME. Furthermore, it has been demonstrated that the IC50 for L-NAME is 2.5 gM against constitutive NO synthase and 20-25 ~tM against inducible NO synthase (39), confirming that a concentration of 30 gM will produce a very substantial inhibition of NO synthesis via either isoform. The discrepancies may, however, relate to the supranormal portal flow rate used by Mittal et al. (40 ml/min), which would have induced a greater shear stress (and thus NO release) than in the current study. Although at the lower end of the normal range in vivo, the basal flow rate of 5 ml/min used in our in vitro experiments resulted in portal pressure recordings similar to those achieved in other perfused rat liver preparations at higher flow rates (40). Our own data showed only a modest increase in portal pressure at increased flow rates, findings also consistent with the highly distensible nature of the portal vascular bed (41). That our preparation was not compromised pharmacologically at 5 ml/min is indicated by the optimal vasoconstriction to bolus injections of NA observed at that flow rate: at higher flow rates there was a progressive decrease in responses (vide infra). L-NAME had no effect on perfusion pressure even at a flow rate of 20 ml/min, suggesting that NO release was independent of flow in our particular experimental system. We do not know whether shear-evoked NO release would be apparent at still higher flows such as the 40 ml/min used by Mittal et al. The important point of agreement between our studies and those of Mittal et al., however, is that portal NO release is not greater in cirrhotic compared with normal liver. This indicates, therefore, that the portal bed 95
R. T. Mathie et al.
is unlikely to participate in any NO-mediated peripheral vasodilatation. In contrast to the augmentation of ATP-mediated vasoconstriction by NO inhibition in the perfused rabbit liver (42), constrictor responses to ATP in the present study were not significantly increased by LNAME in either normal or cirrhotic livers. The magnitude of basal responses to ATP in both series was similar, as was the concentration of L-NAME used to inhibit NO synthesis. Other than species differences and that the hepatic artery was not perfused in the current experiments, therefore, an explanation for the different findings in the two experiments is not obvious. The effect of ATP in the portal bed of cirrhotic animals has not previously been examined. Assessment of the perfusion pressure at different flow rates in our study revealed a similar increase in pressure with step-wise increases in flow between the groups, in contrast to the steeper slope describing the pressure-flow relationship in cirrhosis noted by Zimm o n & Kessler (43). The perfusion pressure at different flow rates was not modulated by L-NAME or glibenclamide. Examination of the constrictor responses of preparations to a single, approximately EDs0, dose of NA at different flow rates revealed that responses were progressively smaller at higher flow rates in all groups. It appears, however, that increased pressure/flow or inhibition of NO at different flow rates has little effect on the hyper-responsiveness to NA in cirrhosis compared with controls. The greater constrictor response to NA of cirrhotic preparations compared with controls in the presence of glibenclamide may be a consequence of activation of K + channels in the portal venous bed in cirrhosis via an unknown mechanism; evidence for the activation of K ÷ channels in the peripheral arterial circulation of cirrhotic rats has been presented previously (19). On the other hand, the maintained responses to NA in the presence of glibenclamide in cirrhotic preparations may simply be a reflection of a sustained hyperresponsiveness to NA in this group. In general, there was good viability of the preparations over the course of the experiment, as evidenced by maintained responsiveness to NA with time, and there is no reason to suppose that the long-term viability of the control preparations was less than that of the cirrhotics. We conclude that portal vasoconstriction associated with NA, but not sympathetic nerve stimulation, methoxamine or ATE is enhanced in experimental cirrhosis. NO does not appear to play a modulatory role. Whether these changes are representative of circumstances in cirrhosis in humans remains to be determined. 96
Acknowledgements The support of the British Heart Foundation and the Hammersmith & Queen Charlotte's Special Health Authority is gratefully acknowledged. We are grateful to Mrs M. Millbourn and Mr J. Godwin, Biological Services Unit, Royal Postgraduate Medical School, for technical help.
