- Email: [email protected]
Contribution of arachidonate metabolites to basal and thyrotropin releasing-hormone-stimulated release of prolactin from purified lactotrophs in primary culture
Llfe Sciences, Vol. 47, pp. 1829-1836 Prlnted in the U.S.A.
Pergamon Press
CONTRIBUTION OF ARACHIDONATE METABOLITES TO BASAL AND THYROTROPIN RELEASING-HORMONE-STIMULATED RELEASE OF PROLACTIN FROM PURIFIED LACTOTROPHS IN PRIMARY CULTURE Marie-Pierre JUNIER, Jean-Maro ISRAEL*, Fernand DRAY*°, Jean-Didier VINCENT* Oregon Regional Primate Research Center, Div. of Neurosciences, Beaverton, OR, 97006, USA, * INSERM U176, 33077 Bordeaux Cedex, France, and ** INSERM U207, Institut Pasteur, 75015 Paris, France (Recelved in flnal form September I0, 1990) Summary
Among the different biochemical pathways which have been suggested to play a role in the control of prolactin (Prl) release from anterior pituitaries, arachidonate and its metabolites have been proposed to be involved in the process of Prl release. In this study we investigated the contribution of arachidonate metabolites to both basal and TRH-stimulated Prl release from perifused lactotrophs in culture (derived from pituitary glands of lactating female rats), which exhibit a high sustained release of Prl in absence of inhibitory input. Inhibition of the general oxidative metabolism of arachidonate by 10-5 M ETYA or of the arachidonate lipoxygenase metabolism by 10-5 M NDGA decreased basal Prl release to 45__.10% (n=3) and 36±4% (n=6) of the control release, respectively. Indomethacin, an inhibitor of the cyclooxygenase pathway, was without effect. Of the lipoxygenase metabolites tested at 10-6 M only 15-HPETE and 15-HETE induced Prl release. 15-HETE elicited prolactin release in a concentration dependent manner with a maximal effect at 10-6 M (10.72+3 ng/ml vs control 5.1±0.8 ng/ml, n=3). The quantity of Prl release induced by TRH was markedly decreased in the presence of NDGA. However, the fraction of Prl release elicited by TRH, calculated as a percentage of the amount of Prl released prior to TRH application, was similar under control conditions, and in the presence of NDGA. In contrast, inhibition of the protein kinases A and G by H8 (10-5 M) failed to alter basal Prl release but inhibited the effect of TRH by 58+6% (n=3). These data suggest that in absence of inhibitory inputs the high sustained release of Prl observed in cultures of lactotrophs derived from lactating female rats depends on the availability of lipoxygenase metabolites, and that the blockade of lipoxygenase reduces the absolute amount of Prl released by TRH without suppressing the ability of TRH to stimulate Prl release. The hormone release properties of highly enriched lactotrophs in culture, obtained from lactating female rats, have been characterized in correlation with their electrophysiological activity (1,2,3). They have been shown to be excitable, and to exhibit high levels of prolactin (Prl) release. The application of dopamine (DA), the main Prl releasing-inhibiting factor (4), induced a hyperpolarizing response, and inhibited Prl release. On the contrary, thyrotropin releasing-hormone (TRH), one of the best known hypothalamic releasing factor of Prl (5), induced a depolarization response characterized by a plateau potential (3), and stimulated Prl release. To further study the mechanisms underlying Prl release in this system we explored in the present study the involvement of arachidonate metabolites in basal and TRH-stimulated Prl release. Indeed, among the different biochemical mechanisms which have been proposed to be implicated in the control of Prl release, the products of membrane phospholipid metabolism have received increased attention. Noteworthy, different studies conducted either on cultures of antepituitary cells or on tumoral cell lines have suggested that, among those products, arachidonate and its metabolites are involved in the control of basal and stimulated Prl release (6,7,8). Moreover, recent studies have shown that arachidonate and its metabolites may modulate ionic channels activity in different biological systems (9,10,11,12). Corresponding author : M.P. Junier, ORPRC, 505 N.W. 185th Av., Beavedon, OR, 97006, USA 0024-3205/90 $3.00 +.00 Copyright (c) 1990 Pergamon Press plc
1830
Arachidonate and Prolactin Release
Vol. 47, No. 20, 1990
Using pharmacological agents which alter arachidonate metabolism, we show that basal Prl release depends on the availability of lipoxygenase arachidonate metabolites, while TRH remains able to elicit Prl release despite suppression of arachidonic acid metabolism. Material and Metho~,~ Animals Anterior pituitary cells were obtained from lactating female rats of the Wistar strain during the last week of lactation. Cell culture Cell dissociation and preparation of the cells for culture have been described in detail previously (13). Briefly, after enzymatic (trypsin 0.5% for 15 min.) and mechanical dissociation of the pituitary tissues, the cells were separated on a continuous gradient of 0.3-2.4% bovin serum albumin and collected in 100 ml fractions numbered 2-9 after rejection of the first 300 ml which contained red blood cells and cellular debris. Ninety five % of the cells collected in fractions 3 to 5 were lactotrophs (2). These cells were harvested, resuspended in Dubbelco's Modified Eagle's Medium (DMEM) containing 10% newborn calf serum, 3.7 g/I NaHCO3 at pH 7.4, and plated on 35 mm Petri dishes (Nunc, Denmark) at 106 cells per dish. Antibiotics (35 mg/I penicillin, 50 mg/I steptomycin) were present during the first 2 days of culture and thereafter omitted. The medium was replaced every 3 days. Prl release experiments After 7 days of culture, perifusion experiments were performed in the original Petri dishes. A continuous-flow perifusion system, which accomodated four dishes at a time, was used with a four channels peristaltic pump (2). Constant temperature (37 C) was maintained by a water heated platform. The perifusion medium consisted of DMEM supplemented with 15 mM Hepes and 12 mM NaHCO3. Flow rate was 0.5 ml.min-1 and medium levels remained at about 1.5 ml per dish. After a 90 min. stabilization period, 2 rain. samples were collected with a fraction collector into tubes containing 50 ml of 1% BSA in PBS. Aliquots were stored at -20 C until radioimmunoassayed. After a 20 min. collection period used to define basal Prl release, test compounds or their vehicle were directly added to the medium for various lengths of time, as indicated in the results section. The total perifusion period did not exceed 60-80 min. Radioimmunoassav The amount of Prl released into the medium was assayed using the RIA kit provided by the National Hormone and Pituitary Programm (Baltimore, MD, USA). Prl values were expressed as nanograms equivalents of rat standard Prl RP-1 per ml. Significant differences between groups were assessed with Student't test. A P of less than 5% was considered to indicate a significant difference. Results are presented as mean+sem. Material The hydroperoxyeicosatetraenoic acids (HPETEs, Sigma), the 15-hydroxyeicosatetraenoic acid (15-HETE, Sigma), and 5,8,11,14 eicosatetraenoic acid (ETYA, a gift from Hoffman La Roche, France) were dissolved in absolute ethanol under an argon blanket at concentrations of 1.5x10-4M, 1.6x10-4M, and 10-2M, respectively. 