Evidence for a role of prostaglandin I2 and ...

1 downloads 0 Views 585KB Size Report
is work was suppor-ted by the Ontario Heart Foundation and the. Up,john Company of Canada. 'visiting Scholar of the Hospital for Sick Children Foundation on.
Evidence for a rule of prostaglandin I, and thromboxane A2 in the ductus venosus of the lambH

Can. J. Physiol. Pharmacol. Downloaded from www.nrcresearchpress.com by 192.64.11.124 on 06/04/13 For personal use only.

A . S . 8. A U E A G B O , ~1. B I S H A H , J~ . LEES, P. Ekl. OI~LEY, AND F. COCEANI' Research Pfzstifute. The Hospital for Sick Chibdren, Toronto, Oni., C'anudl~M56; 1x8 Reccivcd November 23, 1984 ADEAGBO, A. S. $., I. I%ISBIAII, .I.LEES, P. M . OLLEY,and F. CBCEANI. 1985. Evidence for a role of prostaglandin l2 and thromboxane A2 in the ductus venosus of the lamb. Can. J . Physic~I.Pham~acc~I. 63: 1 101- 1 105. The prostaglandin (PG) endoperoxide, PGH2. and thc thronlboxane (TX) A-, analog, 9'11 -epithio-1 I . 12-tnetkano-TXA?, were testcd in vitro on the ductus venosus sphincter from fetal (premature and !nature) and neonatal (I-day-old) lambs. PGH: relaxed the indomethacin-contracted fetal ductus in a dose-dependent manner and its action was reduced aftcr treatment with 15-hydroperesxyarachidc>nic acid. In contrast, reduced glutathione did not affect thc PGW2 relaxation in the indornethacintreated ductus. nor did it rnodlfy the response of the untrcaaed ductus to constrictcsr stirnuli. Unlike gPd;FBPI,, the stable 9a,1 la-epoxymethano-PGHr analog contracted she vessel. Similarly, the TXA2 analog was a contractile agent, its action exceeding that of the PGH2 analog in potency and efficacy. The TXA2 arlalog was active on preparations from both premature (minimum 114 days gestation) and nnature la~nbs,but a maximal sffcct was attained during the perinatal period. These results confirm the existence of a PG-mediated relaxing mechanism.in the ductus veramus and suggest that the active compound is E l 2 . This mechanism is likely responsnble for keeping the ductus patent in the fetus. TXA2. formed within the liver parenchyma. is well suited for playing a role in postnatal closurc of the vcssel. ADEAGBO,A. S . O., I. B ~ S ~ I A J . I LEES, , P. M . OLLEYet F, COCEANI. 1985. Evidence for a role of prostaglandin I2 and thromboxane A2 in the ductus vcnosus of the lamb. Can. .I.Physiol. Pharmacol. 63: 1 I O B - 1 105. On a vksifik les effets irz viaro de I'endopkroxyde de la prostaglandine (PG), PGH2, et de l'analogue de Ia thromboxanc (TX) A2: 9,8 1,dpithio-%1 ,12-m6thano-TXA2, sur le sphincter du canal vcianeux de brebis foekales (prCmatur6es et matures) et nkonatales (ilg6es d'un jour). La PGH, relaxa dc manikre dose-dtpendante le canal toetal contract6 par I'indomCthacinc, et son action fut rkdenite aprks un traitenlent avec l'acide 15-hpdroproxy-arachidonique. A B70ppos6,une dose de glutathione reduite n'affecta pas la relaxation par PGH, dans lc canal trait6 avec j'indomkthacine et elle ne lnodifia pas la rCponse du canal non trait6 aux stimuli du muscle constricteur. C~whraircmenf2 la PGH2, l'analogue stable de la PGH2. 9u, l la-epoxy-rnkthano, contracta le vaisseau. Siinailairernent, l'analogue de lo TXA2, fut un agent contractile, sun action dkpassant celle de I'analsgue de la PGHz en force et en efficacitd. La TXA2 ktait active sur les pr6prations des brebis tana prematurkes (minimum 1 87 jsurs de gestation) que matures, mais un effet maximal fut obtenu durant la phiode p6rinatale. Ces r6sultats confirrmlent l'exisknce d'un rnkcanisrne de relaxation m6diC par la PG dans le canal veineux et sugg2rent que Ba substance active cst Ia PC&. Ce mdcanisme est vraisernblablernent respesnsable du maintien de la persistance du canal dans lc foetus. La TXA2, fornee dam le pmnchyme du foie, est bien appropribe pour jouer un r61e dans la fermekure postnatale du vaisseau. [Traduit par le journal]

