Cd/ Cddm
(1992) 13, 571-580
0 Longman Group UK Ltd 1982
Evidence that basal secretion of relaxin by individual cultured large luteal cells is influenced by mobilization of intracellular calcium: Analysis by a reverse hemolytic plaque assay M.J. TAYLOR and CL. CLARK Department of Veterinary Physiology and Pharmacology, College of Veterinary Medicine, Iowa State University, Ames, Iowa, USA Abstract - Ca2’ redistribution from an intracellular site(s) is a key biochemical event associated with relaxin (RLX) secretion by large luteal cells (LLCs) of porcine origin. However, the functional significance of internal stores of Ca2+ to basal rates of RLX secretion is not well understood. In addition, the identity of the intracellular storage site(s) for Ca2+ within LLCs is not known, nor is it clear if all RLX-releasing LLCs are equally dependent on this pool. In the present study, release of RLX from 24 h cultured luteal cells derived from early pregnant swine was monitored by a reverse hemolytic plaque assay (RHPA). incubation of cultures in the presence of graded concentrations of thapsigar$+n (1 nM - 1 j&l), a plant sesquiterpene lactone that inhibits endoplasmic reticulum Ca ATPase and thereby increases cytosolic Ca2+ concentrations, resulted in a dose-related increase in basal RLX secretion. The stimulatory effect of thapsigargin on RLX production was not abrogated by culture in Ca2+-free medium. Suppression of Ca2’ release from the endoplasmic reticulum of LLCs, achieved by incubating monolayers in medium containing dantrolene (i-100 fl), resulted in dose-related Inhibition of basal RLX release. Taken together, these results suggest that the endoplasmic reticulum serves as a major storage site for Ca2+ redistribution within LLCs and, furthermore, that mobilization from this site is functionally coupled to basal secretion of RLX.
In a previous study, we reported several lines of evidence to suggest that secretion of the insulin-like peptide hormone relaxio (RLX) from porcine large luteal cells (LLCs) is a Ca2+-dependent event In brief, release of RLX was conspicuously stimulated by the Ca2’ iooophore, A23187, in a dose-dependeot fashion while the effect of A23187 (and the
effect of stimulatory secretagogues) was blocked by a Ca2c-ch~nel blocker [l]. Interestingly, basal rates of RLX release were also suppressed by a Ca2+-channel blocker, raising the possibility that basal secretion of RLX was also dependent on Ca2’ redistribution [I]. However, we found that neither basal nor mod571
572
ulated RLX secretion was dependent on the presence of Ca2+ in the medium bathing the monolayers [l]. Accordingly, we concluded that while RLX secretion by LLCs was dependent on Ca2+ redistribution, the source of Ca2+ was likely to be an intracellular site within LLCs, rather than influx of external Ca2+. The endo~lasmic reticulum has been identified as a major Ca ’ storage site in secretory ceils and is believed to release Ca2+ during secretagogue-induced hormone secretion [2]. However, the role of endoplasmic reticulum-stored Ca2+ in controlling basal (non-secretagogue-induced) hormone secretion is less well understood. Pertinently, LLCs posses abundant amounts of endoplasmic reticulum [3, 41. Thus the first objective of the present study was to test the hypothesis that the endoplasmic reticulum of LLCs serves as a Ca2’ store that is functionally linked to RLX production. A second objective was to establish if this store is important to basal rates of RLX secretion, an event that would widen the functional significance of endoplasmic reticulum-stored Ca2’ . To meet these objectives, we report here the results of experiments in which Ca2+ redistribution from the endoplasmic reticulum was manipulated by two separate means. Thapsig in is a plant product a”gL+ [5] that increases cytosolic Ca concentrations in luteal cells [6] by a mechanism that is neither ionophoric nor related to increased inositol 1.4.5trisphosphate (IP3) production [7]. It acts as a specific inhibitor of endoplasmic reticulum-bound Ca2’ATPase and thapsigargin-induced increases in cytosolic Ca2+ are believed to arise from Ca2’ leakage from the endoplasmic reticulum and suppression of sequestration of leaked Ca2’ back into the endoplasmic reticulum because of reduced activity of the Ca2+-ATPase [S]. In contrast, dantrolene has been reported to inhibit mobilization of Ca2’ from the endoplasmic reticulum and thus reduce cytosolic Ca2’ concentrations and Ca2+-dependent hormone secretion [9-131. If Ca2’ mobilization from the endoplasmic reticulum is functionally coupled to basal RLX secretion, then it is logical to predict that thapsigargin and dantrolene would stimulate and inhibit RLX secretion. respectively. The rate of RLX release was monitored by use of a RLX-reverse hemolytic plaque assay, a technique that employs antibody-directed, complement-
CELL CALCIUM
mediated lysis of target erythrocytes around secretory LLCs (cultured in a layer one cell thick) in order to monitor RLX release from individual LLCs. This cell-by-cell approach was selected to identify potential individual cell variations in functional aspects of Ca2’ mobilization.
Materials and Methods Full details of the collection of tissue, cell dispersion and plaque assay methodology are provided elsewhere [l, 141, and only a brief description is given here. Animals Ovaries from pregnant pigs (day 30-40 of gestation) were collected within 30 min of death at the Meat Laboratory, Iowa State University, USA. Gestational age was determined by the day of mating by an intact boar and confirmed by fetal crown-rump length [15]. Both ovaries were placed in medium and transported back to the laboratory within lo-60 min. The medium used in this step and all subsequent steps in cell dispersion and assays was Dulbecco’s Modified Eagles Medium (DMEM) obtained from Grand Island Biological Co., (Gibco, Grand Island, NY, USA) and supplemented with antibiotics (100 U/ml penicillin G and 100 @ml streptomycin sulfate; Gibco) and 0.1% bovine serum albumin (fraction V, Sigma, St Louis, MO, USA). This medium will be referred to hereafter as DMEM-0.1% BSA. Cell dispersion Three to four individual corpora lutea (CL) were incised, the luteal tissue gently enucleated from the fibrous capsule and then washed 2-3 times with fresh DMEM-O.18 BSA. Approximately one quarter of each CL was then placed in a sterile 65 mm Petri dish containing 3 ml of DMEM-0.1% BSA, minced finely with sterile scalpel blades in a laminar flow hood, and washed repeatedly with fresh medium. The luteal fragments were then placed in a Spinners suspension flask (Bellco, Vineland, NJ, USA) containing 15 ml DMEM-0.1%
REVERSE HEMOLYTIC PLAQUE ASSAY & RELAXIN BASAL SECRETION
BSA, 0.12% collagenase (type III; Cooper Biomedical, FEehold, NJ, USA). The tissue was then incubated for 60 min after which time the fragments were drawn in and out of a sterile 10 ml pipette, incubated for a further 10 min and then centrifuged at 1500 g for 10 min. The cells were then washed twice in Spinners Minimum Essential Medium and passed through a 75 pm mesh (Tetco Inc., Elmsford, NY, USA) to remove cell clumps. An aliquot of the mixed cell suspension was taken for LLC cell counts (using a hemocytometer) and cell viability (trypan blue exclusion). Yields were normally about l-2 x lo6 LLCs, and cell viability was 80-95%. The mixed luteal cell suspension was resuspended in DMEM-0.1% BSA at a final concentration of 0.25 x lo6 LLCs/ml. Reverse henwlyticplaque assay (RHPA) Incubation chambers (Cunningham chambers) were constructed by attaching a glass coverslip to a polyL-lysine treated glass microscope slide using doublestick tape. The mixed luteal cell suspension was then mixed 1:l with a 12% suspension of sheep erythrocytes to which Protein A (Sigma) had been chemically attached [16]. This mixture was then infused into each chamber by capillary action and incubated for 45 min (5% CO2/95% air humidified atmosphere) to allow cell attachment. Each chamber was then flushed with fresh medium to form a layer one cell deep on the chamber floor, and then incubated overnight in the presence of DMEM0.1% BSA. The monolayers were then washed again with fresh medium in order to remove cell products that had accumulated overnight (and dead or dying cells which normally detach), and immediately filled with fresh medium containing porcine RLX antiserum (1:80) and thapsigargin or dantmlene sodium. Dantrolene was dissolved in methanol to a 1 mM stock solution and stored at -20°C. Thapsigargin was dissolved in dimethyl sulfoxide (DMSO) and stored as a 1 m.M stock solution at -20°C. Each agent was diluted to the appropriate concentration immediately before experimentation, and the maximal concentration of methanol or DMSO in media was 0.1%. The same concentration of vehicle was added to controls. We have reported previously that these dilutions of