4-bromophenacylbromide (PBPB, Sigma) was dissolved in 0.1% ethanol at the concentration of 10-2 M. N(2-(methylamino)-ethyl)-5-isoquinoline sulfonamide (H8, Calbiochem, France) and nordihydroguaiaretic acid (NDGA, Sigma) were dissolved in dimethylsultoxide at the concentration of 10-2 M. Neomycin sulfate (Sigma) and indomethacin (Sigma) were dissolved in DMEM at the concentration of 10-2 M. All the drugs were aliquoted and stored either at -80 C (HPETEs, 15-HETE, ETYA) or at -20 C except indomethacin which was prepared extemporaneously. Final test concentrations were obtained by diluting the compound with DMEM. Neither ethanol nor DMSO at the concentrations used had an effect on Prl release. Results Arachidonate is released upon the action of either phospholipase A2, which cleaves the phospholipids sn-2-acyl bound, or upon the combined action of phospholipase C and diacylglycerol lipase on phosphatidylinositol (14). To investigate the role of arachidonate in the control of basal Prl
Vol. 47, No. 20, 1990
Arachidonate and Prolactin Release
1831
release, either PBPB, a compound which binds covalently to phospholipase A2 and blocks its activity (15), or neomycin, which inhibits phospholipase C by binding phosphatidylinositols (16), was administered at 10-5 M during a 30 min. perifusion period (figure 1). Both PBPB and neomycin strongly inhibited Prl release, the hormone levels decreasing to 49-+7% and 42+-9% (n=3) of basal release, respectively. Using pharmacological agents which alter arachidonate metabolic pathways we attempted to determine which of them was involved in the maintainance of Prl release in our system. The cells were exposed for 30 min to 10-5M of either ETYA, an inhibitor of the general oxidative metabolism of arachidonate (17), or indomethacin, an inhibitor of the cyclooxygenase pathway (18), or NDGA, an inhibitor of the lipoxygenase pathway (19). Both ETYA and NDGA reduced Prl release to 45+_10%, n=3, and 36-+4°/,,, n=6, of the control levels, respectively. Indomethacin was without effect (figure 2). In order to determine which lipoxygenase metabolite(s) is involved in the maintainance of basal Prl release, the three immediate products of lipoxygenase activity, 5-, 12- and 15-HPETE, were tested at the concentration of 10-6 M. Only 15-HPETE stimulated Prl release (figure 3A). 15-HETE, the direct metabolic product of 15-HPETE, induced a dose-related increase in Prl release with a maximal and statistically significant effect at 10-6 M (10.72:1:1.53 ng/ml vs control 5.56+_0.57 ng/ml, n=4, figure 3B), while 5- and 12-HETE were without effect (data not shown). To examine the involvement of lipoxygenase metabolites in the stimulatory effect of TRH on Prl release, TRH was added to the perifusion medium in the presence of NDGA. Like basal Prl release the absolute amount of Prl released by 10-7 M TRH was markedly reduced in the presence of NDGA, as compared to the effect of 10-7 M TRH alone (figure 4A). However, NDGA did not suppress completely the sfimulatory effect of TRH. Indeed the amount of Prl released during the 10 min. preceeding TRH application doubled in the presence of TRH (figure 4A). To estimate to which extent TRH remained able to elicit Prl release despite the blockade of lipoxygenase, we expressed the amount of Prl release as a percentage of Prl release during the 10 min. preceeding TRH application, i.e. when NDGA exerts its maximal inhibitory effect. The transformed data are presented in figure 4B; the maximal Prl release induced by TRH in the presence of NDGA represented 83__.4%(n=3) of the maximal Prl release induced by TRH in the absence of NDGA. Indomethacin was either without effect or induced a slight potentiation of TRH stimulatory effect (data not shown). For comparison purposes we tested the effect of cyclic nucleotides-dependent protein kinases inhibitors on the TRH-stimulatory effect on Prl release. Indeed, the involvement of cyclic nucleotides in Prl release has been demonstrated (20,21,22,23,24). Using H8, a preferential inhibitor of protein kinases A and G (25), we observed that, contrary to NDGA, H8 (10-5 M) did not affect basal Prl release, but inhibited TRH effect by 58+6% (n=3) (figure 5). Discussion Arachidonate and its lipoxygenase and epoxygenase metabolites are synthesized in the pituitary gland (26,27), and have been suggested to be involved in the basal and stimulated Prl release from antepituitaries (7). In this study we show that lipoxygenase metabolites of arachidonic acid are involved in the maintainance of basal Prl release from lactotrophs in culture, while TRH appears to be able to elicit Prl release in absence of a functional lipoxygenase. Lactating female rats were chosen as donors since their pituitaries yield four times more lactotrophs than others donors. The use of highly purified lactotrophs ensures that the effects observed indeed result from a direct action of the test compounds on lactotrophs. Blockade of the general oxidative metabolism of arachidonate or selective inhibition of lipoxygenase activity, which initiates one of the three known metabolic pathways of arachidonate (28), resulted in the suppression of basal Prl release. In contrast, inhibition of cyclooxygenase activity did not alter Prl release. The specificity of 10 p.M NDGA towards lipoxygenase in cultures of dissociated pituitary cells has been previously shown (29). These findings are in agreement with the observations of others on different models including hemipituitary glands and cloned strain of pituitary cells (30, 8).
1832
Arachldonate and Prolactln Release
A
Vol. 47, No. 20, 1990
NEOMYCPE 12 10-
8 E
==
ii 6'
4-
2-
1o 9
.
. 2O
.
. 30
.
. 40
.
. 50
. 60
i ..... | ..... , ..... 70 80 90 100
Time (min)
PBPB
10
8 7
~'-
..... i . 10
4
.
0
....