We have previously found that indornethacin contracts the ductus V ~ W B S U Sof mature fetal larnabs in vltrs (Adeagbo 46 a / . 1982) and in vdvo (Sideris st cd. 19821, suggesting that an arachidowate cyclooxygenase metabolite actively maintains its patemy. Among the compounds tested, prostaglandin (PG) E? and PGH; are nearly eqlaipotent in relaxing the ductus iam vitro and are both suited for playing this role. dn v i m ) experiments (Adeagbo et ul. 1982) have also shown that stable PC endoperoxide analogs constrict the ductus; however, in those experiments it was not possible to differentiate between a specific effect s f the endoperoxide moiety and an effect mediated through thrsmboxane (TX) receptors. The objective of the present study was twofold: first, to identify the PG relaxing the Quctus venosaas and second, to determine the responsiveness of the ductus t o TX&. A preliminary account of this work has keen reported (Adeagbo et a/. 1984). Nex-term fetal ( 140- 149 days gestation; term 150 days) and new?"is work was suppor-ted by the Ontario Heart Foundation and the Up,john Company of Canada. 'visiting Scholar of the Hospital for Sick Children Foundation on leave of absence from the Department of Pharmacology. University of Ibadan. 'Supported by the Medical Research Council of Canada. 4Author to whom rcquests for reprints should be addressed.

born ( I -day-old) lambs of Wambouillet- SuffoIk crossbreed were used. A few premature animals (I 17- 1 19 or 130- 132 days gestati@na) were also used. General prsceeiure The ductus venosus sphincter was isolated using a published protocol (Adeagbo ei a!. 1982). with special care taken not to touch the luminal surfxe with sharp objects car damage the muscle while separating the vessel from the liver parenchyma. As an additional precaution, the location of the ductus sphincter was ascertained at the start of the dissection not by passing a glass rod through the vessel, as previously reported (Adeagbo a/. 1%82), but by cutting open the porkal sinus The preparation was susper~ded between platinum1 iridium (9: 1 , wjw) hooks in a 10-mL organ bath and its tension was recorded isometrically via a transducer (Grass FT.03C) coupled to a Grass polygraph. The loading tension for term fetal preparations was adjusted to 200-300 rng (mean, 261 mg) and this value was extrap%atedfrom the transnlural pressure in view (see Adcagbo et a / . 1982). Neonatal preparations were subjected to a sl~ghtlygreater load (mean 287 wig; range 250-350 mg), while the load was the same or lower with inlrnature preparations (anean, 243 mg; range, 188-300 mg). The bath was supplied from several reservoirs. and a system of threeway valves allowed the perfusion fluid to be changed rapidly. The perfusion rate was approximately 2 rnljmin, and the tluid temperature was 37°C. Both the resemc~irand the organ bath were continuously bubbled with the required gas naixtenre. A gas mixture of 5% GO2in N2 was used when mounting thc preparation ~nsadethe bath and for a short period (1 7 -39 min) afterwards. Subsequent equilibration and drug tests were conducted with K ~ e b sn~ediumgassed with either 2.5% O2 and 5% C 0 2 in N2 (low Po,) or 5% CO, in O2 (high Po2).

Can. J. Physiol. Pharmacol. Downloaded from www.nrcresearchpress.com by 192.64.11.124 on 06/04/13 For personal use only.