References 1. Bosch J, Pizcueta P, Feu E Fernfindez M, Garcfa-Pag~n JC. Pathophysiology of portal hypertension. Gastroenterol Clin North Am 1992; 21: 1-14. 2. Ring-Larsen H, Hesse B, Henriksen JH, Christensen HJ. Sympathetic nervous activity and renal and systemic hemodynamics in cirrhosis: plasma norepinephrine concentration, hepatic extraction, and renal release. Hepatology 1982; 2: 304-10. 3. Bichet DG, Van Putten VJ, Schrier RW. Potential role of increased sympathetic activity in impaired sodium and water excretion in cirrhosis. New Engl J Med 1982; 307: 1552-7. 4. Esler M, Dudley F, Jennings G, Debinski H, Lambert G, Jones P, Crotty B, Colman J, Willett I. Increased sympathetic nervous activity and the effects of its inhibition with clonidine in alcoholic cirrhosis. Ann Intern Med 1992; 116: 44655. 5. Floras JS, Legault L, Morali GA, Hara K, Blendis LM. Increased sympathetic outflow in cirrhosis and ascites: direct evidence from intraneural recordings. Ann Intern Med 1991; 114: 373-80. 6. Groszmann RJ. Hyperdynamic circulation of liver disease 40 years later: pathophysiology and clinical consequences. Hepatology 1994; 20: 1359-63. 7. Stark ME, Szurszewski JH. Role of nitric oxide in gastrointestinal and hepatic function and disease. Gastroenterology 1992; 103: 1928-49. 8. Ryan J, Sudhir K, Jennings G, Esler M, Dudley E Impaired reactivity of the peripheral vasculature to pressor agents in alcoholic cirrhosis. Gastroenterology 1993; 105:1167-72. 9. Bomzon A, Blendis LM. The nitric oxide hypothesis and the hyperdynamic circulation in cirrhosis. Hepatology 1994; 20: 1343-50. 10. Bhathal PS, Grossman HJ. Reduction of the increased portal vascular resistance of the isolated perfused cirrhotic rat liver by vasodilators. J Hepatol 1985; 1: 325-37. 11. Reichen J. Liver function and pharmacological considerations in pathogenesis and treatment of portal hypertension. Hepatology 1990; 11: 1066-78. 12. Ballet F, Chretien Y, Rey C, Poupon R. Differential response of normal and cirrhotic liver to vasoactive agents. A study in the isolated perfused rat liver. J Pharmacol Exp Ther 1988; 244: 283-9. 13. Zimmermann T, Gardemann A, Machnik G, Dargel R, Jungermann K. Metabolic and hemodynamic responses of bivascularly perfused rat liver to nerve stimulation, noradrenaline, acetylcholine and glucagon in thioacetamide-induced micronodular cirrhosis. Hepatology 1992; 15: 464-70. 14. Grossman HJ, Grossman VL, Bhathal PS. Enhanced vasoconstrictor response of the isolated perfused cirrhotic rat liver to humoral vasoconstrictor substances found in portal venous blood. J Gastroenterol Hepatol 1992; 7: 283-7.
Portal vascular responsiveness in cirrhosis
15. Mittal MK, Lee F-Y, Groszmann RJ. Role of nitric oxide in the intrahepatic circulation of cirrhotic rats. Gastroenterology 1993; 104: A956. 16. Mittal MK, Gupta TK, Lee F-Y, Sieber CC, Groszmann RJ. Nitric oxide modulates hepatic vascular tone in normal rat liver. Am J Physiol 1994; 267: G416-22. 17. Wheatley AM, Stuart ET, Zhao D, Zimmermann A, Gassel H-J, Blumgart LH. Effect of orthotopic transplantation and chemical denervation of the liver on hepatic hemodynamics in the rat. J Hepatol 1993; 19: 442-50. 18. Hassdssian H, Bodin P, Burnstock G. Blockade by glibenclamide of the flow-evoked endothelial release of ATP that contributes to vasodilatation in the pulmonary vascular bed of the rat. Br J Pharmacol 1993; 109: 466-72. 19. Moreau R, Komeichi H, Kirstetter P, Ohsuga M, Cailmail S, Lebrec D. Altered control of vascular tone by adenosine triphosphate-sensitive potassium channels in rats with cirrhosis. Gastroenterology 1994; 106: 1016-23. 20. Proctor E, Chatamra K. High yield micronodular cirrhosis in the rat. Gastroenterology 1982; 83:1183-90. 21. Firth JD, Gove C, Panos MZ, Raine AEG, Williams R, Ledingham JGG. Sodium handling in the isolated perfused kidney of the cirrhotic rat. Clin Sci 1989; 77: 657-61. 22. Altman DG. In: Altman DG. Practical Statistics for Medical Research. London: Chapman & Hall, 1991: 210-2. 23. Ralevic V, Mathie RT, Moore KP, Burnstock G. Vasoconstrictor responsiveness of the rat mesenteric arterial bed in cirrhosis. Br J Pharmacol 1996; in press. 24. Sieber CC, Groszmann RJ. Nitric oxide mediates hyporeactivity to vasopressors in mesenteric vessels of portal hypertensive rats. Gastroenterology 1992; 103: 235-9. 25. Abergel A, Braillon A, Gaudin C, Kleber G, Lebrec D. Persistence of a hyperdynamic circulation in cirrhotic rats following removal of the sympathetic nervous system. Gastroenterology 1992; 102: 656-60. 26. Burnstock G, Crowe R, Kennedy C, T6rrk J. Indirect evidence that purinergic modulation of perivascular adrenergic neurotransmission in the portal vein is a physiological process. Br J Pharmacol 1984; 82: 359-68. 27. Brizzolara AL, Burnstock G. Evidence for noradrenergicpurinergic cotransmission in the hepatic artery of the rabbit. Br J Pharmacol 1990; 99: 835-9. 28. Whittle BJR, Moncada S. Nitric oxide: The elusive mediator of the hyperdynamic circulation of cirrhosis? Hepatology 1992; 16: 1089-92. 29. Sieber CC, Groszmann RJ. In vitro hyporeactivity to methoxamine in portal hypertensive rats: reversal by nitric oxide blockade. Am J Physiol 1992; 262: G996-1001. 30. Lee F-Y, Albillos A, Colombato LA, Groszmann RJ. The role of nitric oxide in the vascular hyporesponsiveness to methoxamine in portal hypertensive rats. Hepatology 1992; 16: 1043--48.