0
10
20
30
40
50
60
70
Time ( r a i n )
80
..... i ..... i ..... i 10 20 30
.
.
.
.
.
.
.
.
40
.
.
.
.
.
.
.
.
50
.
.
.
.
60
i ..... i ..... i ..... 70 80 90 100
Time (min)
FIG• 1
FIG. 2
Effects of 10 p.M neomycin (A) or PBPB (B) on spont-aneous Prl release during a 30min and 20 min perifusion period, respectively• Note the rebound of Prl release after neomycin withdrawal. These curves are representative of 3 experiments•
Effects of 101~1ETYA( [] ), NDGA (z~), or indomethacin ( • ) on spontaneous Prl release during a 30 min application period. These curves are representative of 3 (ETYA, indomethacin) to 6 experiments (NDGA).
Prior to its metabolization, arachidonate is liberated from its storage sites in the membrane phospholipids. This release can result either from the action of phospholipase A2, which cleaves the acyl bond between the phospholipid and the fatty acid, or from the combined action of phospholipase C and diacylglycerol lipase on phosphatidylinositol. To determine the source of arachidonate in our system, we examined the effect of specific inhibitors of each phospholipase• In both cases we observed inhibition of Prl release• This result suggests that either both phospholipases A2 and C may be involved in the mobilization of arachidonate in our system or that both products of phosphatidylinositol metabolism and arachidonate melabolism are implicated in the control of basal Prl release. The inhibitory effect of phospholipase A2 blockade on Prl release corroborates previous studies (30, 31, 6). To our knowledge,
Vol. 47, No. 20, 1990
Arachidonate and Prolactln Release
1833
14 12 10
i
.... • • • • • • 0
10
20
30
40
50
Time (rain)
B
ip,o oo,i 13
9.75
CL
6.5
"E r~ 3.25
0 C 10-SM
C 10-TM
C 10-6M
FIG. 3 A: Effect of 1 ldVl of 5- ( o ), 12- ( • ), and 15-HPETE ( • ) on Prl release during a 10 min application period. These curves are representative of 3 experiments. B: Effect of increasing concentrations of 15-HETE on Prl release (C=control, mean+sem, 15-HETE 10-8M, 10-7M, n=3, 15-HETE 10-6M, n=6). no study has been reported concerning the consequences of phospholipase C inhibition on Prl release. However, phosphatidylinositol turnover has been shown to be altered by DA and TRH in a manner that correlates well with the changes in Prl release (32,33). Among the different lipoxygenase metabolites tested 15-HETE, and its precursor 15-HPETE stimulated Pd release in our system. In contrast, 12-HPETE and 5-HPETE, the two other main precursors of arachidonate lipoxygenase metabolites, were ineffective. A stimulatory effect of 15-HETE on Prl
1834
Arachidonate and Prolactin Release
N D G A ~ -TRH"
A 20 18
B
13_
-TRH-
350
==
16 14 crJ t-
Vol. 47, No. 20, 1990
300 250 -
12 10 8
u) (1= .J~ O
6 4
EL
200 150
,°°t, -
2 0 10
20
30
40
50
60
70
80
Time (min)
40
50 60 70 Time (min)
80
FiG. 4 A: Effect of 10 I~M NDGA on 0.1 ~M TRH-elicited Prl release ( A ) compared to the effect of TRH alone( • ). B: Prl release expressed as % of control release in the presence of NDGA ( ,~ ) or in the absence of
NDGA ( • ) . These curves are representative of 3 experiments.
b
H8
-
-
TRH
30 25
FIG. 5
20 Co c
Effect of 10 tIM H8 on 0.1 I~M TRH-eficited Prl release ( o ) compared to the effect of TRH alone( • ). These curves are representative of 3 experiments.
15
n
t05"
0
•
10
20
30
40
50 60 Time (min)
!
70
•
!
80
•
i
•
90 100
release has also been reported on GH3 cells (8), and on static incubations of dispersed pituitary cells (7). However in this last study 15-HETE stimulated Prl release only at 5 ~M, while 5-HETE efficient concentrations ranged from 1 to 5 MM. 5-HETE being the most unstable of the HETEs, its eventual degradation in our perifusion system shall be envisaged. Different cells donors (non lactating female rats in the cited study) and cell preparation (enriched lactotrophs in this study) could also account for this discrepancy.
Vol. 47, No. 20, 1990
Arachldonate and Prolactln Release
1835
It is interesting to note that recent results have shown the presence of Ca2+ channels active at the resting potential on the lactotrophs (P.M. Lliedo and J.M. Israel, personnal communication). Moreover, in the same model and under similar experimental conditions, the inhibition of Prl release in the presence of DA, which is known to act in a Ca2+-dependent manner (34,35), is analogous to that induced by blockade of lipoxygenase activity (1). This and the previous demonstrations of the ability of arachidonate metabolites to activate ionic channels in neural and cardiac cells (9, 10, 11, 12), suggest a possible link at the molecular level between arachidonate metabolites and the electrical activity of the lactotrophs. Further experiments are required to test this hypothesis. In addition to alter basal Prl release, lipoxygenase metabolites of arachidonic acid have been suggested among other compounds to mediate the stimulatory effect of TRH on Prl release (30). In this study the inhibition of lipoxygenase activity by NDGA induced a strong decrease of the absolute amount of Prl released by TRH. However, considering the relative amount of Pd released by TRH, as compared to the amount of Prl released prior to TRH application (as illustrated in figure 4B), we noticed that the fraction of Prl release elicited by TRH was not markedly altered in the presence of arachidonate metabolism inhibitors. In contrast H8, a preferential inhibitor of protein kinases A and G, failed to alter basal Prl release but strongly reduced the amount of Prl released by TRH. These results show that, although the absolute amount of Prl release triggered by TRH depends from the quantity of basal Prl release, the basal and TRH-stimulated Prl release may be manipulated independently. This suggests that TRH may use metabolic pathways different from the ones involved in the maintainance of basal Prl release. These data do not exclude an involvement of arachidonate metabolites in the TRH effect under normal conditions, especiaity since TRH has been shown to stimulate arachidonate release (7, 36). However these results suggest that more than one metabolic pathway may be used by TRH to elicit Prl release. One of these pathways could involved the cyclic nucleotides-dependent protein kinases A and G as suggested by the inhibitory effect of H8 on TRH-eliclted Prl release. H8 has been found to inhibit both cyclic nucleotide-dependent protein kinases and protein kinase C. However, the H8 concentration required to inhibit protein kinases A and G is far lower than the one needed to inhibit protein kinase C (apparent Ki of 0.48 and 1.2 p.M versus 15 p.M lor protein kinase C), (25). An eventual effect of H8 on protein kinase C can not be totally excluded. But the fact that, on one hand the inhibition of phospholipase C activity, which leads to the formation of a well known stimulator of protein kinase C i.e. diacylglycerol (37), decreased basal Prl release, while on the other hand H8 was ineffective on basal Prl release, indicates that protein kinase C activity was not affected by H8. In conclusion, our results suggest that a) in absence of inhibitory inputs the high sustained release of Prl observed in cultures of lactotrophs derived from lactating female rats depends on the availability of lipoxygenase metabolites, b) among the lipoxygenase metabolites 15-HETE and its precursor 15-HPETE are stimulators of Prl release, c) the blockade of lipoxygenase reduces the absolute amount of Prl released by TRH but does not suppress the ability of TRH to stimulate Prl release, and d) protein kinases A and/or G may be involved in the stimulatory effect of TRH on Prl release. Acknowledaement We thank D. Verrier lor her excellent technical assistance. References 1. 2. 3. 4. 5.
