0270-9139/82/0202-0236$02.00/0 HEPATOLOGY Copyright 0 1982 by the American Association for the Study of Liver Diseases

Vol. 2, No. 2, p. 236, 1982 Printed in U.S.A.

The Relationship Between Prostacyclin Activity and Pressure in the Portal Vein

GEORGE HAMILTON, ROSE CHOW FUNG PHING, RONALD A. HUTTON, PARESH DANDONA, AND KENNETH E. F. HOBBS

Departments of Academic Surgery, Chemical Pathology, and the Katherine Dormandy Haemophilia Centre, Royal Free Hospital School of Medicine, Pond Street, London NW3

ZQG, England

The effect of portal hypertension (induced by partial ligation) on the ability of the portal vein wall to produce prostacyclin (PG12) was studied in rats over a period of 6 weeks. PG12-like activity measured by bioassay was shown to be significantly increased in portal vein segments during established hypertension when compared to control groups. As a collateral circulation developed with consequent fall in portal venous pressure, PGIz -like activity decreased. A positive correlation between PGL-like activity and portal pressure has been demonstrated. Parallel measurements of this activity by bioassay and radioimmunoassay to 6-keto-prostaglandin-F1 a showed a positive correlation (r = 0.88) and suggested that PGIz is the inhibitory prostaglandin involved. The possible significance of these observations in relation to the hemorrhagic manifestations associated with portal hypertension in man is discussed.

Prostacyclin (PG12) is a potent vasodilator and inhib- itor of platelet aggregation. It is produced by vascular endothelium of many species including the rat and man (1,2). As well as being an important mediator of platelet- vessel wall interaction, it may be a circulating hormone, one of the functions of which is to counteract the proag- gregatory property of thromboxane, a prostaglandin de- rivative generated endogenously in stimulated platelets (3) (Figure 1).

The demonstration of alteration in PGI:! stimulating or inhibiting activity of plasma from patients with a number of diseases associated with hemorrhage or thrombotic complications suggests the possibility that PGIz has some clinical relevance (4). Van Hoof et al. have reported groups of patients with liver or renal disease having raised levels of PGIz stimulating factors in their plasma (5). Pace-Asciak et al. found increased PGIz activity in aortas of spontaneously hypertensive rats, and suggested that this might be an adaptive response to hypertension (6). We have demonstrated significantly greater PGI:! activity in rat portal vein wall 1 week after partial ligation of the vein sufficient to induce hypertension (7).

These observations led us to postulate that increased production of PGIz by portal vein wall in response to ~

Received October 2, 1980; accepted November 28, 1981. Dr. Hamilton was supported in this study by a grant from the

Wellcome Trust. Dr. Phing was supported by a grant from the North East Thames

Regional Health Authority. Address reprint requests to: G. Hamilton, Academic Department of

Surgery, Royal Free Hospital School of Medicine, Pond Street, London NW3 2QG, England.

increased intraportal pressure might exacerbate any un- derlying hemorrhagic tendency related, for example, to esophageal varices in patients with portal hypertension. As a step in the investigation of this hypothesis, antiag- gregatory activity produced by the portal vein wall was measured in rats at various times after induction of portal hypertension, and this activity correlated with levels of 6-keto-PGFla, the stable breakdown product of PGIz, as measured by radioimmunoassay.

MATERIALS AND METHODS INDUCTION OF PORTAL HYPERTENSION; COLLECTION OF

SPECIMENS Prehepatic portal hypertension was produced in six

groups of rats (A to F), each group consisting of six animals. The animals were anesthetized with ether, and the peritoneal cavity opened through an upper midline incision. Pressure in the portal vein was measured in centimeters of water by simple manometry, using the portal vein at the porta hepatis as the zero reference point. The portal vein was mobilized with care. A teflon tube or metal needle 1.2 mm in diameter was placed alongside the vessel and a polyamide ligature tied firmly around both as close to the porta hepatis as possible. The tube or needle was then removed leaving the vein con- stricted to a standard diameter. The abdomen was closed, and the animal was allowed to recover.

