INTRODUCTION

Cytosolic Ca²⁺ is an important regulator of many cell functions. Increases in cytosolic Ca²⁺ occur by release of Ca²⁺ from intracellular stores located in the endoplasmic reticulum. There are at least two types of Ca²⁺ channels that regulate release from Ca²⁺ stores: an inositol 1,4,5-trisphosphate (IP₃)¹-sensitive channel (IP₃ receptor) and a ryanodine-sensitive channel (ryanodine receptor). As its name implies, IP₃ opens the IP₃ receptor and is used as a second messenger in many cells to release Ca²⁺. The ryanodine-sensitive channel (RyR) is responsible for calcium-induced calcium release (CICR).

There is solid evidence for CICR in the pancreatic acinar cell (1, 2), in addition to some of the first evidence for IP₃ signaling (3). CICR is important in the acinar cell during submaximal agonist stimulation. Low doses of cholecystokinin or carbachol induce Ca²⁺ waves and oscillations (4). In the two pool model of oscillations, IP₃ acts as an initiator of Ca²⁺ release triggering CICR, generating the Ca²⁺ spike (5). Discovery of regulators of CICR are important for full understanding of Ca²⁺ signaling.

There is a growing list of RyR regulators including, caffeine, cyclic ADP-ribose, procaine, spermine, and long chain acyl-CoA derivatives (6). Even though CICR has been established for the pancreatic acinar cell, there is conflicting data regarding the effects of ryanodine and caffeine (classic activators of the RyR). In rat pancreatic acinar cells, there is data supporting activation (7), inhibition (8), and lack of effect (2) of ryanodine and/or caffeine. Unlike ryanodine and caffeine, cyclic ADP-ribose, another activator of the RyR, is naturally occurring in many cell types, but some mammalian cells, including pancreatic acinar cells, are far less sensitive to cADPR than sea urchin eggs (9). Acyl-CoA are also naturally occurring compounds that release Ca²⁺ (10), but there is no data regarding the presence or the effects of acyl-CoA on acinar cell Ca²⁺ movements.

To study the regulation of CICR in pancreas, we tested the effect of acyl-CoA on Ca²⁺ release mechanisms in permeablized rat pancreatic acini. Also the effect of acyl-CoA in relation to those of ryanodine and caffeine was measured.

EXPERIMENTAL PROCEDURES

Rat pancreatic acini were isolated as described previously (11). The rats were sacrificed by CO₂ induced asphyxiation, followed by cervical dislocation. The pancreas was removed and injected with a collagenase, digestion buffer (in mM: 95 NaCl, 6 KCl, 1 MgCl₂, 4 sodium pyruvate, 11 glucose, 2 NaH₂PO₄, 4 sodium fumarate, 5 glutamate, 25 HEPES, 2 CaCl₂, 2 glutamine, also included 20 units/ml purified collagenase, 2 × minimal Eagle's amino acids, and 0.2% (w/v) bovine serum albumin (BSA) (pH 7.4) (NaOH)). The pancreas was incubated 45 min at 37 °C in a shaking water bath (Dubnoff, Precision Scientific, Chicago, IL) in 5 ml of digestion buffer with three changes of buffer. Acini were dispersed by passage through large and small bore glass pipettes and then washed in the digestion buffer without collagenase and with 4% BSA. Large chunks of undigested tissue were removed and the acini were washed again in the 4% BSA buffer. Acini were suspended in buffer A (in mM: 120 NaCl, 20 HEPES, 5 KCl, 10 sodium pyruvate, 10 ascorbate, 10 glucose, 1 MgCl₂, 1 CaCl₂, 1 mg/ml BSA, 10 mg/liter soybean trypsin inhibitor, pH 7.4) and kept at room temperature.