C. KIRK, These de doctorat de 3eme cycle, Universite de Bordeaux II, France (1986). J.M. ISRAEL, C. KIRK and J.D. VINCENT, J. Physiol. 390 : 1-22 (1987). J.M. ISRAEL, A.P. KUKSTAS and J.D. VINCENT, Neuroendocrinology, in press. R.M. MACLEOD, Frontiers in Neumendocrinoloay, p.169-174, Raven Press, New-York (1974). Y. KATO, N. MATSUSHITA, H. OHTA, K. TOJO, A. SHIMATSU and H. IMURA, The Pitultaw Gland, p. 261-278, Raven Press, New-York (1985). 6. L. GRANDISON, Endocrinology 114 : 1-7 (1984). 7. A.M. JUDD, B.L. SPANGELO, J.T. EHRETH and R.M. MACLEOD, Neuroendocrinology48:407-416
(1988). 8. M. RABIER, C. CHAVIS, A. CRASTES DE PAULET and M. DAMON, Neuroendocrinology47 : 323-328 (1988).
1836
Arachldonate and Prolactln Release
Vol. 47, No. 20, 1990
9. D. PIOMELLI, A. VOLTERRA, N. DALE, S.A. SlEGELBAUM, E.R. KANDEL, J.H. SCHWARTZ and F. BELARDETTI, Nature, Lond., 328 : 38-43 (1987). 10. D. KIM and D.E. CLAPHAM, Science 244 : 1174-1176 (1989). 11. Y. KURACHI, H. ITO, T. SUGIMOTO, T. SHIMIZU, I. MIKI and M. UI, Nature, Lond., 337 : 555-557 (1989). 12. R.W. ORDWAY, J.V. WALSH and J.J. SINGER, Science 244 : 1176-1179 (1989). 13. J.M. ISRAEL, C. DENEF and J.D. VINCENT,Neuroendocrinology37 : 193-199 (1983). 14. M.B. FEINSTEIN and R.I. SHA'AFI, Calcium and Cell Function. p. 337-376, Academic Pess, Newoyork (1983). 15. J. DRENTH, C.M. ENZING, K.H. HALK and J.C. VESSIES, Nature, Lond., 264 : 373-376 (1976). 16. S. COCKCROFT and S. GOMPERTS, Nature, Lond., 314 : 534-536 (1985). 17. C. DUNIEC, J. ROBAK and R. GRIYGLEWSKI, Biochem. Pharmacol. 32 : 2283-2286 (1983). 18. J.R. VANE, Nature, Loncl., 231 : 232-235 (1971). 19. g. HAMBERG, Biochem. Biophys. Acta 431:651-654 (1976). 20. G.D. BARNES, B.I. BROWN, T.G. GARD, D. ATKINSON and R.P. EKINS, Mol. Cell. Enclocrinol. 12 : 273-284 (1978). 21. Z. NAOR, G. SNYDER, C.P. FAWCETT and SM. McCANN, Endocrinology 106 : 1304-1310 (1980). 22. P. ONALI, J. CRONIN and R.M. MACLEOD, Proc. Natl. Acad. Sci. USA 78 : 6531-6534 (1981). 23. L. SWENNEN and C. DENEF, Endocrinology 111 : 398-405 (1982). 24. G. SCHETTINI, J. CRONIN and R.M MACLEOD, Endocrinology 112 : 1801-1807 (1983). 25. H. HIDAKA, M. INAGAKI, S. KAWAMOTO and Y. SASAKI, Biochemistry 23 : 5036-5041 (1984). 26. S. PILOTE, P. VALLERAND, P. BORGEAT, Biochem. Biophys. Res. Comm. 104 : 867-873 (1982). 27. G.D. SNYDER, J. CAPDEVILA, N. CHACOS, S. MANNA and J:R. FALCK, Proc. Natl. Acad. Sci. USA 80 : 3504-3507 (1983). 28. P. NEEDLEMAN, J. TURK, B.A. JAKSCHIK, A.R. MORRISON and J.B. LEFKOWITH, Ann. Rev. e i o c h e m . 5 5 : 69-102 (1986). 29. K. KOIKE, A.M. JUDD, I.S. LOGIN, T. YASUMOTO and R.M..ACLEOD, Neumendocrinology43 : 283-290 (1986). 30. P.L. CANONICO, A.M. JUDD, K. KOIKE, C.A. VALDENEGRO and R.M. MACLEOD, Endocrinology 116 : 218-225 (1985). 31. A.M. CAMORATTOand L. GRANDISON,Endocrinology116 : 1506-1513(1985). 32. P.L. CANONICO,C.A. VALDENEGRO, R.M. MACLEOD,S.B. O'DELLand C.T. HARCUS, Endocrinology113 : 7-14 (1983). 33. C.A. SUTTONand T.F.J. MARTIN, Endocrinology110 : 1273-1280(1982). 34. P. SCHREY,H.J. CLARKand S. FRANKS,J. Endocr. 108 : 423-429(1986). 35. A. MALGAROLI,L. VALLAR, F. R. ELAHI,T. POZZAN,A. SPADAand J. MELDOLESI,J. Biol. Chem. 262 : 13920-13927(1987). 36. A.M. JUDD, K. KOIKEand R.M. MACLEOD,Am. J. Physiol.250 : E288-E295(1986). 37. M.J. BERRIDGE,R.F. IRVINE,Nature,Lond., 312 : 315-321 (1984).
