Activity of prostaglandin biosynthetic pathways in rat pancreatic islets

PROSTAGLANDINS ACTIVITY OF PROSTAGLANDIN BIOSYNTHETIC PATHWAYS IN RAT PANCREATIC ISLETS

Kathleen L. Kellyl and Suzanne G. Laychock2

Department of Pharmacology, Virginia, U.S.A. 23298

Medical

College

of Virginia,

Richmond,

ABSTRACT Isolated pancreatic islets of the rat were either prelabeled with [3H]arachidonic acid, or were incubated over the short term with the concomitant addition of radiolabeled arachidonic acid and a stimulatory concentration of glucose (17mM) for prostaglandin (PG) analysis. In prelabeled islets, radiolabel in 6-keto-PGFl~ , PGE2, and 15-keto-13,14-dihydro-PGF2~ increased in response to a 5 min glucose (17mM) challenge. In islets not prelabeled with arachidonic acid, label incorporation in 6-keto-PGFl~ increased, whereas label in PGE 2 decreased during a 5 min glucose stimulation; after 30-45 min of glucose stimulation labeled PGE levels increased compared to control (2.8mM glucose) levels. Enhanced labelling of PGF2~ was not detected in glucose-stimulated islets prelabeled or not. Isotope dilution with endogenous arachidonic acid probably occurs early in the stimulus response in islets not prelabeled. D-Galactose (17mM) or 2-deoxy-glucose (17mM) did not alter PG production. Indomethacin inhibited islet PG turnover and potentiated glucose-stimulated insulin release. Islets also converted the endoperoxide [3H]PGH2 to 6-keto-PGFl~ , PGF2~ , PGE 2 and PGD 2, in a tlme-dependent manner and in proportions similar to arachidonic acid-derived PGs. In dispersed islet cells, the calcium ionophore ionomycin, but not glucose, enhanced the production of labeled PGs from arachidonic acid. Insulin release paralled PG production in dispersed cells, however, indomethacin did not inhibit ionomycinstimulated insulin release, suggesting that PG synthesis was not required for secretion. In confirmation of islet PGI 2 turnover indicated by 6-keto-PGFle production, islet cell PGl2-1ike products inhibited platelet aggregation induced by ADP. These results suggest that biosynthesis of specific PGs early in the glucose secretion response may play a modulatory role in islet hormone secretion, and that different pools of cellular arachidonic acid may contribute to PG biosynthesis in the microenvironment of the islet.

iPortions of this work were performed in partial fulfillment of the requirements for the Degree of Doctor of Philosophy. 2To whom to address reprint requests. This work was supported by NIH grant AM 25705.

JUNE 1984 VOL. 27 NO. 6

925

PROSTAGLANDINS INTRODUCTION Although prostaglandin (PG) biosynthesis in pancreatic islets has been demonstrated (1,2,3,4) and PGs have been implicated in the regulation of hormone secretion from this organ, controversy exists concerning the modulatory action of PGs in endocrine function. Prostacyclin (PGI2) , and its degradation product 6-keto-PGFl~ , and PGEI which appears to act on the same receptors as PGI2 (5), have been shown to stimulate insulin, glucagon and somatostatin secretion from rat pancreatic islet cells (6,7,8,9,10). In addition, exogenous PGE 2 or PGF2~ stimulate the release of pancreatic islet hormones (11,12,13,10); the endoperoxide PGH 2 also stimulates islet hormone release (14). And yet, since alpha-adrenergic stimulation enhances PGE synthesis during inhibition of insulin secretion, and inhibition of PG biosynthesis is accompanied by enhanced insulin release, the argument has been made that endogenous PGs are inhibitory modulators of islet hormone release (4,15,16). Recently glucose was reported to enhance arachidonic acid turnover in rat islet phospholipids and PGs (2,17,18). However, characterization of PG biosynthesis during the early phase of insulin secretion has not been investigated. Since individual PGs may possess stimulatory or inhibitory actions on secretory processes, it would be valuable to know the relative turnover rates of the parent PGs during secretagogue stimulation of islets. The purpose of the present study is to define the biosynthetic profile and the time course of biosynthesis of islet PGs in challenged islets. The implications of selective PG turnover in islets during stimulation of secretion is discussed. METHODS Isolation of Islets. Pancreata from decapitated male SpragueDawley rats (175-200 g) were excised, minced, disrupted with collagenase and the islets isolated as described previously (17). Groups of 100-400 islets were incubated at 37°C under 02/C02 (95:5) in Krebs Ringer bicarbonate buffer pH 7.4, 0.5 ml, containing 16 mM HEPES, 0.01% fatty acid free BSA (Sigma), and 2.8 mM glucose (KRBH buffer). Following a 30 min preincubation, the islet incubation medium was replaced with 0.5 ml fresh oxygenated KRBH buffer which was supplemented with glucose up to 17 mM for glucose-stimulated samples, and either [l~C]arachidonic acid (0.I ~Ci, 52.2mCi/mmol) (Amersham), [3H]arachidonic acid (0.i ~Ci, 78.2 Ci/mmol) (New England Nuclear), or [3H]PGH 2 0.i ~ Ci, 1.0 uCi/Ng), and the incubation was continued for the times indicated. Islet incubation medium was either extracted immediately following the incubations for PG analysis, or was analyzed for insulin content by radioimmunoassay (17). In certain experiments islets were prelabeled for 90 min with [3H]arachidonic acid (2 ~Ci/ml) in KRBH buffer containing 8.5 mM glucose. The islets were then washed twice over a 30 minute period with KRBH buffer containing 2.8 mM glucose and