573
vehicle do not affect the rate of plaque formation [l]. After infusion of antibody containing agents or control vehicle, the monolayers were incubated for periods of 1.2, 3,4, 8 and 12 h. We have shown in previous studies that maximal plaque formation is achieved by 8-12 h of incubation [l, 141. At the end of each incubation period, each chamber was filled with medium containing guinea pig serum as a source of complement (1:40, Gibco), incubated for 50 min to complete plaque formation, fixed with 8% glutaraldehyde in normal saline, and stored at 4°C immersed in the same fixative in airtight containers. Each monolayer was then examined microscopically to score the percentage of LLCs that formed plaques; at least 200 LLCs per monolayer were counted, and triplicate chambers were prepared (for each treatment or control) at each timepoint. Percentages from the triplicate monolayers were averaged. Small luteal cells (i.e. cells less than 15-18 pm in size) or other luteal cell types were not counted since we have previously established that only LLCs form plaques in the RLX-RKPA [l, 141. In addition, the average size of plaques in each monolayer was measured by superimposing a microscopic image of each monolayer over a digitizing pad and drawing the exact shape of 100-200 plaques/monolayer. Image analysis software (SigmaScan; Jandel Scientific, Sausalito, CA, USA) was used to record, compute and analyze the resulting plaque areas (after calibration). Values obtained from triplicate monolayers were combined and averaged. Drugs and reagents All drugs and reagents were purchased from Sigma. Ca2+-freeand &+-replete medium Medium referred to as Ca2’-free was Minimum Essential Medium with Earle’s salts (MEM; contains no added Ca2’*, Gibco, #320-1380). 0.1% bovine serum albumin (BSA, fraction V, Sigma) and contained 0.2 mM EGTA to chelate adventitious Ca2’. The Ca2+-replete medium was DMEM-0.1% BSA. Both media were supplemented with antibiotics (100 U/ml penicillin G and 100 pg/ml streptomycin sulfate; Gibco). Analysis of the [Ca2+]; of three
CELL cAr.cluM
574 80
=cn
70
z F .E
60
0
Control
v
Thapsigargin
(+Co2+)
(1nM)
v
Thapsigargin
(10nM)
q
Thapsigargin
n
Thapsigargin
*ib 50 Q, is CJ a
40
z
30
Fk 0 ii
20
,o n"
10
0 1
0
2
3
4
5
6
Incubation
7
8
Time
9
10
11
12
13
(h)
Fig. 1 Rate of basal plaque formation in monolayers exposed to calcium-replete medium alone, or the same medium containing graded concentrations of thapsigargin (1 nM - 1 @I)). and infused into micmincubation medium, and the monolayers containing thapsigargin. normal saline).
Monodispersed luteal cells were combined with Pmtein A labelled sheep erythmcytcs
chambers to form monolayers.
After ovemight incubation, the chambers were washed with fresh
then incubated for 1-12 h in the presence of control medium (containing vehicle alone) or medium
Complement was then infused to develop the plaques and the. monolayers fixed with glutaraldehyde (8% in
The data shown arc the mean f SEM from three independent experiments using cells obtained from day 32-36 pmgnaat
pigs. Values denoted by an asterisk am significantly different from controls
(P< 0.05) at the same timepoint. See text for fall details of
the reverse hemolytic plaque assay
separate samples of both media from separate batches by atomic absorption spectra- photometry provided values of 1.6 f 0.1 mM and 6.6 + 0.9 pM [Ca2+]i (mean Z!ISD) in DMEM-0.1% BSA and MBM-0.1% BSA, 0.2mM EGTA, respectively. Controlprocedures Specificity of the porcine RLX antiserum used here has been reported previously [17, 181. Plaque formation was abolished by pre-absorption of RLX
antibody (1:80) with 10 pg/ml puritied porcine RLX and by omission of antibody or complement.
Expression of results Results am expressed as mean St SEM. Differences between control and treatment groups were evaluated by analysis of variance (ANOVA) followed by Newman-Keul range test. A P value less than 0.05 was selected as significant 1191.
575
REVERSE HEMOLYTIC PLAQUE ASSAY & RJXAXIN BASAL SECRETION
of plaque-forming LLCs in monolayers treated with 100 nM thapsigargin was significantly greater than Eflect of inhibitorsof Ca2’ redistributionon the rate control values at 3 and 4 h of incubation, but not at of plaque formation - thapsigargin the earliest stages of incubation (namely, 1 and 2 h of incubation). There was no significant difference between Monolayers were incubated in a Ca2+-replete control and any treatment groups at 8 h of incubatmedium containing graded concentrations (1 nM 1 @I) of thapsigaxgin. Concentrations of 1 and 10 ion (P > 0.05). By 12 h of incubation, all groups nM thapsigargin exerted no significant effect on the had clearly attained maximal plaque formation rate of plaque formation (Fig. 1). However, the (60-65% of all LLCs). This apparent loss of effect of thapsigargin occurs because all RLX-releasing percentage of plaque-forming LLCs in monolayers treated with 1 pM thapsigargin was significantly LLCs will ultimately secrete the threshold amount greater than control values at l-8 h of incubation. of RLX required to form a plaque, regardless of treatment. Interestingly, it was apparent that For example, there were nearly twice as many plaque-forming LLCs in thapsigargin-treated monothapsigargin did not recruit additional LLCs into the layers (43 rtr 4%) than in control monolayers (23 z!z secretory population; note that the percentage of 5%) at 4 h of incubation. LLCs in monolayers treated with 100 nM or 1 pM An intermediate dose of thapsigargin (100 nM) thapsigargin was virtually identical at 8 and 12 h of elicited a time-dependent response. The percentage incubation (Fig. 1). Results
80
i O Control n
Co7trol
o
Thopsigargin
(1,uM;
+Ca2+)
q
Tiapsiqorqin
(lr*.M;
-Ca2+)
70 In =
(+CaZ+) (-Coz’)
z 60 CT .-c
E b 7a, 3 g
50
40
a “0 : 0 L : & CL
30
20
10
0
I0
1
2
3
4
5
6
Incubation
7
8
Time
9
10
11
12
13
(h)
Fig. 2 Rate of plaque formation in monolayers exposed to calcium-replete or calcium-free medium, or the same media containing 1 @I thapsigargin. The experimental paradigm was identical to that dcscribcd for Figure 1. The data shown are the mean f SEM from three independent experiments using cells obtained from day 32-36 pregnant pigs. Values denoted by au asterisk are significantly different from controls (P < 0.05) at the same timepoint; thcrc was no sign&ant effect of lack of calcium under basal or thapsigargin-modulated conditions
CELL CALCIUM
576
On the basis of these data, a concentration of 1 @I thapsigargin was selected for the next series of experiments, in which we compared the effect of thaf$gargin on RLXz+secretionby LLCs bathed in Ca -replete and Ca -free medium Under basal conditions, the percentages of plaque-forming LLCs in monolayers bathed in Ca2+-repleteand Ca2+-free medium were not significantly different at any stage of the incubation (Fig. 2). Similarly, the percentages of plaque-forming LLCs in monolayers bathed in Ca2+-replete and Ca2’-free medium containing 1 pM thapsigargin were not significantly different at any stage of the incubation, although each was significantly greater than corresponding control values at l-4 h of incubation (Fig. 2). To verify the stimulatory effect of thapsigargin on basal RLX secretion, the average areas of plaques (an index of cumulative RLX secretion) were measured in the monolayers of the preceding experiments. The average area of plaques provides
0
Cortrol
direct index of the cumulative amount of RLX released over time [20]. It was found that thapsigargin (1 @I) significantly increased (P < 0.05) the average area of plaques by 24fold over the course of the experiment (Fig. 3). confirmation of a strong stimulatory effect of thapsigargin on RLX production. The absence of Ca2’ in the medium did not significantly affect the amount of RLX secreted under basal conditions, or in response to thapsigargin. a
Eflect of inhibitorsof Ca2+redistributionon the rate of plaque formation - dantrolene
To ascertain the effect of dantrolene (an inhibitor of Ca2’ release from endoplasmic reticulum) on the rate of basal RLX secretion, monolayers were first incubated in the presence of graded concentrations of dantrolene (1, 25 and 100 @4) or medium containing vehicle alone (control). We found that a *
(+co2+)
0 Cortrol (-Co”)
C
n
Thopsigorgin
(1pM;
+Ca+‘)
0
Thoxiga-gin
(1pM;
-Ca+*)
12
3
4
Incubation
5 Time
6
7
8
9
(h)
Fig. 3 Mean ama of plaques (cumulative RLX release) in monolayers exposed to calcium-replete or calcium-free medium (controls) or the same media containing thapsigargin (1 pM) during a timecourse plaque assay.