CAN. J . PHYSHOI>.PHARMACOL. VOI-. 63, t985

0-0

X

@--@

@--a

CONTROL 15-HPETE TREATMENT

GSH TREATMENT

FIG. I. Ductus venosus sphincter of mature fetal lamb. Dc~seresponse curves for the relaxant effect of PGHZ (contact time, 6 min) before and during treatment with either 15-WPETE (3 x lo-'' M ; n = 7) or GSH ( 5 X 10 ' M; n 5 ) . 15-HPETE was injected into the bath 4 min prior to PGH2 applications, while GSH was added to the perfusion fluid and kept in contact with the daactus at least I h before the start of the PGH2 test sequence. ,411 preparations were equilibrated at high Po, and were made to contract with either indornethacin alone (2.8 X 10 M ) or indomethacin pIaas excess potassium (20 ~m'bi).Tcnsion developed was 293 2 36 (top panel) and 545 2 70 mg (bottom panel). Asterisk indicates a statistically significant difference at the 5%) level or better (paired t-test).

-

Solutions and drugs The Krebs solutie~nhad the following composition (in nsillianoles per litre): NaCI, 118; KCl, 4.7; CaCl,, 2.5; KH2FPO4,I ; MgS04, 0.9; dextrose. I 1 . B ; and NaHC03, 25. Potassium- Krebs solution (20 or 55 m M ) was prepared by substituting NaCl with an equinaolar amount of KCI. The pH of the solutiorz was 7.4 aftcr equilibration with gas mixtures containing 5% CO,. The following compounds were used: PGIz sodium salt; PGH2; the stable PGH, analog, 9cu, 1 l a-epoxym e t h a n d 5-hydroxypr05ta5.13-ditnoic acid; the stable TXA2 analog, 9, I I -epithis- 1 1'12methano-TXA2 (Katsura et aE. 1983); 15-hyalrc~peroxyarachidowic acid (15-HPETE); reduced glutathione (GSH, Sigma); and indornethacin (Sigma). PGH2 was prepared as previously reported (Coceani et a / . 1978) and proved to be 87% pure ( K G 2 as a major contaminaist). 15-MPETE was prepared from the parent fatty acid by incubation with soybean lipoxidase (Hamberg and Samuelsson 1967). Both compounds were stored in dicthyl ether at -70°C. The PGH, and 'TXA2 analogs were c%issoIvedin ethanol (2-5 mg/mk) and aliquoks of the ethanol solution (stored at -20°C) were dilutcd with saline on the day of the experiment. PG12 was dissolved directly in ice-cold Tris buffer (50 mM, pH 9) prior to use. For tests of PGH1 and 85-HPETE, an aliquot caf the ether solbatioan was evapo-

FIG.2. Daactus venosus sphincter of mature fetal lamb. Relaxant responses to PGH2 and PG12 (contact time, 6 man) before and during treatment with 15-HPETE (3 X 10 M ) . Records were obtained from different preparations. AIB preparations were equilibrated at high Po2 and were made to contract with either indoanebhacin (2.8 x 10 " A l ) alonc (net tension, 394 mg; top tracing) or indomethacin plus excess ) tension, 412 and 400 ang for middle and potassium (20 I ~ M (net bottona tracing, rcspectively). In this and the subsequent figure, doses are molar concentrations in the bath fluid. rated to dryness in an acetone/solid C 0 2bath, and the dry residue was dissolved in ice-cold saline. Indomethacin was dissolved in ethanol (10 rng/mLji prior to the preparation of the final solution in Krebs medium. Prostaglandins and the TXA2 analog were tested in scqbaential doses, using 3- to 10-fold increments in 6). 1 n ~ kvolu~nes. Doses of all co~npoundsare given in molar concentrations and refer to their final concentration in the bath. The concentration of PGH2 was calculated assuming a 100% purity. Control tests proved that the ethanol-containing vehicle had no effect on the ductaas.