926

JUNE 1984 VOL. 27 NO. 6

PROSTAGLANDINS lacking radiolabel. Following the last wash period, islets were resuspended in KRBH buffer containing i ~Ci/ml [3H]arachidonic acid and the absence or presence of 17 mM glucose, and were allowed to incubate for 5 min for determination of PG production as described above. Dispersed islet cells were prepared by treating islets with 0.2% trypsin (Worthington Biochemicals) and 5 mM EGTA, and vortexing at room temperature for 2 min to disrupt the islet capsule. Trypsin inhibitor (soybean, Sigma) (0.4%) was added to the digestate and the cells centrifuged at 500 rpm for 7 min through 4 ml of KRBH buffer containing 4% BSA. The cell pellet was resuspended in KRBH buffer and the cells counted in a haemocytometer. Cell viability was 95% using trypan blue dye exclusion analysis. Visual inspection showed that the cell suspension contained less than 5% exocrine acinar cells and no identifiable vascular elements which remained at the BSA interface during centrifugatlon. Synthesis of Prostaglandin Endoperoxide. The synthesis of [3H]PGH 2 was carried out using ram seminal vesicle microsomes (Pel Freeze) and [3H-8]arachidonic acid (78.2 Ci/mmol) according to the method of Hamburg and Samuelsson and coworkers (19,20). The specific activity of the PGH 2 was 1.0 ~ C i / ~ g arachidonic acid. [3H]PGH2 (0.01 DCi) was reduced in 0.2 ml diethyl ether containing i mg triphenyl phosphine, and 95% of the radiolabel was subsequently recovered as PGF2~ , thus confirming the identity of PGH 2. Prosta$1andin Isolation and Identification. Islet incubations were stopped by addition of i ml cold buffer (4°C), centrifuged at 500 rpm for 2 min, and the supernatant removed. Supernatants were acidified to pH 3 with formic acid and extracted with ethyl acetate as described previously (3). PG extracts were analyzed by high-performance liquid chromatography (HPLC) silicic acid and reversed phase fatty acid systems as described elsewhere (21) and/or by thin layer chromatography (TLC) on silica gel G plates developed sequentially in (A) chloroform:acetic acid (90:l,v/v) and (B) ethyl acetate: water:isooctane:acetic acid (44:40:20:8,v/v/v/v) (22). Organic extracts were co-chromatographed with authentic standards and identified by iodine staining; PG zones were scraped into vials and counted by liquid scintillation spectrometry to 1.5% error. In all experiments, blank values from samples containing radioactivity but lacking tissue during incubation and contributing background levels of radioactivity in the chromatographic preparations were subtracted prior to quantitation and anlysls of the results. Recovery of PGs was 70%. Bioassay of Prostacyclin. Platelet-rich plasma (PRP) was prepared from New Zealand white rabbits by the method of Siller et al. (23). PRP (0.45mi) was preincubated in a dual-channel aggregometer (Payton Assoc.) for 3 min before stimulation with ADP (10-5M) or an aliquot of KRB buffer with dispersed islet cells. The PGI 2 assay of Bunting et al. (24) was modified such that rat islet cells (5 x 105 cells/ml) in 0.5 ml KRB buffer, pH 7.4, were incubated at

