Arachidonic acid inhibits the store-operated Ca²⁺ current in rat liver cells

Abstract

Vasopressin and other phospholipase-C-coupled hormones induce oscillations (waves) of [Ca²⁺]_cyt (cytoplasmic Ca²⁺ concentration) in liver cells. Maintenance of these oscillations requires replenishment of Ca²⁺ in intracellular stores through Ca²⁺ inflow across the plasma membrane. While this may be achieved by SOCs (store-operated Ca²⁺ channels), some studies in other cell types indicate that it is dependent on AA (arachidonic acid)-activated Ca²⁺ channels. We studied the effects of AA on membrane conductance of rat liver cells using whole-cell patch clamping. We found no evidence that concentrations of AA in the physiological range could activate Ca²⁺-permeable channels in either H4IIE liver cells or rat hepatocytes. However, AA (1–10 μM) did inhibit (IC₅₀=2.4±0.1 μM) Ca²⁺ inflow through SOCs (I_SOC) initiated by intracellular application of Ins(1,4,5)P₃ in H4IIE cells. Pre-incubation with AA did not inhibit I_SOC development, but decreased maximal amplitude of the current. Iso-tetrandrine, widely used to inhibit receptor-activation of phospholipase A₂, and therefore AA release, inhibited I_SOC directly in H4IIE cells. It is concluded that (i) in rat liver cells, AA does not activate an AA-regulated Ca²⁺-permeable channel, but does inhibit SOCs, and (ii) iso-tetrandrine and tetrandrine are effective blockers of CRAC (Ca²⁺-release-activated Ca²⁺) channel-like SOCs. These results indicate that AA-activated Ca²⁺-permeable channels do not contribute to hormone-induced increases or oscillations in [Ca²⁺]_cyt in liver cells. However, AA may be a physiological modulator of Ca²⁺ inflow in these cells.

INTRODUCTION

In non-excitable cells, agonists acting at phospholipase-C-coupled receptors generate repetitive rises in free [Ca²⁺]_cyt (Ca²⁺ concentration in the cytoplasmic space) by releasing Ca²⁺ from intracellular Ca²⁺ stores and activating Ca²⁺ influx through plasma membrane Ca²⁺ channels. It has been generally accepted that the major component of the receptor-activated Ca²⁺ entry into the cell is produced by a store-operated Ca²⁺ entry mechanism, which involves activation of SOCs (store-operated Ca²⁺ channels) in response to emptying intracellular Ca²⁺ stores by Ins(1,4,5)P₃ [1]. At the same time, a significant body of evidence suggests that a non-store-operated Ca²⁺ entry may also be activated by phospholipase-C-coupled receptors [2–4]. In A7r5 vascular smooth muscle cells, avian nasal gland cells and m3HEK293 cells (HEK-293 cells stably transfected with human M₃ muscarinic receptor), non-store-operated Ca²⁺ influx stimulated by AA (arachidonic acid) has been suggested to be a major pathway for Ca²⁺ entry activated by physiological concentrations of specific agonists that induce intracellular Ca²⁺ waves [5–7].

AA, a cis-polyunsaturated fatty acid, is a constituent of membrane phospholipids that can be released by cellular phospholipases, particularly by phospholipase A₂ and diacylglycerol lipase [8]. Free AA has been shown to modulate the activity of a number of ion channels, including a range of Ca²⁺-permeable channels [9–11]. There is no single mechanism of action by which AA modulates ion channels. Thus there is evidence that it binds directly to some channel proteins, but, for other channels, it can also have an indirect effect through its metabolites, free radicals and AA-sensitive protein kinases and phosphatases [12–14]. In some cells in which AA has been shown to inhibit SOCs, it has also been shown to activate a specific Ca²⁺ conductance [15]. Patch-clamping of m3HEK293 cells revealed that, in these cells, AA activates a Ca²⁺ current (I_ARC) that is distinctively different from that activated by store depletion (I_SOC) [16]. This current was implicated in mediating Ca²⁺ oscillations activated by low concentrations of carbachol in these cells [17].

Hepatocytes are polarized epithelial cells in which Ca²⁺ oscillations can be activated by a variety of Ca²⁺-mobilizing hormones. They exhibit a Ca²⁺-selective current mediated by SOCs (I_SOC) that has been characterized by patch-clamp recording [18,19]. The hepatocyte I_SOC exhibits many of the characteristics, including high Ca²⁺ selectivity, of I_CRAC studied in lymphocytes and mast cells [18,19]. Indirect evidence indicates that hormone-induced Ca²⁺ oscillations in hepatocytes are maintained by a store-operated Ca²⁺ entry mechanism, as they are inhibited by known blockers of SOCs [19]. On the other hand, there is evidence that Ca²⁺-mobilizing hormones induce AA release [20].

The present study was designed to elucidate the effects of AA on SOCs in liver cells, and to establish whether AA activates a specific Ca²⁺ conductance in these cells. The results indicate that AA, at concentrations within the estimated physiological range, inhibits I_SOC in rat liver H4IIE cells, but does not itself activate any type of membrane conductance in either H4IIE cells or rat hepatocytes. We also show that iso-tetrandrine, commonly used to inhibit activation of phospholipase A₂ by receptors and therefore AA release [21], is a potent blocker of the I_SOC in H4IIE cells.