The experimental paradigm was a~ described for
Figure 1. Values am the mean f SEM of thme independent experiments involving luteal cells derived from day 32-36 pregnant pigs, and the symbols marked by an asterisk are significantly different (P C 0.05) from control values at the same timepoint of the incubation; them was no significant effect of lack of calcium under baaal or thapsigargin-modulated not be included because of confluence of plaques at this stage
conditions.
Data from the 12 h timepoint could
577
REVERSE HEMOLYTIC PLAQUE ASSAY & RELAXIN BASAL SECRETION
0
60
V
0
1
Control Dontrolene
2
3
(1pM)
4
5
6
incubation
7 Time
8
9
10
11
12
13
(h)
Fig. 4 Rate of plaque formation in monolayers exposed to control medium or graded concentmtions of dantmlene. The data shown are the mean i SEM (triplicate monolayers) from a single representative experiment using luteal cells obtained from a day 39 pregnant pig. Values denoted by an asterisk are significantly d&rent (P< 0.05) from controls at the same timepoint
dose of 1 p.M dantrolene did not significantly affect the rate of plaque formation. However, doses of 25 and 100 @I dantrolene induced a significant reduction (P < 0.05) in the percentage of plaqueforming LLCs present at l-8 h of incubation (Fig, 4). For the reasons alluded to above, this inhibition was no longer present at 12 h of incubation, when maximal plaque formation (about 55%) was reached in all monolayers. Since there was no detectable difference between the inhibitory effects of 25 and 100 pM dantrolene, a concentration of 25 @I dantrolene was selected for use in the next experiments. To verify and quantify the inhibitory effect of 25 pM dantrolene on RLX secretion, we next compared the size of plaques in control and dantrolene-treated monolayers (Fig. 5). The mean area of plaques in dantrolene-treated monolayers over the total course of the experimental incubation was 55.3 f 4.0% of control values, demonstrating that exposure to dantrolene reduced RLX secretion to about half of control values.
Discussion We have shown that while secretion of RLX by porcine LLCs is a Ca2+-dependent phenomenon, neither basal nor stimulated hormone secretion was dependent on a supply of extracellular Ca2+ [l]. Consequently, we concluded that the primary source of Ca2’ coupled to RLX secretion is mobilization from an intracellular compartment, rather than influx of extracellular Ca2’. The goal of the present study, therefore, was to confii this conclusion and, furthermore, to identify the intracellular source responsible for storing, releasing and presumably In other sequestering Ca2+ ions within LLCs. secretory cell types, the endoplasmic reticulum is considered to be the primary Ca2+ store. Our strategy was to use specific pharmacological modulators of Ca2+ release from this organelle, and to ascertain their effect on RLX production. To monitor RLX release, we employed a reverse hemolytic plaque assay that detects the release of hormone from individual LLCs in culture. The rate
CELL CALCIUM
578
of plaque formation in this approach, together with measurement of average plaque areas, provides information on inhibitory and stimulatory secretagogues (see Neill et al. [20], for an extensive review). This method was employed because we wished to examine cell-by-cell responses. The results of the present study demonstrate that thapsigargin induced a strong, dose-related increase in the rate of RLX secretion. Since it is believed that thapsigargin increases cytoplasmic Ca*+ concentrations by specitically inhibiting a Ca*+sensitive ATPase in the endoplasmic reticulum 581, these data imply that mobilization of Ca*’ from the endoplasmic reticulum is of primary importance to RLX production. Our further observation that removal of extracellular Ca*’ failed to abrogate the stimulatory effect of thapsigargin on RLX secretion is consistent with this view, and confirms our earlier inference that extracellular Ca*+ is not required for basal or modulated RLX secretion [l]. It is worth emphasizing, moreover, that the experiments of the current study were carried out wholly under basal (non-stimulated)
conditions, and contrast with a preceding study of hormone production (progesterone) by luteal cells, in which only LH-enhanced hormone production was reported to be sensitive to thapsigargin [6]. The marked effect of thapsigargin on FUX secretion suggests that tonic efflux of Ca*+ from the endoplasmic reticulum is likely to be of fundamental importance to basal RLX production. It was pertinent that the effect of 100 nM thapsigargin on plaque-formation was strikingly time-dependent (Fig. 1). Although an obvious explanation for the delay in functional response is that thapsigargin required l-2 h to penetrate LLCs and elicit a response, a higher dose of thapsigargin (1 pM) was immediately effective. Furthermore, 100 nM thapsigargin was reported to increase Ca*’ mobilization in rat luteal cells within seconds [6]. Thus, a more likely explanation is that of differential responsiveness to thapsigargin by individual LLCS. Put briefly, it is possible that 100 nM thapsigargin represents a sub-effective dose for some LLCs, specifically those which nounally form plaques early in the incubation period (i.e. those that
30
x NE ,s H “,
24 22 20 18 16 14
0
12
3
4
Incubation
Ftg.5
Mean area of
plaques(cumulative
5 Time
6
8
9
(h)
RLX release) in monolayers exposed to medium
dantrolene (25 pM) during a timecourse plaque assay.
7
plusvehicle
(control) or medium containing
Values are the mean f SEM of three independent experiments involving luteal
cells derived from day 31-44 pregnant pigs, and the symbols marked by an asterisk are significantly different (P < 0.05) from control values at the same timepoint of the incubation.
Where no vertical enor bar is shown, the value is less than the size of the symbol.