31. Pizcueta E Piqu6 JM, Fern~indez M, Bosch J, Rodrs J, Whittle BJR, Moncada S. Modulation of the hyperdynamic circulation of cirrhotic rats by nitric oxide inhibition. Gastroenterology 1992; 103: 1909-15. 32. C15xia J, Jimrnez W, Ros J, Asbert M, Castro A, Arroyo V, Rivera F, Rodrs J. Pathogenesis of arterial hypotension in cirrhotic rats with ascites: role of endogenous nitric oxide. Hepatology 1992; 15: 343-9. 33. Guarner C, Soriano G, Tomas A, Bulbena O, Novella MT, Balanzo J, Villardell F, Mourelle M, Moncada S. Increased serum nitrite and nitrate levels in patients with cirrhosis: relationship to endotoxemia. Hepatology 1993; 18: 1139-43. 34. Hori N, Takahashi H, Okanoue T, Mori T, Sakamoto S, Nanbu A, Yoshimura M, Kashima K. Nitric oxide production in patients with chronic liver diseases. Hepatology 1993; 18: 101A. 35. Le Moine O, Soupison T, Sogni P, Husson C, Cruziaux A, Marchant A, Moreau R, Hadengue A, Goldman M, Deviate J, Lebrec D. The role of endotoxins and TNFc~ in the hyperkinetic circulation of cirrhosis. J Hepatol 1993; 18 (Suppl 1): $40. 36. Fern~indez M, Garcfa-Pag~in JC, Casadevall M, Bernardich C, Piera C, Whittle BJR, Piqu6 JM, Bosch J, Rodrs J. Evidence against a role for inducible nitric oxide synthase in the hyperdynamic circulation of portal-hypertensive rats. Gastroenterology 1995; 108: 1487-95. 37. Soubrane O, Lacronique V, Miquerol L, Mignon A, Houssin D. The role of inducible NO synthase (iNOS) in the hyperdynamic circulation of portal hypertension. Hepatology 1995; 22: 156A. 38. Cailmail S, Moreau R, Gadano A, Tazi KA, Lebrec D. The role of calcium channel inhibition in in vitro vascular hyporeactivity to endothelin-1 (ET-1) in rats with cirrhosis. Hepatology 1995; 22: 258A. 39. Mitchell JA, Kohlhaas KL, Sorrentino R, Warner TD, Murad F, Vane JR. Induction by endotoxin of nitric oxide synthase in the rat mesentery: lack of effect on action of vasoconstrictors. Br J Pharmacol 1993; 109: 265-70. 40. Grossman HJ, Grossman VL, Bhathal PS. Hemodynamic characteristics of the intrahepatic portal vascular bed over an extended flow range: a study in the isolated perfused rat liver. Hepatology 1995; 21: 162-8. 41. Lautt WW, Legare DJ. Passive autoregulation of portal venous pressure: distensible hepatic resistance. Am J Physiol 1992; 263: G702-8. 42. Browse DJ, Mathie RT, Benjamin IS, Alexander B. The transhepatic action of ATP on the hepatic arterial and portal venous vascular beds of the rabbit: the role of nitric oxide. Br J Pharmacol 1994; 113: 987-93. 43. Zimmon DS, Kessler RE. Effect of portal venous blood flow diversion on portal pressure. J Clin Invest 1980; 65: 138897.
97