AnaE.vsi.s.qf re.spsponst..s Effects of contractile agents were measured by the fractional rise in tension over the basal tension. Basal tension, which varied depending on the animal preparation and the experinlental condition (see Results), is given after correction for the applied tension. S'CPdetermine relaxant responses, tissues wcre contracted by exposure to high Po2 and indo~nethacin2.8 x lo-" M . Hn certain experiments (exps), the contractile tension was increased further by raising the potassium content of the ~nediurnto 20 m M . Data are expressd as the mean 2 SE. Statistical analysis sf unpaired and paired data has been made using the Student's t-test. hfultiple comparisons have been made with an analysis of variance and a Duncan's maaltiple range Best.

Results The behaviour of the ductus venosus sphincter from term lambs resembled that described in a previous publication (Adeagbo 6.t a&.B 982) in which a different breed of animals was eased. During the initial equilibration in Krebs mediunl gassed with 5% @02 i ~ aN2, the preparation maintained its resting force or, more often, relaxed slightly. Subsequent exposure to either Bow or high Po2 caused a slow contraction, the peak tension being 63 f 22 (nine exps) and 140 2 13 mg (26 exps), respectively. At either Po2, contractile responses were not sustained and, in 0.5- 1.5 k. the muscle tension stabilized at values

TABLEI . Effect of GSH asn contractile responses of the ductus venosus sphincter fronl term fetal lambs

Contractile agent Thromboxane analog' (7 X 168 ') M )

Can. J. Physiol. Pharmacol. Downloaded from www.nrcresearchpress.com by 192.64.11.124 on 06/04/13 For personal use only.

Potassiun~(55 m M )

Before GSH

During GSH"

ph

589250 (3) 341 -+ 112 (31

546569 (3) 291 L86 (3'3

NS NS

N076. Preparat~omwerr equ~lsbratedat a h ~ g hP o 2 . Valucb (mean C SE) ~ndicatethe niaxirnal tension developed (~nilligranas)fcjr the nurnbrr of cxperinacnts given irl parentheses. NS, not significant. "GSH was added to the perfhsion fluid to a final cuncentsztion of 5 X IOV" h4 "Paired I-test. 'Ccpmpounci (9,B I-epithio-1 1 , IZ-1~1ethan(1-T>02~) was applied fo the orgi~r~ bath and kept in contact with the tissue Fvr ;I 6-rnin period.

close to those set initially. Occasionally, the contraction had a rapid onset and subsidence. In contrast to oxygcn alonc. indornethacin (2.8 X 10 &f ) combined with a high Po, caused a contraction (299 -+ 32 mg; 18 exps) that was maintained for as long as the trsatmment was continued. Furthern-nore, indomethacin-treated preparations developed a variable degree (sf spontaneous motility in most cases. 'The neonatal ductus differed from the mature fetal ductras in that it contracted less (58 t I$ mg: three exps) to a high while the piSen~ature fetal d u c t ~ ~reacted s marginally C530 nmg; t h ~ exps) e or not at aH1 (three exps) to eitlwr PCB?value. &@ecas of PGH2 Ja-a the absence of intrinsic tone, the ductus sphincter did not respond to PGH, over a dose range up to ii0Y7M (three exps). In contrast, once the tone of the preparation was raised with indomethacin (with or without 20 m M potassium; see Methods) PGHL rslaxed the tlssue, with a threshold around lo-' M (Fig. 1). The time cc-mrse of the relaxation varied among preparations, cven though relaxation was usually immediate or nearly immediate in onset. Figure 2 exemplifies extreaxies in the spectrum of respcmses to PGHz. Regardless of the tirnc required for the peak cffect, reversal of thc responses was slow and, at the maximal dose employed (10-I M ) , the original tensicpel was restored 6-44 min (mean, 28 mi1aB after washing the preparation. 'IFhe magnitude of the PGH2-induced relaxation was also varied, but this occurrence as well as t8ifferences in the pattern of responses had no obvious relation with either the level of intrinsic tone of the muscle or the use of indomethacin alone versus indomethacin plus excess potassium in raising the tone. Fudhernacare, there was no evidence suggesting the prcsence of tachyphylaxis in the response caf the ductus to PGHL. Unlike the natural endcaperoxide, the 9a,1l a-cpox ymethano endoperoxide analog contracted the ductus and its action was dose dependent (two exps). The thresholc8 concenand a maximal tration ranged between 0.3 and 3 X lo-", contraction f mean, 556 mg) was attained at 10 'Ikf. The latter finding confirms data obtained previously in a different breed of lambs (Adeagbo t9f ak. 1982). I_$-HPETEin a concentration of 3 x 10 ' M had no effect on ductal tone or sometimes produced a modest contraction (Fig. 2). Thc same treatment. however, markedly red~lceclor even abolished the relaxant response to PGHl (Figs. 1 and 2). While interfering with the relaxation, I-HPETE did not unmask any contractile coanponent in the action of the endoperoxide (Fig. 2). Thc inhibitory effect of 15-HfPETE cornpletely subsided upon washing the preparation, but it could be elicited again in successiivs applications. Unlike PGH2, YG4