from the 12 h timepoint could not be included because of confluence of plaques at this stage
Data
REVERSE HEMOLYTIC PLAQUE ASSAY & RELAXIN BASAL SECRETION
release RLX at the highest rates). On the other hand, the same concentration of thapsigargin (100 nM) clearly increased the secretory abilities of other LLCs that secreted RLX at lower rates (and which form plaques in the midpart of the incubation). Differential responsiveness to thapsigargin by individual LLCs implies that cell-to-cell differences in the amount or activity of the endoplasmic reticulum-bound Ca2+-ATRase may exist. It is tempting to speculate that these differences, if translated into cell-to-cell variations in internal Ca2+ mobilization, may provide one mechanistic explanation for the remarkable heterogeneity of basal RLX secretion exhibited by individual LLCs [l, 141. It was also of interest that thapsigargin failed to elicit any RLX secretion from a fraction of LLCs (non-secretory). We can provide no obvious explanation for this phenomenon, although it is possible that these LLCs were selectively damaged by dispersion procedures. We have shown, however, that nonsecretory LLCs contain abundant immunoreactive RLX and express the RLX gene [21], characteristics that do not favor cell damage. Moreover, non-secretory cells exist in several other endocrine cell types [20], suggesting that the existence of non-secretory subpopulations among It is secretory cells is a general phenomenon. possible that thapsigargin failed to elevate cytosolic Ca2+ concentrations in non-secretory LLCs, or alternatively, that Ca2+ mobilization is uncoupled from RLX secretion in this subpopulation of LLCs. Thapsigargin arachidonic stimulates acid production and PGE:! production in non-luteal cells [22], and we cannot rule out the possibility that these or other biochemical changes may have occurred in LLCs in the present study. Nevertheless, we used a second and independent method to influence Ca2+ redistribution from the endoplasmic reticulum Dantrolene is a compound that reportedly reduces Ca2+ efflux from the endo- plasmic reticulum and has been used, therefore, to examine Ca2+-sensitive, in vitro hormone secretion 19-111. In the present study, dantrolene significantly slowed the rate of basal RLX secretion in a dose-related manner. In contrast to thapsigargin, we observed no timedependency of RLX inhibition, although doses intermediate between 1 and 25 @VI (minimal and maximal concentrations, respectively), must be
579
tested to verify this result. As judged by average plaque area (an index of cumulative RLX secretion), a concentration of 25 pM dantrolene reduced the overall amount of RLX produced by about one half. This result is entirely consistent with the effect of thapsigargin on RLX secretion, and strengthens our conclusion that mobilization of Ca2+ from the endoplasmic reticulum is implicated in basal RLX production. In summary, the evidence presented in the current study is consistent with the view that the endoplasmic reticulum serves as a primary storage site of Ca2+ within LLCs that is functionally linked to RLX production. and serves to confirm that extracellular Ca2+ IS ’ not essential for RLX production. Of course, other potential storage sites, such as calciosomes [23], may also participate in Ca2+ handling. Secretion of LH by rat gonadotropes [24] as well as progesterone secretion by granulosa cells [251 and luteal cells [6] were also attributable (at least partially) to redistribution of Ca2+ from an intracellular source. Nevertheless, these prior studies all involved the analysis of secretagogueinduced hormone release rather than basal secretion. The current study appears to be the first to implicate mobilization of intracellular Ca2’ stores with basal hormone secretion, and thapsigargin will be a highly useful tool to investigate this novel phenomenon in more detail, particularly with respect to heterogenous RLX secretion.
Acknowledgements This work was supported by Grant HD-22786 from the National Institutes of Health, USA. We thank Dr David Sherwood. University of Illinois, Urbana-Champaign, USA, for the generous donation of RLX antiserum.
References Taylor MT. Clark CL. (1989) Effect on calcium ionophores and calcium-channel blockers on relaxin release from cultured porcine luteal cells. Endocrinology, 123, 1893-1901. Bet-ridge MJ. (1987) Inositol phosphate and diacylglycerol: two interacting second messengers. Annu. Rev. Biochem., 56, 159-193. Chegini N. Ramani N. Rao CV. (1984) Morphological and
580
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
CELL CALCIUM biochemical characteristics of small and large bovine luteal cells during pregnancy. Mol. Cell Endocrinol., 37,89-102. Niswender GD. Schwa11RM. Fitx TA. Farin CH. Sawyer HR. (1985) Regulation of luteal function in domestic ruminants: new concepts. Rec. Rag. Horm. Res., 41, 101-150. Hakii H. Fujiki H. Saganuma M. et al. (1986) Thapsigargin, a histamine secretagogue. is a non-12-0-tetradecanoylphorbol-13-acetate type tumor promotor in two stage mouse skin carcinogenesis. J. Clin. Cancer Res. Clin. Oncol., 111.177-187. Pepemll JR. Behrman HR. (1990) The calcium-mobilizing agent, thapsigargin, inhibits progesterone production in rat luteal cells by a calcium-independent mechanism. Endocrinology, 127, 1818-1824. Jackson TR. Paterson SI. l’hastrup 0. Hanely MR. (1988) A novel tumor-promotor, thapsigargin, transiently increases cytoplasmic free Ca’+ without generation of inositol phosphates in NGllS-401L neutonal cells. B&hem. J., 253,81-86. Thastrup 0. Cullen PJ. Drobak BK. Hanely MR. Dawson AP. (1990) Thqsigargin, a tumor pmmotor, discharges intracellular Ca’ stores b specific inhibition of the endoplasmic reticulum Ca5 -ATPase . Pmt. Natl. Acad. Sci. USA, 87.2466-2470. Winkle WBV. (1976) Calcium release from skeletal muscle samoplasmic reticulum: site of action of dantmlene sodium. Science, 193, 1130-1131. Comt PM. Staley D. Jimtah H. Bates M. (1985) Molecular mechanisms of gonadotropin releasing hormone action. J. Steroid B&hem., 23.703-710. Lihtmann I. Delarue C. Homo Delarche F. FeuilloIey M. Belanger A. Vaudry H. (1987) Effects of TMB-8 and dantrolene on A(JTH and angiotensin-induced steroidogenesis by frog interrenal glands: evidence for a role of inttacellular calcium in angiotensin action. Cell Calcium, 8,269-282. Metx SA. (1988) Exogenous arachidonic acid promotes insulin release from intact or permeabilixed rat islets by dual mechanisms. Putative activation of Ca” mobilization and protein kinase C. Diabetes, 37, 1453-1469. Veldhuis JD. May W. Juchter D. (1987) Mechanisms subserving hormone action on the ovary: role of calcium ions as assessed by steady state calcium exchange in cultured swine gramtlosa cells. Endocrinology, 120, 445-449. Taylor MJ. Clark CL. Frawley LS. (1987) Analysis of relaxin release from cultured porcine luteal cells by a reverse hemolytic plaque assay: influence of gestational age and prostaglandin Fh. Endocrinology, 120,2085-2091. Evans HE. Sack WO. (1973) Prenatal development of domestic and laboratory animals: growth curves, external
16.
17.
18.
19. 20.
21.
22.
23.
24.
25.
features and selected references. Anat. Hist. EmbryoI., 2, 1l-20. Neil1 JD. Frawley LS. (1983) Detection of hormone release from individual cells in mixed populations using a reverse hemolytic plaque assay. Endocrinology, 112, 1135-I 137. Sherwood OD. (1982) Radioimmunoassay of relaxin. In: Anderson RR. ed Proceedings of the 15th Midwest Conference on Endccrinology and Metabolism - Relaxin. New York, Plenum Press, pp. 315-321. Shetwood OD. Rosentreter KH. Birkhimer ML. (1975) Development of a radioimmunoassay for porcine relaxin using a ‘%abelled polytytosyl-relaxin. Endocrinology, 96, 1110-1115. Steel RGD. Torrie JH. (1960) Principles and Procedures of Statistics. New York, McGraw-Hill. Neil1 JD. Smith PF. Luque EH. Toro MM. Nagy G. Mulchahey JJ. (1987) Detection and measurement of hormone secretion from individual pituitary cell. Rec. Prog. Horm. Res.. 43.175-229. Bagnell CA. Ohleth K. Clark CL. Taylor MJ. (1992) Detection of mlaxin secretion and gene expression in individual porcine htteal cells. In: Lueng PCK. ed. Molecular Basis of Reproductive Endocrinology. New York, Springer Verlag. Ohuchi K. Suguwata T. Watanabe M. et al. (1987) Stimulation of amchidonic acid metabolism in tat peritoneal macrophages by thapsigargin a non-(12 0-tetradecanoyl phorbol-13-acetate) (TPA)-type tumour ptomotor. J. Cancer Res. Clin. Gncol., 113.319-325. Krause KH. Pittet D. Volpe P. Pozzan T., Meldolesi J. Lew DP. (1989) Cal&some, a sarcoplasmic reticulum-like organelle involved in intracelhtlar Ca2’-handling in non-muscle cells: studies in human neutrophils and HL-60 cells. Ceil Calcium, 10, 351-361. Shangold GA. Murphy SN. Miller RJ. (1988) Gonadotropin releasing hormone-induced Ca” transients in single identified gonad0 s requires both intracellular Ca2+ mobilization and m?e Ca + influx. Pmt. Nat]. Acad. Sci. USA, 85.6566-6570. Lee HL. Shangold GA. Larsen AL. Schreiber JR. (1990) The role of exogenous calcium for gonadotropin-stimulated progeatemne production by human granulosa-luteal cells. Fert. Steril., 52.958-964.