Just prior to experimental use, acini were washed in buffer B (in mM: 100 KCl, 20 NaCl, 20 HEPES, 1 MgCl₂, pH 7.2 (KOH)) + 1 mM EGTA, and then permeablized in buffer B + 0.1 mM EGTA and 0.1 unit/ml of streptolysin O (Murex Diagnostics, Norcross, Georgia) for 10 min at room temperature. Acini were then washed once in buffer B, and then washed once in buffer C (buffer B + 3 mM ATP, 10 mM creatine phosphate, 10 units/ml creatine phosphokinase, 10 µM oligomycin, protease inhibitor mixture (5 µg/ml (final) of pepstatin, leupeptin, chymostatin, antipain, and aprotinin)).

The permeablized acini were resuspended in buffer C + 2 µM fura-2. Acini were transferred into a stirred cuvette at 37 °C and Ca²⁺ measurements made in a Shimadzu spectrofluorometer (model RF-1501, Columbia, MD) with E_x = 340 and 380, E_m = 510. Ratiometric calculations of [Ca²⁺] were performed as described previously (12) using a K_D of 224 nM, except for those experiments carried out at 5 °C, where a K_D of 1.5 µM was used.²

The acyl-CoA derivatives were measured as described by Deutsch et al. (13). Briefly, isolated acinar cells were pelleted, resuspended in 100 mM KH₂PO₄ (pH 4.9), and sonicated (20 s). Acyl-CoA derivatives were extracted with 2-propanol and acetonitrile. The acyl-CoA derivatives were pre-purified on an oligonucleotide purification cartridge (Applied Biosystems, Foster City, CA) and eluted with 750 µl of 80% acetonitrile, 20% 25 mM KH₂PO₄ (pH 4.9). This cell extract was then injected onto a Nucleosil C-18, 5-µm HPLC column (100 × 4.6 mm, Altech Associates Inc., Deerfield, IL) at 45 °C. Recovery was monitored by spiking parallel samples with C₁₉-CoA. The elution buffers were: A, 90% 25 mM KH₂PO₄, 10% methanol; and B, 100% acetonitrile. The elution gradient (SP8700 solvent delivery system, Spectra Physics, San Jose, CA) for acetonitrile was a linear increase from 20 to 40% over 2 min, 40% for 3 min, a linear increase from 40 to 75% over 5 min, followed by a linear increase from 75 to 80% over 5 min, with a flow rate of 1 ml/min. The absorbance (@260 nm) was measured (785A Programmable Absorbance Detector, Applied Biosystems) and the peaks integrated (Shimadzu C-R3A). Palmitoyl-CoA and stearoyl-CoA eluted with retention times of 10.5 and 11.1, respectively. The integrated absorbance of authentic acyl-CoAs were used to calculate the nanomoles of acyl-CoA in the cell extract. For measurement of acyl-CoA following stimulation, isolated acinar cells, suspended in solution A, were incubated with stimulant, then pelleted and processed as described above. Times refer to the total time from agonist addition to sonication.

Results are expressed as mean ± S.E., unless otherwise noted. EC₅₀ values were estimated using nonlinear regression to sigmoid curves (Graphpad Prism, Graphpad Software Inc., San Diego, CA). Statistical comparisons were made using Student's t test.

Creatine phosphokinase, BSA, and HEPES were from Boehringer Mannheim (Indianapolis, IN). Creatine phosphate and ATP were from Calbiochem (San Diego, CA). Methanol and 2-propanol were from Fisher Scientific (Pittsburgh, PA). Collagenase was from Worthington Biochemical Corp. (Freehold, NJ). Fura-2 was from Molecular Probes, Inc. (Eugene, OR). All other chemicals were from Sigma.

RESULTS

To study long chain fatty acid-coenzyme A thioester's effect on Ca²⁺ movement in acinar cells, pancreatic acini were permeablized with streptolysin O and Ca²⁺ release measured with fura-2. Fig. 1A shows the effect of the 16-carbon acyl-CoA, palmitoyl-CoA. Increasing doses of palmitoyl-CoA caused more release, followed by saturation of the response. Palmitoyl-CoA (100 µM) added to unpermeablized fura-2-loaded acini had no effect on cytosolic [Ca²⁺] (data not shown). The dose-dependent effect on palmitoyl-CoA was determined by calculating the cumulative dose of palmitoyl-CoA and comparing it to the cumulative Ca²⁺ release (Fig. 1C) (14). The data were fit to a sigmoid curve (r² = 0.92), with an EC₅₀ of 14 ± 1.5 µM and a V_max of 67 ± 3% of total ionomycin releasable Ca²⁺. The curve had a Hill slope of 1.9 ± 0.3, suggesting a cooperative response.