FIG. 4. Diactus venosus sphincter of mature fetal larnb. Dose-- 5 - 10). Responses were not significtantly different at low and high Po. and results were p o l e d

response CUKVG to the TXA2 analog (rz

together. Insert: rcspsnse to a maximally effective concentration of the soanpound at low Po7.

was equally effective in the absence and presence of i 5-HPETE (Fig. 2) (two exps). The latter finding mles csut any unspecific effect of 15-HPETE on muscle reactivity. Brctreatirmemt of indometh:icin-ccb~~tr'dctedtissues with GSHJ (5 X M ) tended to increase the relaxant response tc~PGH2 (Fig. I), but changes were not significant. F~arthennore,GSH only ~narginallyreduced constrictor responses elicited in the normal tissue by either an intermediate dose of the TXA2 analog (see below) or excess potassium (Table l).

E:#~CPS

7XA2 artalog The stable TXA2 analog contracted the 1natur-efetal ductus in a dose-dependent rnanner and its action exceeded that of the endoperoxide analog in both threshold concentration and peak effect. Likewise, the TXA, analog was a more effective constrictor than 5 mM potassium (Tablc 1) (cf. Adeagbo et a!. 1982). As shown in Fig. 3, responses were sustained and had phasic discl~argehof variable annlplitude superirnposcd. Sensitivity to the TXA? analog developed carly in gestation, but maximal potency and efficacy were attained during the perinatal period (Table 2). Pretreatment elf the mature fetal ductus with indo~nlethacin(2.8 X 10 ' M ) resulted in enhanced responsiveness to TXA2, the ED,, for the contraction decreasing from a control value of 2.5 t Om4nM (10 exps) to a value of 0.8 9 0.4 nM Cfour exps) ( P < 0.05. unpaired t-test).

Discussicemn The present investigatiora confirrns the existence of a prostaglandin-mediated relaxing mechanism in the ductus venosus and suggests that the active compound is PCPr12. This rnechanisn-mis likely responsible for keeping the vessel patent in the fetus. Several findings are consistent with this possibility. ( i ) PG12is a potent dknctus dilator, at least in vier() (Adeagbo ut k z l . 1982). (dl) PGIj2, which is the endc~percsxideof choice for assessing the relative activity of terminal enzyrlaes in the cycls-

CAN. J. PHYSHOL. PHARMACOL. VOL. 63. 8985

TABLE2. Contractile response of the ductus venosaas sphincter to the thromboxane analog Response Maximum contraction Threshold ( M )

Age Fetus, 114-119 days (group I)

Can. J. Physiol. Pharmacol. Downloaded from www.nrcresearchpress.com by 192.64.11.124 on 06/04/13 For personal use only.