JUNE 1984 VOL. 27 NO. 6

927

PROSTAGLANDINS 37°C in a shaking water bath with i ~g arachidonic acid for 3 min. A 0.05 ml aliquot of incubation medium was added to 0.45 ml PRP and incubated for 1 min prior to ADP (10-SM) stimulation. Control PRP received the same buffer volume containing araehidonate and ADP. Statistical Analysis. Values shown are mean + S.E. Statistical analyses was performed using Students t-test for--paired or unpaired values of the geometric mean. Probability values < 0.05 were accepted as significant. RESULTS Time Course of PG Biosynthesis. Isolated islets incubated with radiolabeled arachidonic acid and glucose (17 mM) sufficient to evoke maximum insulin release responded with enhanced turnover of PGs. During the first 5 min of glucose stimulation an increase in labeled arachidonic acid turnover in PGI 2 was evidenced by an increase in radiolabel recovered in the product 6-keto-PGFl~ (Fig. i). During the same period, however, radiolabel incorporated into PGE 2 was decreased in glucose-stimulated islets and there was no significant change noted in radiolabeled PGF2~ levels compared to unstimulated islet labeled PG levels (Fig. I). Insulin release from islets exposed to glucose (17 mM) for 5 min was increased by 94 (+ 22)% (p < 0.05) above insulin levels in paired unstimulated islet incubations.

25C

=,., .9

20(

°15o Faw

~0c

E

I/

"K

(i5)

-I'-

E:

0

x

5C I----5 min

I

~---30-45min.---I

Fig. i. Radiolabeled arachidonic acid incorporation into prostaglandins. Isolated islets were incubated in 0.5ml KRBH buffer containing 0.4 HCi [3HI- or [14C]arachidonic acid and either 2.8 mM glucose (control) or 17 mM glucose for 5 or 30-45 min. Organic extracts of islet incubation media were analyzed for PGE 2 (E2), PGF2~ (F2~) and 6-keto-PGFl~ (6KFI~) by TLC. The data are represented as percent of control (~ S.E.M.) for glucose-stimulated values. Significance levels *p< 0.05 using Student's t-test for paired values. Number of independent experimental determinations are in parentheses.

928

JUNE 1984 VOL. 27 NO. 6

PROSTAGLANDINS Unlike the profile of PG turnover in islets incubated with glucose for 5 min, after 30-45 min glucose stimulation enhanced the incorporation of labeled arachidonic acid into PGE2 and 6-keto-PGFld (Fig. i). Since the percentage increase of radiolabel in 6-keto-PGFle during 30-45 min of incubation was only slightly greater than that which occurred during 5 min, the effect of glucose on the incorporation of labeled arachidonic acid into this PG occurred largely during the onset of stimulation. On the other hand, the turnover of PGF 2 in islets appeared unaltered by glucose stimulation (Fig. i). Incubation of islets with 2-deoxyglucose (17mM) and radiolabeled arachidonic acid for 45 min failed to significantly alter labeled PGE 2 (ii0 + 12%), PGF2~ (108 + 5%) or 6-keto-PGFl~ (114 + 6%) levels compare~t0 paired control ~slet PG values (n=4). Lik--ewise, D-galactose (17mM) did not alter the production of PGE 2 (I00 + 14%), PGF2~ (I01 + 7%), or 6-keto-PGF~ (103 + 10%) compared to "untreated control is-let PG levels during 45 min of incubation (n=3). Insulin release from islets after 30 min and 60 min of incubation with glucose (17 mM) was 281(+ 59)% and 391(~ 69)% (p < 0.01), respectively, above insulin levels in paired unstimulated islet incubations. The incorporation of [14C]arachidonlc acid in islet PGs occurred most rapidly during the first 5 min of incubation, and thereafter the cyclooxygenase catalyzed conversion of [14C]arachidonic acid into PGs was time-dependent but of modest proportion (Fig. 2). Basal PG production in control islets also increased rapidly within 5 min after exposure to labeled fatty acid as evidenced by the values in Fig. 2 for PGE 2 and PGF2e which were not increased by the glucose challenge (Fig. i). PGE 2 turnover surpassed that of PGF2e or 6-keto-PGFl~ during 60 min of incubation (Fig. 2). When islets were preincubated with indomethacin (20 ~M) for 30 min prior to addition of radiolabeled arachldonic acid, total PG biosynthesis was inhibited by 91(+5)% after 60 min of islet exposure to radiolabel, which is i~ agreement with the findings of others (18). In addition, indomethacin-treated islets exposed to 8.5mM glucose for 60 mln released 92(+19)% (p< 0.05) more insulin than did islets exposed to glucose alon~; indomethacin did not significantly alter insulin release from islets incubated with 2.8mM glucose. Similar alterations in glucose-stimulated insulin release in response to cyclooxygenase inhibition have been reported by others (2,25). In order to determine if endogenous fatty acid pools contribute to the profile of PG production, islets were prelabeled with [3H]arachidonic acid and a stimulatory concentration of glucose (8.5 mM) for 90 min. Following a subsequent 30 min equilibration at 2.8 mM glucose, islets were treated in a manner identical to the aforementioned experiments with incubation for 5 min in the presence of labeled arachldonic acid and 17 mM glucose or 2.8 mM glucose (control). PG production in prelabeled islets differed in several respects from that observed in islets not prelabeled. Most

JUNE 1984 VOL. 27 NO. 6

929

PROSTAGLANDINS

5 m

~4 E2

~3 F2~

g~

i

O

Fig.