Pergamon Press
CONTRIBUTION OF ARACHIDONATE METABOLITES TO BASAL AND THYROTROPIN RELEASING-HORMONE-STIMULATED RELEASE OF PROLACTIN FROM PURIFIED LACTOTROPHS IN PRIMARY CULTURE Marie-Pierre JUNIER, Jean-Maro ISRAEL*, Fernand DRAY*°, Jean-Didier VINCENT* Oregon Regional Primate Research Center, Div. of Neurosciences, Beaverton, OR, 97006, USA, * INSERM U176, 33077 Bordeaux Cedex, France, and ** INSERM U207, Institut Pasteur, 75015 Paris, France (Recelved in flnal form September I0, 1990) Summary
Among the different biochemical pathways which have been suggested to play a role in the control of prolactin (Prl) release from anterior pituitaries, arachidonate and its metabolites have been proposed to be involved in the process of Prl release. In this study we investigated the contribution of arachidonate metabolites to both basal and TRH-stimulated Prl release from perifused lactotrophs in culture (derived from pituitary glands of lactating female rats), which exhibit a high sustained release of Prl in absence of inhibitory input. Inhibition of the general oxidative metabolism of arachidonate by 10-5 M ETYA or of the arachidonate lipoxygenase metabolism by 10-5 M NDGA decreased basal Prl release to 45__.10% (n=3) and 36±4% (n=6) of the control release, respectively. Indomethacin, an inhibitor of the cyclooxygenase pathway, was without effect. Of the lipoxygenase metabolites tested at 10-6 M only 15-HPETE and 15-HETE induced Prl release. 15-HETE elicited prolactin release in a concentration dependent manner with a maximal effect at 10-6 M (10.72+3 ng/ml vs control 5.1±0.8 ng/ml, n=3). The quantity of Prl release induced by TRH was markedly decreased in the presence of NDGA. However, the fraction of Prl release elicited by TRH, calculated as a percentage of the amount of Prl released prior to TRH application, was similar under control conditions, and in the presence of NDGA. In contrast, inhibition of the protein kinases A and G by H8 (10-5 M) failed to alter basal Prl release but inhibited the effect of TRH by 58+6% (n=3). These data suggest that in absence of inhibitory inputs the high sustained release of Prl observed in cultures of lactotrophs derived from lactating female rats depends on the availability of lipoxygenase metabolites, and that the blockade of lipoxygenase reduces the absolute amount of Prl released by TRH without suppressing the ability of TRH to stimulate Prl release. The hormone release properties of highly enriched lactotrophs in culture, obtained from lactating female rats, have been characterized in correlation with their electrophysiological activity (1,2,3). They have been shown to be excitable, and to exhibit high levels of prolactin (Prl) release. The application of dopamine (DA), the main Prl releasing-inhibiting factor (4), induced a hyperpolarizing response, and inhibited Prl release. On the contrary, thyrotropin releasing-hormone (TRH), one of the best known hypothalamic releasing factor of Prl (5), induced a depolarization response characterized by a plateau potential (3), and stimulated Prl release. To further study the mechanisms underlying Prl release in this system we explored in the present study the involvement of arachidonate metabolites in basal and TRH-stimulated Prl release. Indeed, among the different biochemical mechanisms which have been proposed to be implicated in the control of Prl release, the products of membrane phospholipid metabolism have received increased attention. Noteworthy, different studies conducted either on cultures of antepituitary cells or on tumoral cell lines have suggested that, among those products, arachidonate and its metabolites are involved in the control of basal and stimulated Prl release (6,7,8). Moreover, recent studies have shown that arachidonate and its metabolites may modulate ionic channels activity in different biological systems (9,10,11,12). Corresponding author : M.P. Junier, ORPRC, 505 N.W. 185th Av., Beavedon, OR, 97006, USA 0024-3205/90 $3.00 +.00 Copyright (c) 1990 Pergamon Press plc
1830
Arachidonate and Prolactin Release
Vol. 47, No. 20, 1990
Using pharmacological agents which alter arachidonate metabolism, we show that basal Prl release depends on the availability of lipoxygenase arachidonate metabolites, while TRH remains able to elicit Prl release despite suppression of arachidonic acid metabolism. Material and Metho~,~ Animals Anterior pituitary cells were obtained from lactating female rats of the Wistar strain during the last week of lactation. Cell culture Cell dissociation and preparation of the cells for culture have been described in detail previously (13). Briefly, after enzymatic (trypsin 0.5% for 15 min.) and mechanical dissociation of the pituitary tissues, the cells were separated on a continuous gradient of 0.3-2.4% bovin serum albumin and collected in 100 ml fractions numbered 2-9 after rejection of the first 300 ml which contained red blood cells and cellular debris. Ninety five % of the cells collected in fractions 3 to 5 were lactotrophs (2). These cells were harvested, resuspended in Dubbelco's Modified Eagle's Medium (DMEM) containing 10% newborn calf serum, 3.7 g/I NaHCO3 at pH 7.4, and plated on 35 mm Petri dishes (Nunc, Denmark) at 106 cells per dish. Antibiotics (35 mg/I penicillin, 50 mg/I steptomycin) were present during the first 2 days of culture and thereafter omitted. The medium was replaced every 3 days. Prl release experiments After 7 days of culture, perifusion experiments were performed in the original Petri dishes. A continuous-flow perifusion system, which accomodated four dishes at a time, was used with a four channels peristaltic pump (2). Constant temperature (37 C) was maintained by a water heated platform. The perifusion medium consisted of DMEM supplemented with 15 mM Hepes and 12 mM NaHCO3. Flow rate was 0.5 ml.min-1 and medium levels remained at about 1.5 ml per dish. After a 90 min. stabilization period, 2 rain. samples were collected with a fraction collector into tubes containing 50 ml of 1% BSA in PBS. Aliquots were stored at -20 C until radioimmunoassayed. After a 20 min. collection period used to define basal Prl release, test compounds or their vehicle were directly added to the medium for various lengths of time, as indicated in the results section. The total perifusion period did not exceed 60-80 min. Radioimmunoassav The amount of Prl released into the medium was assayed using the RIA kit provided by the National Hormone and Pituitary Programm (Baltimore, MD, USA). Prl values were expressed as nanograms equivalents of rat standard Prl RP-1 per ml. Significant differences between groups were assessed with Student't test. A P of less than 5% was considered to indicate a significant difference. Results are presented as mean+sem. Material The hydroperoxyeicosatetraenoic acids (HPETEs, Sigma), the 15-hydroxyeicosatetraenoic acid (15-HETE, Sigma), and 5,8,11,14 eicosatetraenoic acid (ETYA, a gift from Hoffman La Roche, France) were dissolved in absolute ethanol under an argon blanket at concentrations of 1.5x10-4M, 1.6x10-4M, and 10-2M, respectively. 