Fetus, 130-132days(groupII) Fetus, 143- 149 days (group 111) Newbom,Bdayold(groupIV)

EDSO

(M)

(nag)"

0.7-3.5X 10 " . 8 ~ 1 . 5 ~ 1 0 " 64) 44) 0.1-1x10" 3.990.8X 10-" (5) (5) I -7 X 10-" 2.520.4X 10 -" (I(]> (10) 1-1.5X10-'0 240.4X10' (3) 43) Statistical analysis (level of significance, P )

Groups compared Group Group Group Group Group Group

Threshold

EDSO

Maximum contraction

I vs. 11 I VS. III I vs. HV I1 vs. III HI vs. IV III vs. IV

NOTE: Figures are aneans + SE for the number of experiments given in parentheses. Responses were not significantiy different at low and high PO, and results were pooled together. "Response obtained at the I0 M concentration.

'

oxygenase pathway, is also a dilator agent and its effectiveness is reduced in 15-HPETE treated preparations; 15-HPETEInterferes with the PG1, synthase (cf. Bunting et al. 1976), while leaving the responsiveness of ductal muscle to PG12unaffected. ( iii) GSW, a compound promoting the synthesis of PGE2 at the expense of other cyclooxygenase products (van Dorp 1967). has an insignificant effect on the PGH2-inducedrelaxation and on contractile responses as well; these findings argue against the presence of a functional PGH2 to PGE2 isomerase in the ductus and suggest that the relaxant action of PCE2, reported previously (Adeagbo al. 8982), is pharmacologic in character. (iv) The prostaglandin mechanism relaxing the ductus is strong enough to curtail the contractile effect of the TXA, analog. While docunaenting the intramural conversion of PGH, to a dilator prostaglandin, conceivably PGI,, our study indicates that the contractile response of the ductus to the endoperoxide analogs reflects a thromboxar~elikeaction. Endoperoxide analogs are known to mimic TXA, in vascular and nonvascular smooth muscle (Coleman et LIC.1981). The TXA2 analog is a potent constrictor, exceeding in its effect all vasoactive agents tested so far (Adeagbo et ak. 1982; Coceani, Adeagbo et a / . 1984). The latter finding touches on the issue of the postnatal closure of the ductus which, in our view (Adeagbo et al. 1982), is also an active event mediated by local and blood-borne stimuli. If indicative of a physiological situation, the observed potency and maximal efficacy of the T%A2analog at the time of birth make TXA2 well suited for being a prime effector of ductus closure. The compound could act alone or in concert with a-adrenoceptor mediated constrictor influences (Coceani, Adeagbo et al. 1984). According to our postulate, TXA,, which is not formed ir~trarnurally(Adeagbo et a / . 1982), originates from the liver parenchyma (Pace-Asciak and Rangaraj

1978) and reaches effective levels in the vessel wall as a result of the postnatal increase in hepatic arterial blood flow and the ensuing rise in tissue oxygenation. Significantly, a functional interaction between parenchymal cells and blood vessels is also thought to be important in the adult liver for controlling the viability of translobar arteries (Mays and Mays 1983) and arteriovenous anastomoses (WlcCuskey 1966). This study reasserts the basic difference between the ductus arteriosus and the ductus venosus, notwithstanding their common dependence on a cycloowygenase product for prenatal patency. While PGE2is likely to maintain patency of the ductus arteriosus (Coceani and Olley 1983), PGI, is ascribed an equivalent role in the ductus venosus. TXA2is a prime candidate for causing closure of the ductus venosus after birth. The ductus arteriosus, on the other hand, is unresponsive to both the natural TXA2 (Coceani et al. 1978) and its synthetic analog (F. Coceani, J. Lees, and P. M. Olley, unpublished data) and may instead be dependent on product(s) of a cytochrorne P-450 linked rnonooxygenase reaction for its closure (Coceani, Hamilton et wk. 1984).