2.

5

i

I0

Time-course

i

i

i

i

i

i

i

i

15 20 25 30 35 ~ Time (rain)

of

i

i

45 50 55 60

prostaglandin

turnover

in

rat

islets.

Islets were incubated in ~ B H buffer containing [14C]arachidonic acid (0.5 ~Ci) and 17 ~ glucose for ~ to 60 min. Organic extracts of media were analyzed for PG content by TLC. Values are mean ~ S . E . M . ) dpm per 200 islets (n = 3-4). U) om

=,,®

180

T

>' 160 O--

m ~ 140 c c

o mo oo

I00

~

80

D_ ~"

60

(5)

6KF=

FZ=

E2

15K-H 2

F2=

0 Fig. 3. Prostaglandin production in prelabeled islets. Islets were prelabeled with [3H]arachidonic acid in KRBH buffer with 8.5 mM glucose for 90 min, and then allowed to re-equilibrate with 2.8 mM glucose for 30 min prior to stimulation with 17 mM glucose for 5 min for PG analysis. Control islets were treated identically except that 2.8 mM glucose was present during the final 5 min incubation. Organic extracts of islet incubations were analyzed for PGs (see legend fig. i); 15-keto-13,14-dihydro-PGF2a (15K-H2-F2~). Significance levels *p < 0.05, **p < 0.02. Number of independent experimental determinations are in parentheses.

930

JUNE 1984 VOL. 27 NO. 6

PROSTAGLANDINS 65 60

[]

5rnin

5~

[]

ZOmia

5O

~ m o

45

,o

~.

25

~

~'0

:%

5

~

,o

I

f 6K PGFI.

P GFZ,~

PGE z

PGD z

Fig. 4. Time-dependent metabolism of [3H]PGH2 to prostaglandlns. Isolated islets were incubated in 1 ml KRBH buffer containing [3H]PGH2 (0.01 ~Ci) and 2.8 - 17 mM glucose for 5 and 20 min intervals. Organic extracts of islet incubation medium were analyzed by HPLC and then TLC to determine conversion of [3H]PGH2 to PGs. Values are mean (~ S.E.M.) dpm per 200 islets recovered in PGs (n = 5). Significance levels **p < 0.05 using Student's t-test for unpaired values at 5 min.

notably, prelabeled islets responded to glucose stimulation with an increase in PGE 2 production (Fig. 3), and the increase in labeled PGE 2 (37%) approximated the decrease in label appearing in PGE 2 (33%) in islets not prelabeled (Fig. i). These data suggest that the decreased label in PGE 2 of islets not prelabeled with arachidonate was the result of isotope dilution by endogenous unlabeled arachidonic acid stores mobilized by glucose, In addition, the level of radiolabel in PGF2~ was not significantly altered by glucose stimulation in prelabeled islets, however, there was a significant increase in the metabollte migrating on TLC like 15-keto-13,14-dihydro-PGF2~ (Fig.3). Pretreatment with indomethacin (20 ~M) inhibited production of this metabolite by 92 + 8% of control. Endoperoxide Conversion to Islet PG. The endoperoxide PGH 2 is the immediate precursor of the prostaglandins subsequent to the cyclooxygenase step. The metabolism of exogenously administered [3H]PGH 2 therefore bypasses the rate limiting step in the synthesis of PGs, release of arachidonic acid from phospholipids. [3H]PGH2 was incubated with isolated islets for 5 or 20 min, and PGs were identified by HPLC and TLC. Figure 4 illustrates the tlme-dependent conversion of [3H]PGH 2 to 6-keto-PGFle , PGF2~ , PGE2, and PGD 2. In