4-bromophenacylbromide (PBPB, Sigma) was dissolved in 0.1% ethanol at the concentration of 10-2 M. N(2-(methylamino)-ethyl)-5-isoquinoline sulfonamide (H8, Calbiochem, France) and nordihydroguaiaretic acid (NDGA, Sigma) were dissolved in dimethylsultoxide at the concentration of 10-2 M. Neomycin sulfate (Sigma) and indomethacin (Sigma) were dissolved in DMEM at the concentration of 10-2 M. All the drugs were aliquoted and stored either at -80 C (HPETEs, 15-HETE, ETYA) or at -20 C except indomethacin which was prepared extemporaneously. Final test concentrations were obtained by diluting the compound with DMEM. Neither ethanol nor DMSO at the concentrations used had an effect on Prl release. Results Arachidonate is released upon the action of either phospholipase A2, which cleaves the phospholipids sn-2-acyl bound, or upon the combined action of phospholipase C and diacylglycerol lipase on phosphatidylinositol (14). To investigate the role of arachidonate in the control of basal Prl
Vol. 47, No. 20, 1990
Arachidonate and Prolactin Release
1831
release, either PBPB, a compound which binds covalently to phospholipase A2 and blocks its activity (15), or neomycin, which inhibits phospholipase C by binding phosphatidylinositols (16), was administered at 10-5 M during a 30 min. perifusion period (figure 1). Both PBPB and neomycin strongly inhibited Prl release, the hormone levels decreasing to 49-+7% and 42+-9% (n=3) of basal release, respectively. Using pharmacological agents which alter arachidonate metabolic pathways we attempted to determine which of them was involved in the maintainance of Prl release in our system. The cells were exposed for 30 min to 10-5M of either ETYA, an inhibitor of the general oxidative metabolism of arachidonate (17), or indomethacin, an inhibitor of the cyclooxygenase pathway (18), or NDGA, an inhibitor of the lipoxygenase pathway (19). Both ETYA and NDGA reduced Prl release to 45+_10%, n=3, and 36-+4°/,,, n=6, of the control levels, respectively. Indomethacin was without effect (figure 2). In order to determine which lipoxygenase metabolite(s) is involved in the maintainance of basal Prl release, the three immediate products of lipoxygenase activity, 5-, 12- and 15-HPETE, were tested at the concentration of 10-6 M. Only 15-HPETE stimulated Prl release (figure 3A). 15-HETE, the direct metabolic product of 15-HPETE, induced a dose-related increase in Prl release with a maximal and statistically significant effect at 10-6 M (10.72:1:1.53 ng/ml vs control 5.56+_0.57 ng/ml, n=4, figure 3B), while 5- and 12-HETE were without effect (data not shown). To examine the involvement of lipoxygenase metabolites in the stimulatory effect of TRH on Prl release, TRH was added to the perifusion medium in the presence of NDGA. Like basal Prl release the absolute amount of Prl released by 10-7 M TRH was markedly reduced in the presence of NDGA, as compared to the effect of 10-7 M TRH alone (figure 4A). However, NDGA did not suppress completely the sfimulatory effect of TRH. Indeed the amount of Prl released during the 10 min. preceeding TRH application doubled in the presence of TRH (figure 4A). To estimate to which extent TRH remained able to elicit Prl release despite the blockade of lipoxygenase, we expressed the amount of Prl release as a percentage of Prl release during the 10 min. preceeding TRH application, i.e. when NDGA exerts its maximal inhibitory effect. The transformed data are presented in figure 4B; the maximal Prl release induced by TRH in the presence of NDGA represented 83__.4%(n=3) of the maximal Prl release induced by TRH in the absence of NDGA. Indomethacin was either without effect or induced a slight potentiation of TRH stimulatory effect (data not shown). For comparison purposes we tested the effect of cyclic nucleotides-dependent protein kinases inhibitors on the TRH-stimulatory effect on Prl release. Indeed, the involvement of cyclic nucleotides in Prl release has been demonstrated (20,21,22,23,24). Using H8, a preferential inhibitor of protein kinases A and G (25), we observed that, contrary to NDGA, H8 (10-5 M) did not affect basal Prl release, but inhibited TRH effect by 58+6% (n=3) (figure 5). Discussion Arachidonate and its lipoxygenase and epoxygenase metabolites are synthesized in the pituitary gland (26,27), and have been suggested to be involved in the basal and stimulated Prl release from antepituitaries (7). In this study we show that lipoxygenase metabolites of arachidonic acid are involved in the maintainance of basal Prl release from lactotrophs in culture, while TRH appears to be able to elicit Prl release in absence of a functional lipoxygenase. Lactating female rats were chosen as donors since their pituitaries yield four times more lactotrophs than others donors. The use of highly purified lactotrophs ensures that the effects observed indeed result from a direct action of the test compounds on lactotrophs. Blockade of the general oxidative metabolism of arachidonate or selective inhibition of lipoxygenase activity, which initiates one of the three known metabolic pathways of arachidonate (28), resulted in the suppression of basal Prl release. In contrast, inhibition of cyclooxygenase activity did not alter Prl release. The specificity of 10 p.M NDGA towards lipoxygenase in cultures of dissociated pituitary cells has been previously shown (29). These findings are in agreement with the observations of others on different models including hemipituitary glands and cloned strain of pituitary cells (30, 8).
1832
Arachldonate and Prolactln Release
A
Vol. 47, No. 20, 1990
NEOMYCPE 12 10-
8 E
==
ii 6'
4-
2-
1o 9
.
. 2O
.
. 30
.
. 40
.
. 50
. 60
i ..... | ..... , ..... 70 80 90 100
Time (min)
PBPB
10
8 7
~'-
..... i . 10
4
.
0
....
0
10
20
30
40
50
60
70
Time ( r a i n )
80
..... i ..... i ..... i 10 20 30
.
.
.
.
.
.
.
.
40
.
.
.
.
.