Laparotomy, portal vein manometry, and excision of 2 to 3 cm of portal vein splenad (i.e., hypertensive segment) to the ligature were carried out in each group at 24 hr (Group A), 4 days (Group B), 1 week (Group C), 2 weeks

236

Vol. 2, No. 2, 1982 PGIi ACTIVITY AND PRESSURE I N THE POHTAL VEIN 237

ARACHIDONIC A C I D

t’ CYCLIC ENDOPEROX 1 DES

(PGG2 PGH+

I \

6 - keto PGFlH

FIG. 1. Metabolism and structures of PGZz and thromboxane (TX.4, ), and their stable deriuatiues. The cyclic endoperoxides PGGa and PGH2 art: produced from arachidonic acid, a component of cell memhrane phospholipids, by the action of cyclooxygenase (1). In vas- cular endothelium, PGG2 and PGH, are mainly metabolized by pros- tacyclin synthetase (2) to PGIL, an unstable compound which rapidly breaks down nonenzymatically to its stable derivative, 6-keto-PGFI a. In platelets, however, thromboxane synthetase (3) metabolizes cyclic endoperoxides to thromboxane, also a very labile compound, which rapidly breaks down to its stable derivative TXB2. In other tissues, 1’GG.L and PCHL are mainly metabolized to PGL),, PGE,, and PGF2 which are stable.

(Group D), 4 weeks (Group E), and 6 weeks (Group F). Control groups consisting of equal numbers of animals had laparotomy without manipulation of the portal vein, and then were reanesthetized at the same time intervals as Groups A to F and portal veins removed after pressure measurement.

In two additional groups of 6 animals each, 6 under- went partial ligation of the portal vein as described above (Group G) and 6 had laparotomy with portal vein dissec- tion alone (Group H). These animals were reanesthetized at 1 week and portal veins removed after pressure mea- surement.

Tail vein blood samples were taken under ether anes- thesia from control rats following laparotomy alone (n = 6), from rats having portal vein ligation prior to first laparotomy (n = 6), and at 1 (n = 6), 3 (n = 6), 5 (n = 6), and 7 days (n = 6) thereafter. Serum aspartate amino- transferast: (SGOT) and alanine aminotransferase (SGPT) levels were measured in these samples using an “optimized standard method” U-V system kit [Boehrin- ger Mannheim GmbH Diagnostica (Catalog numbers 124362 and 124524), Indianapolis Ind.].

BIOASSAY FOR PGL-LIKE ACTIVITY PGIZ-like activity from portal veins was measured as

described by Hutton et al. (8). Because PGIz is labile at room temperature, the dissected venous tissue was im-

mediately washed in Tris buffer at 4°C (0.05 M pH 8.0), cleaned of connective tissue, chopped into rings of 1-mm length, dried, and weighed. These rings were then resus- pended in fresh Tris saline at a temperature of 4°C and a concentration of 50 mg weight per ml of buffer. After 3 min incubation at room temperature, the supernatant fluid was removed from the tissue and stored at a tem- perature of 4°C until assayed a few minutes later.

The inhibitory effect of the supernatant fluid on ADP- induced platelet aggregation was measured. A 0.4-ml platelet-rich plasma (300 to 400 X lo9 per liter) was incubated in a Payton aggregometer at 37°C with 0.05 ml of tissue extract (or Tris buffer as control) for 60 sec, at which time 0.05 ml of ADP (100/pmoles per liter) was added, and the fall in optical density was determined 3 min after the addition of ADP. This was repeated with PG12 as standard, and the results were expressed as nanograms of PGIz-like activity per 50 mg of tissue. The same supernatants were also incubated at 37°C for 20 min (to destroy any PGL) and their PGL-like activity reassessed.

RADIOIMMUNOASSAY FOR 6-KETO-PGFl Samples from Groups G and H were similarly prepared

and analyzed. In addition to bioassay, 6-keto-PGF1 a, the stable breakdown product of PGL, was measured in the supernatants by a radioimmunoassay technique de- scribed by Salmon (9) and the results expressed as nan- ograms of 6-keto-PGF1a per 50 mg tissue. The 6-keto- PGF, a antiserum was supplied by Dr. J. A. Salmon, The Wellcome Research Laboratories, Beckenham, Kent, U.K.: tritiated (‘H) 6-keto-PGFI a was obtained from New England Nuclear Ltd., Southampton, U.K.