[View Larger Version of this Image (19K GIF file)]

Palmitoyl-CoA could release Ca²⁺ even after IP₃ could no longer release Ca²⁺ (Fig. 1B). The converse was true, IP₃ could release Ca²⁺ after palmitoyl-CoA response was saturated (Fig. 1A). This suggests separate Ca²⁺ pools, sensitive to either palmitoyl-CoA or IP₃.

To further characterize the separation and overlap of these pools, maximal doses of IP₃ and palmitoyl-CoA were added to permeablized acinar cells alone or together and compared with the total Ca²⁺ pool releasable by 2 µM ionomycin. As seen in Table I, 100 µM palmitoyl-CoA released 67 ± 2%, and 6 µM IP₃ released 45 ± 6% of total ionomycin releasable Ca²⁺. When added together these agents released 83 ± 2% of the total pool. Simple calculations show that 17% of the total pool is insensitive to either compound, and 29% is sensitive to both compounds.

Table I. Palmitoyl-CoA (PCaA) and IP3 release Ca2+ from distinct and overlapping pools The peak levels of Ca2+ released from permeabilized acinar cells were measured after addition of maximal doses of either palmitoyl-CoA (100 µM PCoA) or inositol-1,4,5-trisphosphate (6 µM IP3) added individually, or simultaneously (PCoA + IP3). Values are given as a percentage of the total ionomycin releasable pool which was determined by addition of 10 µM ionomycin at the end of the experiment. In order to measure that portion of the pool sensitive to palmitoyl-CoA but not to IP3, 100 µM PCoA was added after the peak response to IP3 was attained (PCoA after IP3). Alternatively, the palmitoyl-CoA sensitive, IP3-insensitive pool was calculated from the difference in the maximal responses achieved after the combined addition of PCoA + IP3 less the response achieved by addition of IP3 alone. Complementary determinations and calculations were made for the IP3-sensitive, palmitoyl-CoA-insensitive pool (sensitive to IP3 only). That portion of the pool sensitive to both PCoA and IP3 was calculated from the difference in the amount released by palmitoyl-CoA added alone less the amount released by PCoA after IP3 addition (the complementary calculation using the IP3 response gave an identical answer). Values are the mean ± SD of three to five individual determinations made on three separate days. Additions (store sensitivity) Ionomycin-releasable Ca2+ Measured Calculated % PCoA (100 µM) 67  ± 2 IP3 (6 µM) 45  ± 6 PCoA + IP3 83  ± 2 PCoA after IP3 (sensitive to PCoA only) 38  ± 3 38 IP3 after PCoA   Sensitive to IP3 only 16  ± 3 16   Sensitive to both IP3 & PCoA 29   Insensitive to either IP3 or PCoA 17

Calculations also predict that 38% of total Ca²⁺ stores are sensitive only to palmitoyl-CoA and should be releasable after depletion of the IP₃-sensitive store. Table I shows that experimentally, 38 ± 3% is, in fact, released by palmitoyl-CoA after IP₃ stimulation.

A similar prediction from the calculations is that 16% of the total pool is sensitive to IP₃ only. Experimentally, IP₃ addition after depletion of palmitoyl-CoA stores released 16 ± 3% of the Ca²⁺ stores. These data support the idea that the pools are distinct but overlapping.