.
.
.
50
.
.
.
.
60
i ..... i ..... i ..... 70 80 90 100
Time (min)
FIG• 1
FIG. 2
Effects of 10 p.M neomycin (A) or PBPB (B) on spont-aneous Prl release during a 30min and 20 min perifusion period, respectively• Note the rebound of Prl release after neomycin withdrawal. These curves are representative of 3 experiments•
Effects of 101~1ETYA( [] ), NDGA (z~), or indomethacin ( • ) on spontaneous Prl release during a 30 min application period. These curves are representative of 3 (ETYA, indomethacin) to 6 experiments (NDGA).
Prior to its metabolization, arachidonate is liberated from its storage sites in the membrane phospholipids. This release can result either from the action of phospholipase A2, which cleaves the acyl bond between the phospholipid and the fatty acid, or from the combined action of phospholipase C and diacylglycerol lipase on phosphatidylinositol. To determine the source of arachidonate in our system, we examined the effect of specific inhibitors of each phospholipase• In both cases we observed inhibition of Prl release• This result suggests that either both phospholipases A2 and C may be involved in the mobilization of arachidonate in our system or that both products of phosphatidylinositol metabolism and arachidonate melabolism are implicated in the control of basal Prl release. The inhibitory effect of phospholipase A2 blockade on Prl release corroborates previous studies (30, 31, 6). To our knowledge,
Vol. 47, No. 20, 1990
Arachidonate and Prolactln Release
1833
14 12 10
i
.... • • • • • • 0
10
20
30
40
50
Time (rain)
B
ip,o oo,i 13
9.75
CL
6.5
"E r~ 3.25
0 C 10-SM
C 10-TM
C 10-6M
FIG. 3 A: Effect of 1 ldVl of 5- ( o ), 12- ( • ), and 15-HPETE ( • ) on Prl release during a 10 min application period. These curves are representative of 3 experiments. B: Effect of increasing concentrations of 15-HETE on Prl release (C=control, mean+sem, 15-HETE 10-8M, 10-7M, n=3, 15-HETE 10-6M, n=6). no study has been reported concerning the consequences of phospholipase C inhibition on Prl release. However, phosphatidylinositol turnover has been shown to be altered by DA and TRH in a manner that correlates well with the changes in Prl release (32,33). Among the different lipoxygenase metabolites tested 15-HETE, and its precursor 15-HPETE stimulated Pd release in our system. In contrast, 12-HPETE and 5-HPETE, the two other main precursors of arachidonate lipoxygenase metabolites, were ineffective. A stimulatory effect of 15-HETE on Prl
1834
Arachidonate and Prolactin Release
N D G A ~ -TRH"
A 20 18
B
13_
-TRH-
350
==
16 14 crJ t-
Vol. 47, No. 20, 1990
300 250 -
12 10 8
u) (1= .J~ O
6 4
EL
200 150
,°°t, -
2 0 10
20
30
40
50
60
70
80
Time (min)
40
50 60 70 Time (min)
80
FiG. 4 A: Effect of 10 I~M NDGA on 0.1 ~M TRH-elicited Prl release ( A ) compared to the effect of TRH alone( • ). B: Prl release expressed as % of control release in the presence of NDGA ( ,~ ) or in the absence of
NDGA ( • ) . These curves are representative of 3 experiments.
b
H8
-
-
TRH
30 25
FIG. 5
20 Co c
Effect of 10 tIM H8 on 0.1 I~M TRH-eficited Prl release ( o ) compared to the effect of TRH alone( • ). These curves are representative of 3 experiments.
15
n
t05"
0
•
10
20
30
40
50 60 Time (min)
!
70
•
!
80
•
i
•
90 100
release has also been reported on GH3 cells (8), and on static incubations of dispersed pituitary cells (7). However in this last study 15-HETE stimulated Prl release only at 5 ~M, while 5-HETE efficient concentrations ranged from 1 to 5 MM. 5-HETE being the most unstable of the HETEs, its eventual degradation in our perifusion system shall be envisaged. Different cells donors (non lactating female rats in the cited study) and cell preparation (enriched lactotrophs in this study) could also account for this discrepancy.
Vol. 47, No. 20, 1990
Arachldonate and Prolactln Release
1835
It is interesting to note that recent results have shown the presence of Ca2+ channels active at the resting potential on the lactotrophs (P.M. Lliedo and J.M. Israel, personnal communication). Moreover, in the same model and under similar experimental conditions, the inhibition of Prl release in the presence of DA, which is known to act in a Ca2+-dependent manner (34,35), is analogous to that induced by blockade of lipoxygenase activity (1). This and the previous demonstrations of the ability of arachidonate metabolites to activate ionic channels in neural and cardiac cells (9, 10, 11, 12), suggest a possible link at the molecular level between arachidonate metabolites and the electrical activity of the lactotrophs. Further experiments are required to test this hypothesis. In addition to alter basal Prl release, lipoxygenase metabolites of arachidonic acid have been suggested among other compounds to mediate the stimulatory effect of TRH on Prl release (30). In this study the inhibition of lipoxygenase activity by NDGA induced a strong decrease of the absolute amount of Prl released by TRH. However, considering the relative amount of Pd released by TRH, as compared to the amount of Prl released prior to TRH application (as illustrated in figure 4B), we noticed that the fraction of Prl release elicited by TRH was not markedly altered in the presence of arachidonate metabolism inhibitors. In contrast H8, a preferential inhibitor of protein kinases A and G, failed to alter basal Prl release but strongly reduced the amount of Prl released by TRH. These results show that, although the absolute amount of Prl release triggered by TRH depends from the quantity of basal Prl release, the basal and TRH-stimulated Prl release may be manipulated independently. This suggests that TRH may use metabolic pathways different from the ones involved in the maintainance of basal Prl release. These data do not exclude an involvement of arachidonate metabolites in the TRH effect under normal conditions, especiaity since TRH has been shown to stimulate arachidonate release (7, 36). However these results suggest that more than one metabolic pathway may be used by TRH to elicit Prl release. One of these pathways could involved the cyclic nucleotides-dependent protein kinases A and G as suggested by the inhibitory effect of H8 on TRH-eliclted Prl release. H8 has been found to inhibit both cyclic nucleotide-dependent protein kinases and protein kinase C. However, the H8 concentration required to inhibit protein kinases A and G is far lower than the one needed to inhibit protein kinase C (apparent Ki of 0.48 and 1.2 p.M versus 15 p.M lor protein kinase C), (25). An eventual effect of H8 on protein kinase C can not be totally excluded. But the fact that, on one hand the inhibition of phospholipase C activity, which leads to the formation of a well known stimulator of protein kinase C i.e. diacylglycerol (37), decreased basal Prl release, while on the other hand H8 was ineffective on basal Prl release, indicates that protein kinase C activity was not affected by H8. In conclusion, our results suggest that a) in absence of inhibitory inputs the high sustained release of Prl observed in cultures of lactotrophs derived from lactating female rats depends on the availability of lipoxygenase metabolites, b) among the lipoxygenase metabolites 15-HETE and its precursor 15-HPETE are stimulators of Prl release, c) the blockade of lipoxygenase reduces the absolute amount of Prl released by TRH but does not suppress the ability of TRH to stimulate Prl release, and d) protein kinases A and/or G may be involved in the stimulatory effect of TRH on Prl release. Acknowledaement We thank D. Verrier lor her excellent technical assistance. References 1. 2. 3. 4. 5.