JUNE 1984 VOL. 27 NO. 6

931

PROSTAGLANDINS

certain experiments glucose (17 mM) was added to the islets with [3H]PGH2, however, glucose did not have a significant quantitative effect on PG biosynthesis. The conversion of substrate to PGs was much greater in these preparations using [3H]PGH 2 (approximately 60% of label recovered as PGs) as compared to those experiments which employed radiolabeled arachidonic acid (approximately 1% of label recovered as PGs), but the relative profile of PGs synthesized by the islets w a s the same for both substrates. After 5 min, most of the [3H]PGH2 was recovered as PGF2~ (47 + 1%) and PGE 2 (30 + 1%), followed by PGD2 (22 + 1%) and 6-keto-P~F I (3 + 0.4%). After 20 min, PGE2 and PGF2~ still constitute the major PG~ synthesized, however, there is relatively more PGE 2 (44 + 4% of radiolabel incorporated into PGs) than PGF2~ (37 + 5%), P~D2 (18 + 1%), or 6-keto-PGFl~ (I + 0.1%). PGD2 was identified in these ~reparations by a combination of HPLC and TLC techniques as distinct from PGE 2 or the metabolite 15-keto-13,14-dihydro-PGF2~. Prosta$1andin Production by Dispersed Islet Cells. When cells from disrupted rat islets were incubated with [3H]arachidonic acid PG biosynthesis occurred in similar proportions to those observed in intact islets (Fig. 5). However, unlike intact islets, dispersed cell PG biosynthesis was not stimulated by glucose (Fig. 5). On the other hand, the cation ionophore ionomycin stimulated arachidonate conversion to 6-keto-PGFl~ and PGF2~ (Fig. 5). Insulin release from dispersed islet cells paralleled the responsivity Of the prostaglandin biosynthetic pathway. Insulin release from cells exposed to 17 mM glucose was not significantly different from unstimulated cells (94 ! 20% of control), whereas i ~ M ionomycin enhanced insulin release 2-fold (195 ~ 13% of control, p < 0.01) after 45 min of incubation. However, insulin release from islets pretreated with indomethacin (20 ~M) for 30 min prior to incubation with ionomycin for 45 min was not significantly different (18 + 16%, n=8) from release in control islets treated with ionomycin. Bioassay of the products synthesized by islet cells was carried out in order to further characterize the production of PGI 2. Dispersed cells incubated with arachidonic acid for 3 min were added to platelets aggregated by ADP. Prostacyclln-like products produced by islet cells inhibited ADP-induced platelet aggregation by 26% which corresponded to 3.2 ng PGI2/5 x 105 ceils (Fig. 6). DISCUSSION We have shown previously that homogenates of rat islets synthesize the parent prostaglandins PGE2, PGF2~ , and PGI 2. The present studies demonstrate the time-dependent biosynthesis of PGs in intact islets, and further support a role for PGs in islet endocrine function since selective incorporation of arachidonic acid into 6-keto-PGFl~ and PGE 2 is stimulated by the insulin secretagogue glucose. Increases in the 15-keto-13,14-dihydro-PGF2~

932

JUNE 1984 VOL. 27 NO. 6

PROSTAGLANDINS

I 8

?,'. ..', ',,..

¢

T

_gE4

o"o

// //

~'.'I

//:-~..,

,

'.:/l

C

(*) ,~e)i(~)]J C

G

I

I- 6 K pG FI=.--~

C

G

C

I

G

I

: ' ~ PG E2---~

t,-.,-- pG F 2 ~

Fig. 5. Dispersed rat islet cell PG biosynthesis. Dispersed cells (1.5 x 105 cells/ml) were incubated in i ml KRBH buffer with [3H]arachidonic acid (0.i ~ Ci) and control (C) glucose (2.8 mM), high glucose (G) (17 mM) or ionomycin (I) (10-6M) for 45 min at 37°C° PGs were extracted from islet cell incubation media and identified by TLC. Values are mean (+ S.E.M.) dpm in PGs/I.5 x 105 cells/ml. Significance levels **p ~ 0.01 using Student's t-test. Numbers in parentheses represent independent determinations.

ADP

tOO =

o

._~

75

E

o 5C

-~

-FIslet

i+ pGglz

2rig

Cells

~

25

0

b2'o ;o~o 8'o ,do "" Time

; 2'o 4 do ;o,~o' (seconds)

Fig. 6. Inhibition of platelet aggregation by rat islet prostaglandins. Platelet-rich rabbit plasma was aggregated by ADP (10-5M), and aggregation was inhibited by 1 or 2 ng PGI2, or islet incubation medium (see Methods).