ASSAY FOR ADPASE ADPase activity in the tissue extracts was determined

as described by Hutton et al. (8). Each extract was incubated with an equal volume of ADP (100 pmoles per liter), and the residual ADP was assayed at intervals by the firefly luminescence method (10) and by measuring platelet aggregation induced by the residual nucleotide.

SPLENIC VENOGRAPHY Splenic venograms were carried out on five rats follow-

ing partial portal vein ligation at intervals of 1, 2, 3, 4, and 6 weeks in order to observe the development of collateral circulation. Five milliliters of 45% Hypaque contrast medium was injected into the splenic pulp under anesthesia with intraperitoneal urethane (0.6 ml per 100 gm body weight of 25% solution of urethane) and radio- graphs taken immediately. These rats were not used for PGIz determinations.

RESULTS There were no areas of gut or liver necrosis or conges-

tion noted in any of the animals. SGOT levels were raised on the first day in both ligated and control animals but there was no significant difference between the two groups. SGPT levels remained normal in both groups (Figure 2).

Portal venous pressures immediately before removal

238 HAMILTON ET AL. HEPATOLOGY

10

20

cm. H$

10

I

- = P. V. Ligation - = Laparotomy alone

P Meant s. E. M. (n.61

FIG. 2. Serial hepatic enzyme levels in rat tail vein blood samples after partial portal vein ligation or laparotomy alone. SGOT levels were moderately raised in both groups after 24 hr but thereafter returned to normal; there was no change in SGPT levels in both groups.

SGOT

T I

SGPT

I I I

1 3 5

TIM (Days)

I

1

f ---)-- Laparotomy alone I I II it

P v Ligation meant9.E. Nn-6).

I I I 1 I I 1 2 3 4 5 6

TIME (weeks)

FIG. 3. Portal vein pressureprior to removal ofportal vein for assay (portal vein pressure 60 see after ligation is shown in the figure at zero time). Pressure was significantly increased from 24 hr to 2 weeks after partial portal vein ligation (p < 0.01). Thereafter, pressure had stabilized aL levels approaching those of the control groups; throughout the experiment there was no significant change in portal vein pressure in the control groups trom normal levels. Portal vein pressures were measured prior to ligation and fell within the normal range [normal portal vein pressure = 9.2 rt 1.3 (S.D.) cm HzO].

Vol. 2, No. 2, 1982 PGI, ACTIVITY AND PRESSURE IN THE PORTAL VEIN

3 0 -

2 0 -

ng 15omg

10-

239

f P. v Ligation. mean9.E. M.ln-6)

--)--~igarotomyaione I , ,,

------- T

I I I 8 < ?r 1 2 3 4 5 6

TINE (woks) FIG. 4. Portal vein PGL activity from 24 hr to 6 weeks after partial portal vein ligation or laparotomy alone. PGI:! activity is significantly

increased in the ligated groups from Day 1 to Week 2 when compared to the control laparotomized groups (p < 0.01). Thereafter, PGIz activity approaches levels seen in the control groups. PGIl activity in the control groups is significantly greater than that found in anesthesia alone [mean = 4.3 f 2.6 (S.D.)] (p < 0.01).

of portal veins are shown in Figure 3. In the control groups, no significant changes were noted from normal portal vein pressure. There was a highly significant in- crease in portal venous pressure between 24 hr and 1 week after ligation (p < 0.01). By the second week, the pressure had fallen to a mean of 15 cm of water and by the fourth week, it had stabilized at 11 to 12 cm of water.

PGIz-like activity measured in the portal vein of ani- mals following either laparotomy or induction of portal hypertension at 1, 4, and 7 days, and 2, 4, and 6 weeks are shown in Figure 4. These results show that in animals with portal hypertension, PGIz -like activity is increased above control values after 24 hr, reaches a maximum between 4 and 7 days, begins to diminish after 2 weeks, and at 6 weeks approaches control levels. The difference reaches statistical significance only during the period between 24 hr and 2 weeks after the initial procedure (p < 0.01).