C. KIRK, These de doctorat de 3eme cycle, Universite de Bordeaux II, France (1986). J.M. ISRAEL, C. KIRK and J.D. VINCENT, J. Physiol. 390 : 1-22 (1987). J.M. ISRAEL, A.P. KUKSTAS and J.D. VINCENT, Neuroendocrinology, in press. R.M. MACLEOD, Frontiers in Neumendocrinoloay, p.169-174, Raven Press, New-York (1974). Y. KATO, N. MATSUSHITA, H. OHTA, K. TOJO, A. SHIMATSU and H. IMURA, The Pitultaw Gland, p. 261-278, Raven Press, New-York (1985). 6. L. GRANDISON, Endocrinology 114 : 1-7 (1984). 7. A.M. JUDD, B.L. SPANGELO, J.T. EHRETH and R.M. MACLEOD, Neuroendocrinology48:407-416
(1988). 8. M. RABIER, C. CHAVIS, A. CRASTES DE PAULET and M. DAMON, Neuroendocrinology47 : 323-328 (1988).
1836
Arachldonate and Prolactln Release
Vol. 47, No. 20, 1990
9. D. PIOMELLI, A. VOLTERRA, N. DALE, S.A. SlEGELBAUM, E.R. KANDEL, J.H. SCHWARTZ and F. BELARDETTI, Nature, Lond., 328 : 38-43 (1987). 10. D. KIM and D.E. CLAPHAM, Science 244 : 1174-1176 (1989). 11. Y. KURACHI, H. ITO, T. SUGIMOTO, T. SHIMIZU, I. MIKI and M. UI, Nature, Lond., 337 : 555-557 (1989). 12. R.W. ORDWAY, J.V. WALSH and J.J. SINGER, Science 244 : 1176-1179 (1989). 13. J.M. ISRAEL, C. DENEF and J.D. VINCENT,Neuroendocrinology37 : 193-199 (1983). 14. M.B. FEINSTEIN and R.I. SHA'AFI, Calcium and Cell Function. p. 337-376, Academic Pess, Newoyork (1983). 15. J. DRENTH, C.M. ENZING, K.H. HALK and J.C. VESSIES, Nature, Lond., 264 : 373-376 (1976). 16. S. COCKCROFT and S. GOMPERTS, Nature, Lond., 314 : 534-536 (1985). 17. C. DUNIEC, J. ROBAK and R. GRIYGLEWSKI, Biochem. Pharmacol. 32 : 2283-2286 (1983). 18. J.R. VANE, Nature, Loncl., 231 : 232-235 (1971). 19. g. HAMBERG, Biochem. Biophys. Acta 431:651-654 (1976). 20. G.D. BARNES, B.I. BROWN, T.G. GARD, D. ATKINSON and R.P. EKINS, Mol. Cell. Enclocrinol. 12 : 273-284 (1978). 21. Z. NAOR, G. SNYDER, C.P. FAWCETT and SM. McCANN, Endocrinology 106 : 1304-1310 (1980). 22. P. ONALI, J. CRONIN and R.M. MACLEOD, Proc. Natl. Acad. Sci. USA 78 : 6531-6534 (1981). 23. L. SWENNEN and C. DENEF, Endocrinology 111 : 398-405 (1982). 24. G. SCHETTINI, J. CRONIN and R.M MACLEOD, Endocrinology 112 : 1801-1807 (1983). 25. H. HIDAKA, M. INAGAKI, S. KAWAMOTO and Y. SASAKI, Biochemistry 23 : 5036-5041 (1984). 26. S. PILOTE, P. VALLERAND, P. BORGEAT, Biochem. Biophys. Res. Comm. 104 : 867-873 (1982). 27. G.D. SNYDER, J. CAPDEVILA, N. CHACOS, S. MANNA and J:R. FALCK, Proc. Natl. Acad. Sci. USA 80 : 3504-3507 (1983). 28. P. NEEDLEMAN, J. TURK, B.A. JAKSCHIK, A.R. MORRISON and J.B. LEFKOWITH, Ann. Rev. e i o c h e m . 5 5 : 69-102 (1986). 29. K. KOIKE, A.M. JUDD, I.S. LOGIN, T. YASUMOTO and R.M..ACLEOD, Neumendocrinology43 : 283-290 (1986). 30. P.L. CANONICO, A.M. JUDD, K. KOIKE, C.A. VALDENEGRO and R.M. MACLEOD, Endocrinology 116 : 218-225 (1985). 31. A.M. CAMORATTOand L. GRANDISON,Endocrinology116 : 1506-1513(1985). 32. P.L. CANONICO,C.A. VALDENEGRO, R.M. MACLEOD,S.B. O'DELLand C.T. HARCUS, Endocrinology113 : 7-14 (1983). 33. C.A. SUTTONand T.F.J. MARTIN, Endocrinology110 : 1273-1280(1982). 34. P. SCHREY,H.J. CLARKand S. FRANKS,J. Endocr. 108 : 423-429(1986). 35. A. MALGAROLI,L. VALLAR, F. R. ELAHI,T. POZZAN,A. SPADAand J. MELDOLESI,J. Biol. Chem. 262 : 13920-13927(1987). 36. A.M. JUDD, K. KOIKEand R.M. MACLEOD,Am. J. Physiol.250 : E288-E295(1986). 37. M.J. BERRIDGE,R.F. IRVINE,Nature,Lond., 312 : 315-321 (1984).