Please send reprint requests to : DC Michael J. Taylor, Department of Veterinary Physiology and Pharmacology, College of Veterinary Medicine, Iowa State University, Ames IA 50011, USA Received : 14 November 1991 Accepted : 9 April 1992
(1992) 13, 571-580
0 Longman Group UK Ltd 1982
Evidence that basal secretion of relaxin by individual cultured large luteal cells is influenced by mobilization of intracellular calcium: Analysis by a reverse hemolytic plaque assay M.J. TAYLOR and CL. CLARK Department of Veterinary Physiology and Pharmacology, College of Veterinary Medicine, Iowa State University, Ames, Iowa, USA Abstract - Ca2’ redistribution from an intracellular site(s) is a key biochemical event associated with relaxin (RLX) secretion by large luteal cells (LLCs) of porcine origin. However, the functional significance of internal stores of Ca2+ to basal rates of RLX secretion is not well understood. In addition, the identity of the intracellular storage site(s) for Ca2+ within LLCs is not known, nor is it clear if all RLX-releasing LLCs are equally dependent on this pool. In the present study, release of RLX from 24 h cultured luteal cells derived from early pregnant swine was monitored by a reverse hemolytic plaque assay (RHPA). incubation of cultures in the presence of graded concentrations of thapsigar$+n (1 nM - 1 j&l), a plant sesquiterpene lactone that inhibits endoplasmic reticulum Ca ATPase and thereby increases cytosolic Ca2+ concentrations, resulted in a dose-related increase in basal RLX secretion. The stimulatory effect of thapsigargin on RLX production was not abrogated by culture in Ca2+-free medium. Suppression of Ca2’ release from the endoplasmic reticulum of LLCs, achieved by incubating monolayers in medium containing dantrolene (i-100 fl), resulted in dose-related Inhibition of basal RLX release. Taken together, these results suggest that the endoplasmic reticulum serves as a major storage site for Ca2+ redistribution within LLCs and, furthermore, that mobilization from this site is functionally coupled to basal secretion of RLX.
In a previous study, we reported several lines of evidence to suggest that secretion of the insulin-like peptide hormone relaxio (RLX) from porcine large luteal cells (LLCs) is a Ca2+-dependent event In brief, release of RLX was conspicuously stimulated by the Ca2’ iooophore, A23187, in a dose-dependeot fashion while the effect of A23187 (and the
effect of stimulatory secretagogues) was blocked by a Ca2c-ch~nel blocker [l]. Interestingly, basal rates of RLX release were also suppressed by a Ca2+-channel blocker, raising the possibility that basal secretion of RLX was also dependent on Ca2’ redistribution [I]. However, we found that neither basal nor mod571
572
ulated RLX secretion was dependent on the presence of Ca2+ in the medium bathing the monolayers [l]. Accordingly, we concluded that while RLX secretion by LLCs was dependent on Ca2+ redistribution, the source of Ca2+ was likely to be an intracellular site within LLCs, rather than influx of external Ca2+. The endo~lasmic reticulum has been identified as a major Ca ’ storage site in secretory ceils and is believed to release Ca2+ during secretagogue-induced hormone secretion [2]. However, the role of endoplasmic reticulum-stored Ca2+ in controlling basal (non-secretagogue-induced) hormone secretion is less well understood. Pertinently, LLCs posses abundant amounts of endoplasmic reticulum [3, 41. Thus the first objective of the present study was to test the hypothesis that the endoplasmic reticulum of LLCs serves as a Ca2’ store that is functionally linked to RLX production. A second objective was to establish if this store is important to basal rates of RLX secretion, an event that would widen the functional significance of endoplasmic reticulum-stored Ca2’ . To meet these objectives, we report here the results of experiments in which Ca2+ redistribution from the endoplasmic reticulum was manipulated by two separate means. Thapsig in is a plant product a”gL+ [5] that increases cytosolic Ca concentrations in luteal cells [6] by a mechanism that is neither ionophoric nor related to increased inositol 1.4.5trisphosphate (IP3) production [7]. It acts as a specific inhibitor of endoplasmic reticulum-bound Ca2’ATPase and thapsigargin-induced increases in cytosolic Ca2+ are believed to arise from Ca2’ leakage from the endoplasmic reticulum and suppression of sequestration of leaked Ca2’ back into the endoplasmic reticulum because of reduced activity of the Ca2+-ATPase [S]. In contrast, dantrolene has been reported to inhibit mobilization of Ca2’ from the endoplasmic reticulum and thus reduce cytosolic Ca2’ concentrations and Ca2+-dependent hormone secretion [9-131. If Ca2’ mobilization from the endoplasmic reticulum is functionally coupled to basal RLX secretion, then it is logical to predict that thapsigargin and dantrolene would stimulate and inhibit RLX secretion. respectively. The rate of RLX release was monitored by use of a RLX-reverse hemolytic plaque assay, a technique that employs antibody-directed, complement-
CELL CALCIUM
mediated lysis of target erythrocytes around secretory LLCs (cultured in a layer one cell thick) in order to monitor RLX release from individual LLCs. This cell-by-cell approach was selected to identify potential individual cell variations in functional aspects of Ca2’ mobilization.
Materials and Methods Full details of the collection of tissue, cell dispersion and plaque assay methodology are provided elsewhere [l, 141, and only a brief description is given here. Animals Ovaries from pregnant pigs (day 30-40 of gestation) were collected within 30 min of death at the Meat Laboratory, Iowa State University, USA. Gestational age was determined by the day of mating by an intact boar and confirmed by fetal crown-rump length [15]. Both ovaries were placed in medium and transported back to the laboratory within lo-60 min. The medium used in this step and all subsequent steps in cell dispersion and assays was Dulbecco’s Modified Eagles Medium (DMEM) obtained from Grand Island Biological Co., (Gibco, Grand Island, NY, USA) and supplemented with antibiotics (100 U/ml penicillin G and 100 @ml streptomycin sulfate; Gibco) and 0.1% bovine serum albumin (fraction V, Sigma, St Louis, MO, USA). This medium will be referred to hereafter as DMEM-0.1% BSA. Cell dispersion Three to four individual corpora lutea (CL) were incised, the luteal tissue gently enucleated from the fibrous capsule and then washed 2-3 times with fresh DMEM-O.18 BSA. Approximately one quarter of each CL was then placed in a sterile 65 mm Petri dish containing 3 ml of DMEM-0.1% BSA, minced finely with sterile scalpel blades in a laminar flow hood, and washed repeatedly with fresh medium. The luteal fragments were then placed in a Spinners suspension flask (Bellco, Vineland, NJ, USA) containing 15 ml DMEM-0.1%
REVERSE HEMOLYTIC PLAQUE ASSAY & RELAXIN BASAL SECRETION
BSA, 0.12% collagenase (type III; Cooper Biomedical, FEehold, NJ, USA). The tissue was then incubated for 60 min after which time the fragments were drawn in and out of a sterile 10 ml pipette, incubated for a further 10 min and then centrifuged at 1500 g for 10 min. The cells were then washed twice in Spinners Minimum Essential Medium and passed through a 75 pm mesh (Tetco Inc., Elmsford, NY, USA) to remove cell clumps. An aliquot of the mixed cell suspension was taken for LLC cell counts (using a hemocytometer) and cell viability (trypan blue exclusion). Yields were normally about l-2 x lo6 LLCs, and cell viability was 80-95%. The mixed luteal cell suspension was resuspended in DMEM-0.1% BSA at a final concentration of 0.25 x lo6 LLCs/ml. Reverse henwlyticplaque assay (RHPA) Incubation chambers (Cunningham chambers) were constructed by attaching a glass coverslip to a polyL-lysine treated glass microscope slide using doublestick tape. The mixed luteal cell suspension was then mixed 1:l with a 12% suspension of sheep erythrocytes to which Protein A (Sigma) had been chemically attached [16]. This mixture was then infused into each chamber by capillary action and incubated for 45 min (5% CO2/95% air humidified atmosphere) to allow cell attachment. Each chamber was then flushed with fresh medium to form a layer one cell deep on the chamber floor, and then incubated overnight in the presence of DMEM0.1% BSA. The monolayers were then washed again with fresh medium in order to remove cell products that had accumulated overnight (and dead or dying cells which normally detach), and immediately filled with fresh medium containing porcine RLX antiserum (1:80) and thapsigargin or dantmlene sodium. Dantrolene was dissolved in methanol to a 1 mM stock solution and stored at -20°C. Thapsigargin was dissolved in dimethyl sulfoxide (DMSO) and stored as a 1 m.M stock solution at -20°C. Each agent was diluted to the appropriate concentration immediately before experimentation, and the maximal concentration of methanol or DMSO in media was 0.1%. The same concentration of vehicle was added to controls. We have reported previously that these dilutions of