We also determined what percentage of the IP₃ and palmitoyl-CoA Ca²⁺ stores were sensitive to thapsigargin. Thapsigargin irreversibly inhibits the Ca²⁺-ATPase that catalyzes the filling of the Ca²⁺ stores. We found that thapsigargin released 80.2 ± 8.6% of the ionomycin releasable stores. Adding 100 µM palmitoyl-CoA after thapsigargin depletion still caused release of 8% of the ionomycin releasable store. Since in these cells approximately 65% of the total ionomycin-sensitive store was sensitive to palmitoyl-CoA, this data indicates that 12% of the palmitoyl-CoA-sensitive store is thapsigargin-insensitive, while 88% is sensitive. Similar experiments using 6 µM IP₃ indicate that 96% of the IP₃-sensitive store is thapsigargin-sensitive.

To determine if palmitoyl-CoA acts to release Ca²⁺ by inhibiting the Ca²⁺-ATPase, we added thapsigargin just prior to palmitoyl-CoA addition. While thapsigargin caused slow Ca²⁺ release, Ca²⁺ release induced by palmitoyl-CoA was not diminished (data not shown).

To investigate the possible palmitoyl-CoA interactions with IP₃ signaling, we compared the effect of submaximal IP₃ and palmitoyl-CoA, as well as the effect of heparin, which competes with IP₃ and blocks the Ca²⁺ release through the IP₃ receptor. The results in Fig. 2 show that there is no difference between the sum of the Ca²⁺ release due to submaximal doses of IP₃ and palmitoyl-CoA alone versus the Ca²⁺ release when both are added together, the effects were additive. Furthermore, as shown in Fig. 3, 200 µg/ml heparin decreased the potency and effectiveness of IP₃ to release Ca²⁺ (Fig. 3A). However, at the same concentration heparin had no effect on the ability of palmitoyl-CoA to release Ca²⁺ (Fig. 3B). These data suggest that palmitoyl-CoA acts independently from IP₃.

[View Larger Version of this Image (36K GIF file)]

[View Larger Version of this Image (13K GIF file)]

To investigate the effect of the acyl group on Ca²⁺ release, we tested the dose-dependent effect of different chain lengths and saturation of acyl-CoA. At a dose of 70 µM, short chain acyl-CoA derivatives had no or only a small effect on Ca²⁺ release (Fig. 4A). Acyl-CoA derivatives with 12 carbons or longer caused significant Ca²⁺ release. Maximal efficacy occurred with acyl-CoA derivatives of 16 and 18 carbons. Shorter and longer derivatives were less efficacious.

[View Larger Version of this Image (40K GIF file)]

We also compared the potency of the acyl-CoA, using the cumulative dosage method shown in Fig. 1, A and C. The potency of the acyl-CoA derivatives paralleled their efficacy with the lowest EC₅₀ values occurring with 16 and 18 carbon derivatives. Again, the shorter and longer acyl derivatives were less potent (Fig. 4B). The degree of acyl chain saturation had little effect on potency or efficacy. All of the Ca²⁺ releasing derivatives had a similar apparent cooperativity of response, with a Hill slope coefficient of 2. Neither coenzyme A nor palmitate alone at similar concentrations had any effect (data not shown).

To investigate the possibility that palmitoyl-CoA was indirectly effecting Ca²⁺ release, we tested the effect of heptadecan-2-onyldethio-CoA, a non-hydrolysable analogue of palmitoyl-CoA (15). As shown in Fig. 5A, this analogue is at least as potent as palmitoyl-CoA in eliciting Ca²⁺ release. To confirm this finding we also carried out release experiments at 5 °C. Permeablized acini were first equilibrated at 37 °C to load the Ca²⁺ stores, then the temperature was shifted to 5 °C. Addition of palmitoyl-CoA at 5 °C still caused Ca²⁺ release with an EC₅₀ slightly greater than that observed at 37 °C (Fig. 5B). These two experiments suggest that acyl-CoA act directly on the Ca²⁺ stores rather than enzymatically.