The fall in the level of PGL-like activity occurs at the same time as the fall in portal vein pressure. This rela- tionship is demonstrated in Figure 5. This shows a posi- tive correlation between PGIz activity in the portal vein and portal pressure between Days 4 and 7 (correlation coefficient = 0.76).

PGIz-like activity measured by bioassay and 6-keto- PGFl a measured by radioimmunoassay in the same sam- ples taken from Groups G and H are shown in Figure 6. These results show increased production of 6-keto- PGFla as well as increased PGI2-like activity in the animals with portal hypertension. There is a strong cor- relation between PG12 -like activity measured by bioassay

4 0 1

Oe/ 0 :

/ 0 , I I I 1

0 10 20 30 40

P. V. pressure (cm H20) FIG. 5. Relationship between portal vein pressure and PGh activity.

There is a strongly positive correlation between pressure and PGI:! activity in the portal vein (r = +0.76) (values at 4 and 7 days after procedure).

and 6-keto-PGF1 a measured by radioimmunoassay (cor- relation coefficient = 0.88; Figure 7).

ADPase activity could not be detected in m y of the extracts tested.

HAMILTON ET AL. HEPATOLOGY

measured by RIA (n-6) I + SD)

0 measured by Bioassay (11-61 (+ SD)

h

Group H Group G FIG. 6. Portal vein PGI, activity measured by bioassay and radioim-

munoassay (RIA) 1 week after partial portal vein ligation or laparotomy alone. In both control (H) and portal hypertensive (G) groups, 6-keto-YGF, a levels were approximately 60% of the antiaggre- gatory activity measured by bioassay and expressed as PGIJike activ- ity.

l5 1

r = a a

0 5 10 15 PGIdike activity ng/50mg

FIG. 7. Relationship between PGIz -like activity measured by bioas- say and 6-keto-PGFl a levels in portal vein supernatant fluids. There is a strongly positive correlation between the antiaggregatory activity measured by bioassay, and 6-keto-PGF, a levels in supernatant fluids obtained after incubation of portal vein segments from both control and ligated animals (r = 0.88).

There was no antiaggregatory activity in supernatants which had been incubated at 37°C for 20 min.

Splenic venography demonstrated retroperitoneal and splenorenal portasystemic collateral formation 2 weeks

after partial portal vein ligation, which was well estab- lished by 4 weeks (Figure 8).

DISCUSSION These results indicate that sustained partial ligation of

the portal vein sufficient to induce portal hypertension is associated with a significant increase in local production of antiaggregatory activity when compared to sham-op- erated animals. The production of inhibitory prosta- glandins other than PGIz by vascular tissue is low (1, 31, and no ADP degrading activity was detected in any of the samples tested. Moreover, the antiaggregatory activ- ity was, like that of pure PGIz, destroyed on incubation of the extracts at 37°C for 20 min, and we therefore believe that the major inhibitory substance generated by the vessels was PG12. This has been corroborated by the results of measurements using both bioassay and ra- dioimmunoassay on samples from Groups G and H.

Previous studies have shown that part of this response may be a nonspecific effect of surgical manipulation unrelated to hypertension (7). In the present study, how- ever, during the first 2 weeks of observation, PGL-like activity remained significantly greater in portal hyper- tensive animals than in those subjected to surgical ma- nipulation alone, and this trend persisted until portal vein pressure in these animals returned to control levels. Furthermore, a strong correlation between portal vein pressure and PGIz-like activity has been established (r = 0,76). These observations suggest that this is a specific response to increased pressure and not to surgery alone.

This demonstration of a direct relationship between PGL-like activity and pressure in the portal vein, and reports of similar increases in PGIz from aortas of rats with spontaneous (6) and experimentally induced hyper- tension (11) suggest that increased PGIz activity may be a normal adaptive respone to increased pressure by vas- cular endothelium.

We have shown PGIz-like activity and pressure in the portal vein falling simultaneously with the development of portasystemic collateral circulation as confirmed by splenic venography. Asa result of its potent vasodilator effect, this increased production of PGIz may promote formation of a collateral circulation, thus aiding return to normal pressure.