JUNE 1984 VOL. 27 NO. 6

933

PROSTAGLANDINS

metabolite level may indicate the rapid turnover and metabolism of PGF2~ , however, other enzymatic pathways such as the metabolism of PGE by 9-keto reductase (26) and 15-hydroxyprostanoate oxidoreductase (27) may contribute to the formation of PGF-type metabolites. Glucose stimulation may alter the production of cofactors or allosteric effectors which regulate enzyme activities associated with PG turnover (28). The release of 6-keto-PGFl~ from islets, as determined by radioimmunoassay, is both time-dependent and stimulated by glucose within i h (18). The present studies demonstrate that labeled arachidonic acid recovery in 6-keto-PGFla is elevated within 5 min after glucose stimulation. These data are consistent with the evidence that arachidonic acid turnover in specific islet cell phospholipids is stimulated within 5 min of glucose challenge ( 1 7 ) . Since a major part of the labeled arachidonic acid is incorporated into PGs during the first 5 or 30-45 min of stimulation when biphasic insulin secretion occurs, there may exist a close coupling of glucose recognition and PG biosynthesis. A similar early PG synthesis response to secretagogue stimulation was noted in adrenal cells in vitro (29). The data derived from islets prelabeled and not prelabeled with arachidonate suggest that endogenous pool(s) of arachidonic acid are mobilized upon glucose stimulation and that unlabeled fatty acid is metabolized to PGE and PGF, thus diluting the labeled arachldonate activity in those PGs soon after glucose stimulation. Since glucose-stimulated arachidonic acid turnover in islet phospholipids continues for at least 35 min (17), the increase in labeled PGE observed after 30 min in glucose-stimulated but not prelabeled islets may be largely derived from arachldonic acid mobilized from an endogenous pool labeled during prolonged stimulation. Evans et al. (2) have demonstrated increases in islet labeled PGE 2 and PGF2~ biosynthesis after 60 min of glucose stimulation in islets prelabeled with arachidonic acid. An explanation for the more rapid label incorporation in 6-keto-PGFl~ as opposed to the other PGs in islets not prelabeled may include different pools of islet arachidonic acid available for compartmentalized PG production. Quantitation of the metabolite 6-keto-PGFle also includes essentially all of the short-lived PGI 2 produced, and may have allowed for detection of qualitative changes in the metabolite in islets prelabeled or not. In many tissues, including the islet, the role of PGI 2 in eliciting biological responses relies upon the activation of adenylate cyclase and cAMP generation as the second messenger (5,6). Since labeled arachidonic acid conversion to PGI 2 occurs early on in the sequence of events initiated by glucose stimulation, as indicated by the appearance of radiolabeled 6-keto-PGFl~ , PGI 2 may play a role in initiating insulin secretion (6) or affecting the release of other islet hormones through a cAMP-mediated mechanism. A specific role for PGE 2 in the release of islet hormones has not been elucidated to date, however, investigators have

934

JUNE 1984 VOL. 27 NO. 6

PROSTAGLANDINS hypothesized that this PG behaves in a negative feedback capacity (16). The potentiation of glucose-induced insulin release by indomethacin in the present studies supports the hypothesis that PGs modulate insulin release in a negative manner. Alternatively, both or neither of the PGs may directly mediate insulin release. The metabolism of [3H]PGH 2 in islets allowed us to demonstrate a rapid, dynamic and selective synthesis of PGs, including PGD 2 which was difficult to quantitate in islet homogenates (3) but is synthesized by islets from arachidonate over the long term (2). The data showing that the relative proportion of PGs synthesized from [3H]PGH2 and labeled arachidonic acid are similar suggests that a strict control of enzyme activities subsequent to cyclooxygenase and responsible for specific PGE2, PGF2~ and PGI 2 biosynthesis in islets appears to be operative. In these studies we could discern no effect of glucose on the conversion of PGH 2 to PGs, and other investigators have demonstrated that exogenous endoperoxides are not converted to PGs in a manner analogous to arachidonic acid in tissue exposed to a stimulus (30). The "flooding" of enzymatic pathways with supraphysiologic concentrations of PGH 2 may disrupt the delicate balance of enzyme activities regulated by glucose, and arachidonic acid mobilization and metabolism in response to glucose (2,17) may dilute out an effect of glucose on [3H]PGH 2 conversion to PGs. The cation ionophore ionomycin stimulated PG production in islet cells due to stimulation of phospholipase activity (31) and subsequent arachidonate mobilization, and perhaps altered cofactor interaction with arachidonate metabolizing enzymes (33,34). The loss of glucose responsiveness in dispersed cells due to trypsinization and morphological derangement (34) was evident with regard to insulin and PG release, suggesting that certain PGs may amplify the stimulus-secretion signal. However, since indomethacin pretreatment of intact rat islets potentiated glucose-induced insulin release in the absence of PG biosynthesis, a complex picture for PG mediation of insulin release emerges. In the total absence of PG biosynthesis insulin release may be stimulated by glucose in the absence of predominantly negative feed-back effects of PGE (25). On the other hand, ionomycin-stimulated prostacyclin and PGF2~ production was not required for insulin release. Altered Ca 2 homeostasis induced by the ionophore, however, may be expected to have many cellular effects which affect secretory activity (35). These results suggest that enhanced PG production is not essential to insulin release and PGs may function intracellularly in a capacity other than a required mediator of secretion. Corroborative evidence of PGI 2 synthesis in dispersed islet cells is provided by the results demonstrating synthesis of an inhibitor to ADP-induced platelet aggregation. Heretofore, we have reported the biosynthesis of PGI 2 in whole islets containing