[View Larger Version of this Image (13K GIF file)]

To identify the relationship between palmitoyl-CoA-sensitive Ca²⁺ stores and those involved in CICR, we tested the effect of raising the resting [Ca²⁺]. The effect of palmitoyl-CoA (10 µM) could be potentiated by higher resting [Ca²⁺] (Fig. 6A). To control for the possible artifact of increased pool loading at higher resting [Ca²⁺], in parallel experiments, we increased the resting Ca²⁺ and added a maximal dose of palmitoyl-CoA (100 µM) to estimate the total palmitoyl-CoA-sensitive pool. Using this data we found that at [Ca²⁺] = 240 nM ± 40 nM, 10 µM palmitoyl-CoA released 20 ± 10% of the palmitoyl-CoA-sensitive pool. At [Ca²⁺] = 583 nM ± 117 nM, 10 µM palmitoyl-CoA released 50 ± 5% of the palmitoyl-CoA-sensitive pool (p = 0.015). These data suggest an interaction with the CICR channel.

[View Larger Version of this Image (22K GIF file)]

To further explore CICR, acyl-CoA, and their relationship to the RyR, we pretreated permeablized acini with a low concentration of ryanodine (10 µM) and caffeine (10 mM), two activators of the ryanodine receptor. The data in Fig. 6B show that the effect of palmitoyl-CoA was potentiated by pretreatment, with an 89% greater response than the Ca²⁺ increase due to palmitoyl-CoA alone (95% confidence interval of the mean percent increase: 41-137%, n = 5). These data suggest that palmitoyl-CoA is acting through a ryanodine-like channel.

Cyclic ADP-ribose, which has been reported to synergistically interact with palmitoyl-CoA (16) and RyR, had a strictly additive effect in our system. In our experiment 40 µM cADPR was added alone or together with a dose of palmitoyl-CoA which gave a similar submaximal response. The combined effect of palmitoyl-CoA and cADPR was not significantly different from the sum of the responses of the two agents added alone. We could not determine whether or not cADPR caused release from the same Ca²⁺ pool as palmitoyl-CoA since we failed to observe saturation of the cADPR effect at concentrations up to 100 µM (data not shown). In addition, neither high concentrations of ryanodine (100 µM) nor ruthenium red (10 µM), classic inhibitors of CICR, had any effect on the efficacy or potency of palmitoyl-CoA.

To determine whether palmitoyl-CoA or other long chain acyl-CoA might be important in modulating Ca²⁺ stores in pancreatic acini, we determined the apparent concentrations of acyl-CoA normally present in acini. The two acyl-CoA derivatives found were palmitoyl-CoA and stearoyl-CoA (C₁₈-CoA) with estimated concentrations of 2.94 ± 0.41 and 7.23 ± 0.84 nmol/mg of protein, respectively, which total 10.17 ± 2.96 nmol/mg of protein (95% confidence interval). Deeny et al. (17) estimated the free acyl-CoA concentration by determining the amount of cytosolic binding and the K_D for that binding. If we use those equations,³ and assume that 95% of total acyl-CoA is in the mitochondria (17), and the cytosolic volume is 2.6 µl/mg of protein (18), we estimate that the free concentration of acyl-CoA was 3 µM (1-10 µM, 95% confidence interval).

Acyl-CoA levels in acini did not change following stimulation with carbachol (100 µM) for 1, 3, or 10 min, with bombesin (400 nM) for 1 min, or with forskolin (30 µM) for 2 min (n = 2 for each condition).

DISCUSSION

The results above show that palmitoyl-CoA causes rapid dose-dependent Ca²⁺ release from permeablized pancreatic acinar cells, and that the release is both distinct from and overlapping with the IP₃-sensitive Ca²⁺ pool. Heparin had no effect on palmitoyl-CoA-stimulated Ca²⁺ release and at submaximal doses, the effects of palmitoyl-CoA on Ca²⁺ release are additive with those of IP₃, suggesting that the Ca²⁺ release mechanisms are independent of the IP₃R. The effect of acyl-CoA derivatives is moderately specific with maximal effects occurring with hydrocarbon chain lengths of 16-18. Both extravesicular Ca²⁺ and ryanodine/caffeine potentiate the effect of palmitoyl-CoA. The concentration of palmitoyl-CoA in acinar cells is high enough to regulate Ca²⁺ release.