I t has been suggested that PGIz may have a role as a circulating hormone (12), but measurement of circulating levels in the portal system was not made in this study. However, from this data, it is probable that PG12 exerts its effect locally on the vessel wall throughout a hyper- tensive vascular system. This requires further study both in the experimental animal and in man.

If PGIz is persistently elevated in portal hypertension in man, its potent vasodilator properties may predispose to the development of the clinically important potasys- temic collaterals and its antiaggregatory activity may locally impair the hemostatic mechanism. Together these may contribute to the excessive bleeding from esophageal varices which commonly complicates this disease.

Acknowledgments: We are indebted to Upjohn Lim- ited, Kalamazoo, Michigan for the gift of pure PG12 and

Vol. 2, No. 2,1982 PGIz ACTIVITY AND PRESSURE IN THE PORTAL VEIN 241

FIG. 8. Portasystemic collateral formation demonstrated by splenic venography. Splenic venograms (using 458 Hypaque contrast medium) in rats; A, normal, B, 1 week after partial ligation of portal vein; C, 2 weeks after partial ligation of portal vein (the site of ligation of the portal vein is shown by an arrow; D, 4 weeks after partial ligation. There is minimal collateral formation at 1 week, but by the second week, large splenorenal, retroperitoneal, and periesophageal collaterals have opened up. At 4 weeks after ligation, the collateral circula!ion b well established with a large splenorenal collateral draining into the left renal vein.

242 HAMILTON ET AL. HEPATOLOGY

to the Wellcome Research Laboratories Buckingham, Kent, U.K. for the gift of 6-keto-PGFla antiserum.

REFERENCES

1. Moncada S, Higgs EA, Vane JR. Human arterial and venous tissues generate prostacyclin (prostaglandin X) a potent inhibitor of plate- let aggregation. Lancet 1977; 1:18-20. Moncada S, Vane JR, Whittle BJR. Relative Dotencv of Drostacv- ~- din, prostaglandin E and D2 as inhibitors of piatelet aggregation in several species. J Physiol (Lond) 19774 273:28-48. Moncada S, Vane JR. Unstable metabolites of arachidonic acid, their role in haemostasis and thrombosis. Br Med Bull 1978; 34:129- 135. Remuzzi G, Misiani R, Marchedi D, et al. Haemolytic-uraemic syndrome: deficiency of plasma factor(s) regulating prostacyclin activity? Lancet 1978 2871-872. Van Hoof A, Chaumone DAF, Vermylen J. Plasma from patients with hepatic or renal failure stimulates prostacyclin release from

“exhausted” rat aorta slices. Thromb Haemostasis 1979; 4243. 6. Pace-Asciak CR, Carrara MC, Rangaraj G, et al. Enhanced for-

mation of PGI2, a potent hypotensive substance, by aortic rings and homogenates of the spontaneously hypertensive rat. Prostaglandins

7. Hamilton G, Rosza I, Hutton R, et al. Portal vein prostacyclin activity in experimental portal hypertension in rats. Clin Sci 1981; 60:327-329.

8. Hutton RA, Dandona P, Chow FPR, et al. Inhibition of platelet aggregation by placental extracts. Thromb Res 1980; 17:465-471.

9. Salmon JA. A radioimmunoassay for 6 keto-prostaglandin FI . Pros- taglandins 1978; 15383-398.

10. Holmsen H, Holmsen I, Bernhardsen A. Microdetermination of adenosine diphosphate and adenosine triphosphate in plasma with the firefly luciferase system. Anal Biochem 1966; 17:456-473.

11. Friedman LA, Webster J, Hensby CN, et al. Prostacyclin produc- tion in arterial hypertension. In: Lewis PJ, OGrady J, e&. Clinical pharmacology of prostacyclin. New York Raven Press, 1981: 97- 101.

12. Moncada S, Korbut R, Bunting S, et al. Prostacyclin is a circulating hormone. Nature (Lond) 1978; 273:767-768.

1978; 15:1005-1012.