JUNE 1984 VOL. 27 NO. 6

935

PROSTAGLANDINS

vascular elements capable of synthesizing PGI 2. In the platelet aggregation studies, the PGl2-1ike inhibitor was derived from cells of disrupted islets lacking identifiable vascular elements. The amount of PGI 2 synthesized by islets based upon the approximate 2000 cells per islet is 12 pg/islet, which is similar in magnitude to the level of 6-keto-PGFl~ reported by others using radioimmunoassay (18). These studies demonstrate that labeled arachidonic acid conversion to PGs in rat pancreatic islets is a rapidly initiated event responsive to physiologic stimulation. The differential production of certain PGs during the sequence of events in glucose-stimulated hormone release may play a regulatory role in the secretory process.

ACKNOWLEDGEMENTS The authors thank Gene R. Bryson for his technical assistance.

REFERENCES i.

2.

3.

4.

5. 6.

7.

8.

936

Luyckx, A.S., and Lefebvre, P.J. Endogenous Prostaglandins Modulate Glucagon Secretion by Isolated Guinea-Pig Islets. Adv. Prostaglandin and Thromboxane Res. 8: 1299-1302, 1980. Evans, M.H., Pace, C.S., and Clements, R.S., Jr. Endogenous Prostaglandin Synthesis and Glucose-Induced Insulin Secretion from the Adult Rat Pancreatic Islet. Diabetes 32: 509-515, 1983. Kelly, K.L., and Laychock, S.G. Prostaglandin Synthesis and Metabolism in Isolated Pancreatic Islets of the Rat. Prostaglandins 21: 759-769, 1981. Metz, S.A., Robertson, R.P., and Fujimoto, W.Y. Inhibition of Prostaglandin E Synthesis Augments Glucose-lnduced Insulin Secretion in Cultured Pancreas. Diabetes 30: 551-557, 1981. Moncada, S. Biological Importance of Prostacyclin PGI2, PGE I, and PGD 2. Br. J. Pharmac. 76: 3-31, 1982. Heaney, T.P., and Larkins, R.G. The Effect of Prostacyclin and 6-Keto-Prostaglandin FI~ on Insulin Secretion and Cyclic Adenosine 3',5'-Monophosphate Content in Isolated Rat Islets. Diabetes 30: 824-828, 1981. Burr, I.M, and Sharp, R. Effects of Prostaglandin E l and of Epinephrine on the Dynamics of Insulin Release In Vitro. Endocrinology 94: 835-839, 1974. Metz, S., Fujimoto, W., and Robertson, R.P. Modulation of Insulin Secretion by Cyclic AMP and Prostaglandin E. Metabolism 31: 1014-1022, 1982.

JUNE 1984 VOL. 27 NO. 6

PROSTAGLANDINS

9.

Nishi, S., Seino, Y., Seino, S., Tsuda, K., Takemura, J., Ishida, H., and Imura, H. Prostaglandin El: Stimulation of Somatostatin and Insulin Response to Glucose from the Isolated Perfused Rat Pancreas. Prostaglandins, Leukotrienes and Med. 8: 593-597, 1982. i0. ~ek, S., Tai, T-Y., and Elster, A. Stimulatory Effects of Prostaglandin El, E2, and F 2 ~ on Glucagon and Insulin Release In Vitro. Diabetes 27: 801-809, 1978. ii. Matsuyama, T., Horie, H., Namba, M., Nonaka, K., and Tarui, S. Glucose Dependent Stimulation by Prostaglandin D 2 of Glucagon and Insulin in Perfused Rat Pancreas. Life Sciences 32: 979-982, 1983. 12. Schusdziarra, V., Rouiller, D., Harris, V., Wasada, T., and Unger, R.H. Effect of Prostaglandin E 2 Upon Release of Pancreatic Somatostatin-like Immunoreactivity. Life Sciences 28: 2099-2102, 1981. 13. Pek, S., Tai, T-Y., Elster, A., and Fajans, S.S. Stimulation by Prostaglandin E 2 of Glucagon and Insulin Release From Isolated Rat Pancreas. Prostaglandins i0: 493-502, 1975. 14. Pek, S., Lands, W.E.M., Akpan, J., and Hurley, M. Effects of Prostaglandins H2, D2, I2, and Thromboxane on In Vitro Secretion of Glucagon and Insulin. Adv. Prostaglan---ain--and Thromboxane Res. 8: 1295-1298, 1980. 15. Metz, S.A., and Robertson, R.P. Prostaglandin Synthesis Inhibitors Reverse -Adrenergic Inhibition of Acute Insulin Response to Glucose. Am. J. Physiol. 239: E490-E500, 1980. 16. Robertson, R.P. Hypothesis: PGE, Carbohydrate Homeostasis, and Insulin Secretion. Diabetes 32: 231-234, 1983. 17. Laychock, S.G. Fatty Acid Incorporation into Phospholipids of Isolated Pancreatic Islets of the Rat. Diabetes 32: 6-13, 1983. 18. Yamamoto, S., Ishii, M., Nakadate, T., and Kato, R. Stimulation of 6-Keto-Prostaglandin FI~ Release From Isolated Pancreatic Islets by an Insulinotropic Concentration of ~ Glucose. Biochem. Biophys. Res. Comm. I14: 1023-1027, 1983. 19. Hamberg, M., and Samuelsson, B. Detection and Isolation of an Endoperoxide Intermediate in Prostaglandin Biosynthesis. Proc. Natl. Acad. Sci. 70: 899-903, 1973. 20. Hamberg, M., Svensson, J., Wakabayashi, T., and Samuelsson, B. Isolation and Structure of two Prostaglandin Endoperoxides that Cause Platelet Aggregation. Proc. Natl. Acad. Sci. 71: 345-359, 1974. 21. Whorton, A.R., Smigel, M., Oates, J.A., and Frolich, J.C. Regional Differences in Prostacyclin Formation by the Kidney: Prostacyclin is a Major Prostaglandin of Renal Cortex. Biochim. Biophys. Acta 529: 176-180, 1978. 22. Sun, F.F., Chapman, J.P., and McGuire, J.C. Metabolism of Prostaglandin Endoperoxide in Animal Tissue. Prostaglandins 14: 1055-1074, 1977.