573
vehicle do not affect the rate of plaque formation [l]. After infusion of antibody containing agents or control vehicle, the monolayers were incubated for periods of 1.2, 3,4, 8 and 12 h. We have shown in previous studies that maximal plaque formation is achieved by 8-12 h of incubation [l, 141. At the end of each incubation period, each chamber was filled with medium containing guinea pig serum as a source of complement (1:40, Gibco), incubated for 50 min to complete plaque formation, fixed with 8% glutaraldehyde in normal saline, and stored at 4°C immersed in the same fixative in airtight containers. Each monolayer was then examined microscopically to score the percentage of LLCs that formed plaques; at least 200 LLCs per monolayer were counted, and triplicate chambers were prepared (for each treatment or control) at each timepoint. Percentages from the triplicate monolayers were averaged. Small luteal cells (i.e. cells less than 15-18 pm in size) or other luteal cell types were not counted since we have previously established that only LLCs form plaques in the RLX-RKPA [l, 141. In addition, the average size of plaques in each monolayer was measured by superimposing a microscopic image of each monolayer over a digitizing pad and drawing the exact shape of 100-200 plaques/monolayer. Image analysis software (SigmaScan; Jandel Scientific, Sausalito, CA, USA) was used to record, compute and analyze the resulting plaque areas (after calibration). Values obtained from triplicate monolayers were combined and averaged. Drugs and reagents All drugs and reagents were purchased from Sigma. Ca2+-freeand &+-replete medium Medium referred to as Ca2’-free was Minimum Essential Medium with Earle’s salts (MEM; contains no added Ca2’*, Gibco, #320-1380). 0.1% bovine serum albumin (BSA, fraction V, Sigma) and contained 0.2 mM EGTA to chelate adventitious Ca2’. The Ca2+-replete medium was DMEM-0.1% BSA. Both media were supplemented with antibiotics (100 U/ml penicillin G and 100 pg/ml streptomycin sulfate; Gibco). Analysis of the [Ca2+]; of three
CELL cAr.cluM
574 80
=cn
70
z F .E
60
0
Control
v
Thapsigargin
(+Co2+)
(1nM)
v
Thapsigargin
(10nM)
q
Thapsigargin
n
Thapsigargin
*ib 50 Q, is CJ a
40
z
30
Fk 0 ii
20
,o n"
10
0 1
0
2
3
4
5
6
Incubation
7
8
Time
9
10
11
12
13
(h)
Fig. 1 Rate of basal plaque formation in monolayers exposed to calcium-replete medium alone, or the same medium containing graded concentrations of thapsigargin (1 nM - 1 @I)). and infused into micmincubation medium, and the monolayers containing thapsigargin. normal saline).
Monodispersed luteal cells were combined with Pmtein A labelled sheep erythmcytcs
chambers to form monolayers.
After ovemight incubation, the chambers were washed with fresh
then incubated for 1-12 h in the presence of control medium (containing vehicle alone) or medium
Complement was then infused to develop the plaques and the. monolayers fixed with glutaraldehyde (8% in
The data shown arc the mean f SEM from three independent experiments using cells obtained from day 32-36 pmgnaat
pigs. Values denoted by an asterisk am significantly different from controls
(P< 0.05) at the same timepoint. See text for fall details of
the reverse hemolytic plaque assay
separate samples of both media from separate batches by atomic absorption spectra- photometry provided values of 1.6 f 0.1 mM and 6.6 + 0.9 pM [Ca2+]i (mean Z!ISD) in DMEM-0.1% BSA and MBM-0.1% BSA, 0.2mM EGTA, respectively. Controlprocedures Specificity of the porcine RLX antiserum used here has been reported previously [17, 181. Plaque formation was abolished by pre-absorption of RLX
antibody (1:80) with 10 pg/ml puritied porcine RLX and by omission of antibody or complement.
Expression of results Results am expressed as mean St SEM. Differences between control and treatment groups were evaluated by analysis of variance (ANOVA) followed by Newman-Keul range test. A P value less than 0.05 was selected as significant 1191.
575
REVERSE HEMOLYTIC PLAQUE ASSAY & RJXAXIN BASAL SECRETION
of plaque-forming LLCs in monolayers treated with 100 nM thapsigargin was significantly greater than Eflect of inhibitorsof Ca2’ redistributionon the rate control values at 3 and 4 h of incubation, but not at of plaque formation - thapsigargin the earliest stages of incubation (namely, 1 and 2 h of incubation). There was no significant difference between Monolayers were incubated in a Ca2+-replete control and any treatment groups at 8 h of incubatmedium containing graded concentrations (1 nM 1 @I) of thapsigaxgin. Concentrations of 1 and 10 ion (P > 0.05). By 12 h of incubation, all groups nM thapsigargin exerted no significant effect on the had clearly attained maximal plaque formation rate of plaque formation (Fig. 1). However, the (60-65% of all LLCs). This apparent loss of effect of thapsigargin occurs because all RLX-releasing percentage of plaque-forming LLCs in monolayers treated with 1 pM thapsigargin was significantly LLCs will ultimately secrete the threshold amount greater than control values at l-8 h of incubation. of RLX required to form a plaque, regardless of treatment. Interestingly, it was apparent that For example, there were nearly twice as many plaque-forming LLCs in thapsigargin-treated monothapsigargin did not recruit additional LLCs into the layers (43 rtr 4%) than in control monolayers (23 z!z secretory population; note that the percentage of 5%) at 4 h of incubation. LLCs in monolayers treated with 100 nM or 1 pM An intermediate dose of thapsigargin (100 nM) thapsigargin was virtually identical at 8 and 12 h of elicited a time-dependent response. The percentage incubation (Fig. 1). Results
80
i O Control n
Co7trol
o
Thopsigargin
(1,uM;
+Ca2+)
q
Tiapsiqorqin
(lr*.M;
-Ca2+)
70 In =
(+CaZ+) (-Coz’)
z 60 CT .-c
E b 7a, 3 g
50
40
a “0 : 0 L : & CL
30
20
10
0
I0
1
2
3
4
5
6
Incubation
7
8
Time
9
10
11
12
13
(h)
Fig. 2 Rate of plaque formation in monolayers exposed to calcium-replete or calcium-free medium, or the same media containing 1 @I thapsigargin. The experimental paradigm was identical to that dcscribcd for Figure 1. The data shown are the mean f SEM from three independent experiments using cells obtained from day 32-36 pregnant pigs. Values denoted by au asterisk are significantly different from controls (P < 0.05) at the same timepoint; thcrc was no sign&ant effect of lack of calcium under basal or thapsigargin-modulated conditions
CELL CALCIUM
576
On the basis of these data, a concentration of 1 @I thapsigargin was selected for the next series of experiments, in which we compared the effect of thaf$gargin on RLXz+secretionby LLCs bathed in Ca -replete and Ca -free medium Under basal conditions, the percentages of plaque-forming LLCs in monolayers bathed in Ca2+-repleteand Ca2+-free medium were not significantly different at any stage of the incubation (Fig. 2). Similarly, the percentages of plaque-forming LLCs in monolayers bathed in Ca2+-replete and Ca2’-free medium containing 1 pM thapsigargin were not significantly different at any stage of the incubation, although each was significantly greater than corresponding control values at l-4 h of incubation (Fig. 2). To verify the stimulatory effect of thapsigargin on basal RLX secretion, the average areas of plaques (an index of cumulative RLX secretion) were measured in the monolayers of the preceding experiments. The average area of plaques provides
0
Cortrol
direct index of the cumulative amount of RLX released over time [20]. It was found that thapsigargin (1 @I) significantly increased (P < 0.05) the average area of plaques by 24fold over the course of the experiment (Fig. 3). confirmation of a strong stimulatory effect of thapsigargin on RLX production. The absence of Ca2’ in the medium did not significantly affect the amount of RLX secreted under basal conditions, or in response to thapsigargin. a
Eflect of inhibitorsof Ca2+redistributionon the rate of plaque formation - dantrolene
To ascertain the effect of dantrolene (an inhibitor of Ca2’ release from endoplasmic reticulum) on the rate of basal RLX secretion, monolayers were first incubated in the presence of graded concentrations of dantrolene (1, 25 and 100 @4) or medium containing vehicle alone (control). We found that a *
(+co2+)
0 Cortrol (-Co”)
C
n
Thopsigorgin
(1pM;
+Ca+‘)
0
Thoxiga-gin
(1pM;
-Ca+*)
12
3
4
Incubation
5 Time
6
7
8
9
(h)
Fig. 3 Mean ama of plaques (cumulative RLX release) in monolayers exposed to calcium-replete or calcium-free medium (controls) or the same media containing thapsigargin (1 pM) during a timecourse plaque assay.