Acknowledgements We gratefully acknowledge the excellent assistance of Fr'dncis Hamilton, Thomas VanHelder, Carole Breen, and Anne Johnston. Prostaglandins and the thromboxane analog were kindly provided by, respectively, the UpgoAn Company and Ono Pharmaceutical. ADEAGBO, A. %. o.,F. COCEAWI, I. BISHAI,and P. M. QELEY. 1984. Involvement of PG12 in prenatal patency of the lamb ciucl~ls venosus. Pa International Conference on Prostaglandins and L u k o trienes. Edited b?pl B. M. Bailey. Abstr. No. 244. ADEAGBO,A. S. O., F. COCEANI,and P. M. ~ L L E Y1982. . The response of the lamb ductus venosus to prostaglandins and

Can. J. Physiol. Pharmacol. Downloaded from www.nrcresearchpress.com by 192.64.11.124 on 06/04/13 For personal use only.

anhibitors of prostaglandin and thromboxane synthesis. Circ. Res. 51: 580-586. S. MONCADA, and J . R. VANE.1976. BUNTING. S., R. GRYGLEWSKI, Arterial walls generate from prostaglandin endoperoxides a substance (prostaglandin X) which relaxes strips of mesenteric and coeliac arteries and inhibits platelet aggregation. Prostaglandins, 12: 897-913. CWEAR'I,F.. A. S. 0. ADEAGBO, E. CUTZ.,and P. M. e)Lr.EY. 1984. Autonomic mechanisn~sin the ductus venosus of the lamb. Am. J. Physiol. 247: H17-H24. CCKEANI, F., I. BISHAI,E. WHITE,E. BODACH.and P. M. OLLEY. 1978. Action of prostaglandins, endoperoxides, and thrornboxanes on the lamb ductus arteriosus. Am. J . Physiol. 234: H117 -H 122. C C ~ E A NF.. I , N. C. HI~MILTON, J. LABUC,and P. M. OLLEY.1984. Cytochron~eP4so-Iinked monooxygenase: involvement in the lamb ductus arteriosus . Am. J. Physiol. 246: H640 -H643. COCEANI, F., and P. M. OL,LEY.1983. Prostaglandins and the ductus arteriosus. Pediatr. Cardiol. 4: 33 - 37. COLEM.L~N, R. A . , P. P. A. HUMPHREY, I. KENNEDY, G . P. LEVY,and P. LUMLEY.1981. Comparison of the actions of U-46619, a prostaglandin Hz-analog. with those of prostaglandin H2 and thromboxane A- on some isolated smooth ~nusclepreparations. Br. J . Pharmacol. 73: 773-778.

HAMBERG, M., and B. SAMUELSSON. 1967. On the specificity of the oxygenation of unsaturated fatty acids catalysed by soybean lipoxidase. J . Biol. Chem. 242: 5329-5335. KATSURA,M., T. MIYAMOTO,N. HAMANAKA, K. K o N ~ T. , TEWADA, Y . OHGAKI.A. KAWASJ~KI, and M. TSUBOSIIIMA. 1983. bra vitro and 8'11 vivo effects of new powedu1 tkrcm~boxaneantagonists (3-alkylamino pinane derivatives). IH Advances in prostaglandin, thromboxane, and leukotriene research. Vol. 11. Edited by B . Samuelsson, W. Paoletti, and P. Ramwell. Raven Press. New Ymk. pp. 351-357. MAYS,E. T.. 11, and E. T. MAYS. 1983. Are hepatic arteries endarteries'? J . Anat. 13'7: 637-3544. MCCUSKEY, W. S. 1966. A dynamic and static study of hepatic arterioles and hepatic sphincters. Am. J . Anat. 119: 455-478. PACE-ASCIAK, C. R., and G. RANGARAJ. 1978. Distribution of prostaglandin biosynthetic pathways in organs and tissues of the fetal lamb. Biochim. Biophys. Acta. 528: 512-514. SIDERIS, E. B., K. YOKCBCMI, T. VANHEEDER. F. COCEANI, and P. M. OLLEY.1982. Effects of indornethacin and prostaglandin E2 (PGE2) on the lamb fetal ductus venosus. Circulation, 66: II- 112. VAN DORY,D. A. 1967. Aspects of the biosynthesis of prostaglandins. Prog. Biochem. Pharmacol. 3: 7 1-82.