JUNE 1984 VOL. 27 NO. 6

937

PROSTAGLANDINS

23.

24.

25. 26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

Silver, M.J., Smith, J.B., Ingerman, C., and Kocsis, J.J. Arachldonlc Acid Induced Human Platelet Aggregation and Prostaglandin Production. Prostaglandins 4: 863-875, 1973. Bunting, S., Gryglewski, R., Moncada, S., and Vane, J.R. Arterial Walls Generate from Prostaglandin Endoperoxides A Substance (Prostaglandin X) Which Relaxes Strips of Mesenterlc and Coeliac Arteries and Inhibits Platelet Aggregation. Prostaglandins 12: 897-913, 1976. Robertson, R.P. Prostaglandins, glucose homeostasis, and diabetes mellitus. Med. Clin. N. Amer. 65: 759-771, 1981. Lee, S.C. and Levlne, L. Prostaglandin Metabolism. II. Identification of two 15-hydroxyprostaglandin dehydrogenase types. J. Biol. Chem. 250: 548-552, 1975. Leslie, C. and Levine, L. Evidence for the Presence of a Prostaglandin E 2 -9-keto reductase in rat organs. Biochem. Biophys. Res. Comm. 52: 717-724, 1973. Sih, C.J. and Takeguchi, C.A. Biosynthesis. In The Prostaglandins (ed. P.W. Ramwell), vol.l, Plenum Press, N.Y. 89-91, 1973. Laychock, S.G., Warner, W., and Rubin, R.P. Further Studies on the Mechanisms Controlling Prostaglandln Biosynthesis in the Cat Adrenal Cortex: The Role of Calcium and Cyclic AMP. Endocrinology I00: 74-81, 1977. Birkle, D.L., Wright, K.F., Ellis, C.K. and Ellis, E.F. Prostaglandin levels i~ isolated brain mlcrovessels and in normal and norepinephrine-stimulated cat brain homogenates. Prostaglandins 21: 865-877, 1979. Wong, P.Y-K., and Cheung,W.Y. Calmodulin Stimulates Human Platelet Phospholipase A 2. Biochim. Biophys. Res. Comm. 90: 473-480, 1979. Bokoch, G., and Reed, P.W. Stimulation of Arachidonic Acid Metabolism in the Polymorphonuclear Leukocyte by an N-Formyfated Peptide. J. Biol. Chem. 255: 10223-10226, 1980. Wong, P.Y-K., Lee, W.H., and Chao, P.H-W. The Role of Calmodulin in Prostaglandln Metabolism. Ann. N.Y. Acad. Sci. 356: 179-189, 1980. Pipeleers, D., Veld, P.I., Maes, E., and Van de Wlnkel, M. Glucose-induced Insulin Release Depends on Functional Cooperation Between Islet Cells. Proc. Natl. Acad. Sci. 79: 7322-7325, 1982. Rubin, R.P. Calcium and Cellular Secretion. Plenum Press, New York, 1982.

Editor: Eo Paul Robertson

938

Received: 9-8-83

Accepted: 5-14-84

JUNE 1984 VOL 27 NO. 6