The experimental paradigm was a~ described for
Figure 1. Values am the mean f SEM of thme independent experiments involving luteal cells derived from day 32-36 pregnant pigs, and the symbols marked by an asterisk are significantly different (P C 0.05) from control values at the same timepoint of the incubation; them was no significant effect of lack of calcium under baaal or thapsigargin-modulated not be included because of confluence of plaques at this stage
conditions.
Data from the 12 h timepoint could
577
REVERSE HEMOLYTIC PLAQUE ASSAY & RELAXIN BASAL SECRETION
0
60
V
0
1
Control Dontrolene
2
3
(1pM)
4
5
6
incubation
7 Time
8
9
10
11
12
13
(h)
Fig. 4 Rate of plaque formation in monolayers exposed to control medium or graded concentmtions of dantmlene. The data shown are the mean i SEM (triplicate monolayers) from a single representative experiment using luteal cells obtained from a day 39 pregnant pig. Values denoted by an asterisk are significantly d&rent (P< 0.05) from controls at the same timepoint
dose of 1 p.M dantrolene did not significantly affect the rate of plaque formation. However, doses of 25 and 100 @I dantrolene induced a significant reduction (P < 0.05) in the percentage of plaqueforming LLCs present at l-8 h of incubation (Fig, 4). For the reasons alluded to above, this inhibition was no longer present at 12 h of incubation, when maximal plaque formation (about 55%) was reached in all monolayers. Since there was no detectable difference between the inhibitory effects of 25 and 100 pM dantrolene, a concentration of 25 @I dantrolene was selected for use in the next experiments. To verify and quantify the inhibitory effect of 25 pM dantrolene on RLX secretion, we next compared the size of plaques in control and dantrolene-treated monolayers (Fig. 5). The mean area of plaques in dantrolene-treated monolayers over the total course of the experimental incubation was 55.3 f 4.0% of control values, demonstrating that exposure to dantrolene reduced RLX secretion to about half of control values.
Discussion We have shown that while secretion of RLX by porcine LLCs is a Ca2+-dependent phenomenon, neither basal nor stimulated hormone secretion was dependent on a supply of extracellular Ca2+ [l]. Consequently, we concluded that the primary source of Ca2’ coupled to RLX secretion is mobilization from an intracellular compartment, rather than influx of extracellular Ca2’. The goal of the present study, therefore, was to confii this conclusion and, furthermore, to identify the intracellular source responsible for storing, releasing and presumably In other sequestering Ca2+ ions within LLCs. secretory cell types, the endoplasmic reticulum is considered to be the primary Ca2+ store. Our strategy was to use specific pharmacological modulators of Ca2+ release from this organelle, and to ascertain their effect on RLX production. To monitor RLX release, we employed a reverse hemolytic plaque assay that detects the release of hormone from individual LLCs in culture. The rate
CELL CALCIUM
578
of plaque formation in this approach, together with measurement of average plaque areas, provides information on inhibitory and stimulatory secretagogues (see Neill et al. [20], for an extensive review). This method was employed because we wished to examine cell-by-cell responses. The results of the present study demonstrate that thapsigargin induced a strong, dose-related increase in the rate of RLX secretion. Since it is believed that thapsigargin increases cytoplasmic Ca*+ concentrations by specitically inhibiting a Ca*+sensitive ATPase in the endoplasmic reticulum 581, these data imply that mobilization of Ca*’ from the endoplasmic reticulum is of primary importance to RLX production. Our further observation that removal of extracellular Ca*’ failed to abrogate the stimulatory effect of thapsigargin on RLX secretion is consistent with this view, and confirms our earlier inference that extracellular Ca*+ is not required for basal or modulated RLX secretion [l]. It is worth emphasizing, moreover, that the experiments of the current study were carried out wholly under basal (non-stimulated)
conditions, and contrast with a preceding study of hormone production (progesterone) by luteal cells, in which only LH-enhanced hormone production was reported to be sensitive to thapsigargin [6]. The marked effect of thapsigargin on FUX secretion suggests that tonic efflux of Ca*+ from the endoplasmic reticulum is likely to be of fundamental importance to basal RLX production. It was pertinent that the effect of 100 nM thapsigargin on plaque-formation was strikingly time-dependent (Fig. 1). Although an obvious explanation for the delay in functional response is that thapsigargin required l-2 h to penetrate LLCs and elicit a response, a higher dose of thapsigargin (1 pM) was immediately effective. Furthermore, 100 nM thapsigargin was reported to increase Ca*’ mobilization in rat luteal cells within seconds [6]. Thus, a more likely explanation is that of differential responsiveness to thapsigargin by individual LLCS. Put briefly, it is possible that 100 nM thapsigargin represents a sub-effective dose for some LLCs, specifically those which nounally form plaques early in the incubation period (i.e. those that
30
x NE ,s H “,
24 22 20 18 16 14
0
12
3
4
Incubation
Ftg.5
Mean area of
plaques(cumulative
5 Time
6
8
9
(h)
RLX release) in monolayers exposed to medium
dantrolene (25 pM) during a timecourse plaque assay.
7
plusvehicle
(control) or medium containing
Values are the mean f SEM of three independent experiments involving luteal
cells derived from day 31-44 pregnant pigs, and the symbols marked by an asterisk are significantly different (P < 0.05) from control values at the same timepoint of the incubation.
Where no vertical enor bar is shown, the value is less than the size of the symbol.