There are two contrary reported acyl-CoA effects on Ca²⁺ movements, increased uptake and increased release. Deeney et al. (17) found that overnight incubation of clonal -cells (HIT cells) in palmitic acid (20 µM) decreased basal cytosolic free [Ca²⁺] from 100 to 60 nM, and 1 µM palmitoyl-CoA added to permeablized -cells decreased steady-state [Ca²⁺] from 80 to 50 nM within 4 min. Rys-Sikora et al. (19), studying a smooth muscle cell line (DDT₁MF-2), found that low dose palmitoyl-CoA (0.1-1 µM) inhibited GTP induced Ca²⁺ release. But, they also found that higher doses (3-30 mM) decreased Ca²⁺ uptake with or without GTP, these results were only briefly noted along with the suggestion of possible nonspecific detergent effects of acyl-CoA. Dawson and colleagues (20) also studied the effect of palmitoyl-CoA on GTP and Ca²⁺ in rat liver microsomes. Using BSA adsorption of free fatty acids and measuring conversion of CoA to palmitoyl-CoA, they conclude that CoA acts through the acyl-CoA derivatives to release Ca²⁺. This work confirmed similar data in liver microsomes and permeablized hepatocytes by Fulceri et al. (21). Their EC₅₀ for acyl-CoA induced Ca²⁺ release of 35-100 µM is somewhat higher than what was found by Bindoli et al. (10) in sarcoplasmic reticulum and higher than that reported here (both EC₅₀ values ~14 µM).

One concern with our results was that the lipids could be releasing the Ca²⁺ by nonspecific detergent effects on the Ca²⁺ pools. Biophysical experiments estimate that the acyl-CoA long chain critical micelle concentration is 20-250 µM (22), depending on conditions (e.g. buffers, ions). But these experiments were done without cell proteins or cell lipids present, so the biophysical determination of critical micelle concentration will not be an accurate measurement of the concentration in cells above which detergent effects must occur. The detergent effect should be determined empirically.

With our data, we argue against nonspecific detergent effects. First, there is reuptake into Ca²⁺ pools following addition of submaximal doses, suggesting adequate membrane integrity (Fig. 6). Second, after addition of saturating doses of palmitoyl-CoA, IP₃ and ionomycin were still able to release Ca²⁺ (Fig. 1, Table I), meaning those pools are still intact. Third, addition of palmitoyl-CoA to whole cells did not cause Ca²⁺ to leak into the cells through a solublized plasma membrane. Finally, from data in the literature (23, 24), detailed electrical recordings in lipid bilayer experiments could be performed in the presence of acyl-CoA (1-100 µM).

Another concern was that palmitoyl-CoA was acting through a metabolite. We found that palmitoyl-CoA was just as potent at low temperature as at high temperature, which suggests that the acyl-CoA are not acting through an enzymatic process. IP₃ has a similar temperature independent effect on the IP₃R (14). The fact that the effect of the non-hydrolyzable analogue can also release Ca²⁺ lends further proof that its action is not through a metabolite, e.g. palmitoylation of an enzyme. Similar effects with non-hydrolyzable analogues were seen previously by Fulceri et al. (29) in skeletal muscle sarcoplasmic reticulum (heptadecan-2-onodethio-CoA), and by Rys-Sikori (19) in cultured smooth muscle (2-oxopentadecyl-CoA). These long chain coenzyme As did not change [Ca²⁺] when added to intact cells and their action was not inhibitable by heparin in permeablized cells, so the acyl derivatives cannot be acting directly on a receptor-IP₃-Ca²⁺ transduction pathway. Finally, the fact that thapsigargin coaddition does not effect the response suggests palmitoyl-CoA does not act by inhibiting the Ca²⁺-ATPase. While Deeney et al. (17) found that palmitoyl-CoA could lower resting [Ca²⁺] and that that effect could be blocked with thapsigargin, Fulceri et al. (21) found no inhibition of the Ca²⁺-ATPase, and Bindoli et al. (10) found increased Ca²⁺-ATPase activity with doses <50 µM.