from the 12 h timepoint could not be included because of confluence of plaques at this stage
Data
REVERSE HEMOLYTIC PLAQUE ASSAY & RELAXIN BASAL SECRETION
release RLX at the highest rates). On the other hand, the same concentration of thapsigargin (100 nM) clearly increased the secretory abilities of other LLCs that secreted RLX at lower rates (and which form plaques in the midpart of the incubation). Differential responsiveness to thapsigargin by individual LLCs implies that cell-to-cell differences in the amount or activity of the endoplasmic reticulum-bound Ca2+-ATRase may exist. It is tempting to speculate that these differences, if translated into cell-to-cell variations in internal Ca2+ mobilization, may provide one mechanistic explanation for the remarkable heterogeneity of basal RLX secretion exhibited by individual LLCs [l, 141. It was also of interest that thapsigargin failed to elicit any RLX secretion from a fraction of LLCs (non-secretory). We can provide no obvious explanation for this phenomenon, although it is possible that these LLCs were selectively damaged by dispersion procedures. We have shown, however, that nonsecretory LLCs contain abundant immunoreactive RLX and express the RLX gene [21], characteristics that do not favor cell damage. Moreover, non-secretory cells exist in several other endocrine cell types [20], suggesting that the existence of non-secretory subpopulations among It is secretory cells is a general phenomenon. possible that thapsigargin failed to elevate cytosolic Ca2+ concentrations in non-secretory LLCs, or alternatively, that Ca2+ mobilization is uncoupled from RLX secretion in this subpopulation of LLCs. Thapsigargin arachidonic stimulates acid production and PGE:! production in non-luteal cells [22], and we cannot rule out the possibility that these or other biochemical changes may have occurred in LLCs in the present study. Nevertheless, we used a second and independent method to influence Ca2+ redistribution from the endoplasmic reticulum Dantrolene is a compound that reportedly reduces Ca2+ efflux from the endo- plasmic reticulum and has been used, therefore, to examine Ca2+-sensitive, in vitro hormone secretion 19-111. In the present study, dantrolene significantly slowed the rate of basal RLX secretion in a dose-related manner. In contrast to thapsigargin, we observed no timedependency of RLX inhibition, although doses intermediate between 1 and 25 @VI (minimal and maximal concentrations, respectively), must be
579
tested to verify this result. As judged by average plaque area (an index of cumulative RLX secretion), a concentration of 25 pM dantrolene reduced the overall amount of RLX produced by about one half. This result is entirely consistent with the effect of thapsigargin on RLX secretion, and strengthens our conclusion that mobilization of Ca2+ from the endoplasmic reticulum is implicated in basal RLX production. In summary, the evidence presented in the current study is consistent with the view that the endoplasmic reticulum serves as a primary storage site of Ca2+ within LLCs that is functionally linked to RLX production. and serves to confirm that extracellular Ca2+ IS ’ not essential for RLX production. Of course, other potential storage sites, such as calciosomes [23], may also participate in Ca2+ handling. Secretion of LH by rat gonadotropes [24] as well as progesterone secretion by granulosa cells [251 and luteal cells [6] were also attributable (at least partially) to redistribution of Ca2+ from an intracellular source. Nevertheless, these prior studies all involved the analysis of secretagogueinduced hormone release rather than basal secretion. The current study appears to be the first to implicate mobilization of intracellular Ca2’ stores with basal hormone secretion, and thapsigargin will be a highly useful tool to investigate this novel phenomenon in more detail, particularly with respect to heterogenous RLX secretion.
Acknowledgements This work was supported by Grant HD-22786 from the National Institutes of Health, USA. We thank Dr David Sherwood. University of Illinois, Urbana-Champaign, USA, for the generous donation of RLX antiserum.
References Taylor MT. Clark CL. (1989) Effect on calcium ionophores and calcium-channel blockers on relaxin release from cultured porcine luteal cells. Endocrinology, 123, 1893-1901. Bet-ridge MJ. (1987) Inositol phosphate and diacylglycerol: two interacting second messengers. Annu. Rev. Biochem., 56, 159-193. Chegini N. Ramani N. Rao CV. (1984) Morphological and
580
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
CELL CALCIUM biochemical characteristics of small and large bovine luteal cells during pregnancy. Mol. Cell Endocrinol., 37,89-102. Niswender GD. Schwa11RM. Fitx TA. Farin CH. Sawyer HR. (1985) Regulation of luteal function in domestic ruminants: new concepts. Rec. Rag. Horm. Res., 41, 101-150. Hakii H. Fujiki H. Saganuma M. et al. (1986) Thapsigargin, a histamine secretagogue. is a non-12-0-tetradecanoylphorbol-13-acetate type tumor promotor in two stage mouse skin carcinogenesis. J. Clin. Cancer Res. Clin. Oncol., 111.177-187. Pepemll JR. Behrman HR. (1990) The calcium-mobilizing agent, thapsigargin, inhibits progesterone production in rat luteal cells by a calcium-independent mechanism. Endocrinology, 127, 1818-1824. Jackson TR. Paterson SI. l’hastrup 0. Hanely MR. (1988) A novel tumor-promotor, thapsigargin, transiently increases cytoplasmic free Ca’+ without generation of inositol phosphates in NGllS-401L neutonal cells. B&hem. J., 253,81-86. Thastrup 0. Cullen PJ. Drobak BK. Hanely MR. Dawson AP. (1990) Thqsigargin, a tumor pmmotor, discharges intracellular Ca’ stores b specific inhibition of the endoplasmic reticulum Ca5 -ATPase . Pmt. Natl. Acad. Sci. USA, 87.2466-2470. Winkle WBV. (1976) Calcium release from skeletal muscle samoplasmic reticulum: site of action of dantmlene sodium. Science, 193, 1130-1131. Comt PM. Staley D. Jimtah H. Bates M. (1985) Molecular mechanisms of gonadotropin releasing hormone action. J. Steroid B&hem., 23.703-710. Lihtmann I. Delarue C. Homo Delarche F. FeuilloIey M. Belanger A. Vaudry H. (1987) Effects of TMB-8 and dantrolene on A(JTH and angiotensin-induced steroidogenesis by frog interrenal glands: evidence for a role of inttacellular calcium in angiotensin action. Cell Calcium, 8,269-282. Metx SA. (1988) Exogenous arachidonic acid promotes insulin release from intact or permeabilixed rat islets by dual mechanisms. Putative activation of Ca” mobilization and protein kinase C. Diabetes, 37, 1453-1469. Veldhuis JD. May W. Juchter D. (1987) Mechanisms subserving hormone action on the ovary: role of calcium ions as assessed by steady state calcium exchange in cultured swine gramtlosa cells. Endocrinology, 120, 445-449. Taylor MJ. Clark CL. Frawley LS. (1987) Analysis of relaxin release from cultured porcine luteal cells by a reverse hemolytic plaque assay: influence of gestational age and prostaglandin Fh. Endocrinology, 120,2085-2091. Evans HE. Sack WO. (1973) Prenatal development of domestic and laboratory animals: growth curves, external
16.
17.
18.
19. 20.
21.
22.
23.
24.
25.
features and selected references. Anat. Hist. EmbryoI., 2, 1l-20. Neil1 JD. Frawley LS. (1983) Detection of hormone release from individual cells in mixed populations using a reverse hemolytic plaque assay. Endocrinology, 112, 1135-I 137. Sherwood OD. (1982) Radioimmunoassay of relaxin. In: Anderson RR. ed Proceedings of the 15th Midwest Conference on Endccrinology and Metabolism - Relaxin. New York, Plenum Press, pp. 315-321. Shetwood OD. Rosentreter KH. Birkhimer ML. (1975) Development of a radioimmunoassay for porcine relaxin using a ‘%abelled polytytosyl-relaxin. Endocrinology, 96, 1110-1115. Steel RGD. Torrie JH. (1960) Principles and Procedures of Statistics. New York, McGraw-Hill. Neil1 JD. Smith PF. Luque EH. Toro MM. Nagy G. Mulchahey JJ. (1987) Detection and measurement of hormone secretion from individual pituitary cell. Rec. Prog. Horm. Res.. 43.175-229. Bagnell CA. Ohleth K. Clark CL. Taylor MJ. (1992) Detection of mlaxin secretion and gene expression in individual porcine htteal cells. In: Lueng PCK. ed. Molecular Basis of Reproductive Endocrinology. New York, Springer Verlag. Ohuchi K. Suguwata T. Watanabe M. et al. (1987) Stimulation of amchidonic acid metabolism in tat peritoneal macrophages by thapsigargin a non-(12 0-tetradecanoyl phorbol-13-acetate) (TPA)-type tumour ptomotor. J. Cancer Res. Clin. Gncol., 113.319-325. Krause KH. Pittet D. Volpe P. Pozzan T., Meldolesi J. Lew DP. (1989) Cal&some, a sarcoplasmic reticulum-like organelle involved in intracelhtlar Ca2’-handling in non-muscle cells: studies in human neutrophils and HL-60 cells. Ceil Calcium, 10, 351-361. Shangold GA. Murphy SN. Miller RJ. (1988) Gonadotropin releasing hormone-induced Ca” transients in single identified gonad0 s requires both intracellular Ca2+ mobilization and m?e Ca + influx. Pmt. Nat]. Acad. Sci. USA, 85.6566-6570. Lee HL. Shangold GA. Larsen AL. Schreiber JR. (1990) The role of exogenous calcium for gonadotropin-stimulated progeatemne production by human granulosa-luteal cells. Fert. Steril., 52.958-964.
Please send reprint requests to : DC Michael J. Taylor, Department of Veterinary Physiology and Pharmacology, College of Veterinary Medicine, Iowa State University, Ames IA 50011, USA Received : 14 November 1991 Accepted : 9 April 1992