Acyl-CoA is most likely acting through a ryanodine-like receptor/CICR channel. One hallmark of a RyR/CICR receptor is that ryanodine/caffeine pretreatment can potentiate Ca²⁺ release (6). Our data show that this potentiation is true for the pancreatic acinar cell (Fig. 6). The same dose of palmitoyl-CoA in the presence of ryanodine and caffeine causes almost two times the Ca²⁺ release compared with control stimulation. Ca²⁺ can also potentiate the effects of acyl-CoA, which is consistent with a CICR channel/RyR. Doubling the resting [Ca²⁺], doubled the amount of Ca²⁺ released from the palmitoyl-CoA-sensitive pool.

Previous studies in other cell types also suggest that the action of acyl-CoA is through the ryanodine receptor. Fulceri et al. (21) and Chini and Dousa (16) found that caffeine could cross-desensitize liver and sea urchin egg homogenates, respectively, to the effect of acyl-CoA. In our hands caffeine caused only a small transient response, so we could not deplete the CICR pool and test cross-desensitization. Fulceri et al. (29) in skeletal muscle and Coronado and colleagues (23, 24) in both skeletal and cardiac muscle, found that palmitoyl-CoA could increase ryanodine binding (increased binding is taken to be an indicator of the channel's open state, for review, see Ref. 6), and the acyl derivatives could increase the Ca²⁺ channel open probability. Our data taken, with these previous reports, suggest that the pancreatic acinar cell has a ryanodine-like receptor and that acyl-CoA are acting through that protein.

The most likely role for acyl-CoA in Ca²⁺ signaling is to potentiate CICR. The fact that we do not measure a change of acyl-CoA with stimulation raises the possibility that acyl-CoA is not regulated prior to Ca²⁺ mobilization. But, the two observations that Ca²⁺ can be released from the CICR pool even at the higher resting [Ca²⁺], and that palmitoyl-CoA can release that Ca²⁺, suggest that CICR is desensitized at the higher resting [Ca²⁺] and palmitoyl-CoA can resensitize the pool. A bolder hypothesis is that palmitoyl-CoA is required for CICR, because the act of raising the resting [Ca²⁺] was not sufficient to activate CICR and deplete the Ca²⁺ store. Since we do find acyl-CoA in the pancreatic acinar cell, at apparent concentrations high enough to stimulate Ca²⁺ release, they may play one of these roles.

The acyl-CoA-binding protein is another factor that could be involved. Fulceri et al. (25) have suggested that the acyl-CoA-binding protein·acyl-CoA complex is the mediator of the effect of acyl-CoA on Ca²⁺ release in skeletal muscle. Kolmer et al. (26) has found acyl-CoA-binding protein RNA (also called the diazapam binding inhibitor) in human pancreas. So the regulation of acyl-CoA-binding protein could provide subtle modification of acyl-CoA's augmentation of CICR. Since we have used a permeablized cell system, there may be other factors or protein phosphorylation states that could change the exact role played by acyl-CoA derivatives.

Finally, while we have not measured a change in total acyl-CoA with hormone stimulation, acyl-CoA may be involved in pathologic conditions. Whitmer et al. (27) have found increased long chain acyl-CoA levels in ischemic heart. Others have found that the metabolic disorder palmitoyl transferase II deficiency, a lack of the enzyme that transports acyl-carnitine out of the cytosol and into the mitochondria, causes elevated palmitoyl-carnitine and causes paralysis and muscle pain when triggered by exercise, cold, and fever (23). These results raise the interesting possibility that acyl-CoA and acyl-carnitine are communication signals from metabolism to the signal transduction pathways. More studies would be required to demonstrate this idea and the roles long chain acyl-CoA plays in physiologic and pathologic conditions.