EXPERIMENTAL

Cell culture

H4IIE cells (A.T.C.C. CRL 1548) were cultured at 37 °C in 5% (v/v) CO₂ in air in DMEM (Dulbecco's modified Eagle's medium; Gibco) supplemented with penicillin (100 units/ml), streptomycin (100 μg/ml), 10 mM Hepes (pH 7.4) and 10% (v/v) foetal bovine serum (complete DMEM) [22]. The cells were subcultured for a maximum of 15 passages. Hepatocytes were isolated from male Hooded Wistar rats by liver perfusion with collagenase, plated on glass coverslips in complete DMEM (as above) and used for patch clamping on the following day [23].

Electrophysiology

Whole-cell patch clamping was performed at room temperature (24 °C) using a computer-based patch-clamp amplifier (EPC-9, HEKA Electronics, Lambrecht/Pfalz, Germany) and PULSE software (HEKA Electronics). The usual bath solution contained 140 mM NaCl, 4 mM CsCl, 10 mM CaCl₂, 2 mM MgCl₂, 10 mM glucose and 10 mM Hepes, adjusted to pH 7.4 with NaOH. The internal solution 1 contained 120 mM caesium glutamate, 5 mM CaCl₂, 5 mM MgCl₂, 1 mM MgATP, 10 mM EGTA and 10 mM Hepes, adjusted to pH 7.2 with NaOH. The internal solution 2 contained 130 mM caesium glutamate, 10 mM CsCl, 0.5 mM CaCl₂, 6 mM MgATP, 1 mM EGTA and 10 mM Hepes, adjusted to pH 7.2 with NaOH. The calculated internal free Ca²⁺ concentration for each internal solution was approx. 100 nM (EQCAL; Biosoft, Cambridge, U.K.). Depletion of the intracellular Ca²⁺ stores was achieved by addition of 20 μM Ins(1,4,5)P₃ (D-myo-inositol 1,4,5-trisphosphate hexapotassium salt; Sigma) to internal solution 1 or 1 μM thapsigargin (Sigma) to the bath solution. Patch pipettes were pulled from borosilicate glass, coated with Sylgard and fire-polished; pipette resistance ranged between 2 and 4 MΩ. In order to monitor the development of I_SOC, voltage ramps between −138 and +102 mV were applied every 2 s, starting immediately after achieving the whole-cell configuration. Acquired currents were filtered at 2.7 kHz and sampled at 10 kHz. Traces presented in the Figures were further digitally filtered at 1.5 kHz. All voltages shown have been corrected for the liquid junction potential of −18 mV between the bath and electrode solutions (estimated by JPCalc [24]). The holding potential was −18 mV throughout. Cell capacitance was compensated automatically by the EPC9 amplifier. AA and tetrandrine were purchased from Sigma; iso-tetrandrine was purchased from Calbiochem.

RESULTS

First, the effect of AA on the membrane conductance of H4IIE rat liver cells was investigated to establish if AA activates a specific Ca²⁺ current. As it has been suggested that pre-activation of capacitative Ca²⁺ entry may inhibit AA-induced Ca²⁺ entry [15], in these experiments, we used an intracellular solution containing 6 mM Mg ATP and Ca²⁺ buffered to 100 nM with 1 mM EGTA. This would prevent the spontaneous development of I_SOC, and would also minimize the contribution of Mg²⁺-regulated non-selective cation current attributed to TRPM7 (transient receptor potential melastatin-7 cation channel) [25]. Addition of 10 μM AA to the bath solution under these conditions failed to activate any current within 10 min of recording (Figure 1A).

Figure 1. AA fails to activate any current in either H4IIE cells or in rat hepatocytes.

Time courses of the inward and outward membrane currents measured at −118 mV (bottom trace) and 82 mV (top trace) in H4IIE cells (n=5) (A) and in rat hepatocytes (n=5) (B). Application of 10 μM AA in the bath is indicated by a horizontal bar. Internal solution 2 (see the Experimental section) was used in the pipette in order to prevent development of I_SOC.

H4IIE cells are likely to have lost receptors that are normally present in rat hepatocytes, as they do not respond to ATP (<100 μM) or vasopressin (G. Rychkov and G. Barritt, unpublished work). It might be argued that AA-activated Ca²⁺ entry requires the presence of functional G-protein-coupled receptors on the plasma membrane, or that AA-activated Ca²⁺ channels themselves have been lost in this particular cell line. Therefore similar experiments were performed on rat hepatocytes in primary culture that are known to generate cytoplasmic Ca²⁺ waves in response to a variety of hormones [19,26,27]. AA (<20 μM) in the bath solution failed to activate any current in rat hepatocytes within 10 min of recording (Figure 1B), indicating that the failure to observe any AA-activated current in H4IIE cells is unlikely to be due to the loss of expression of G-protein-coupled receptors, and that the absence of AA-activated channels is not unique to H4IIE cells.

The fact that AA does not activate Ca²⁺ conductance in liver cells does not imply that it has no role in Ca²⁺ signalling in these cells. AA has been shown to inhibit store-operated Ca²⁺ entry in various cell types [15,28–30]. However, its effects on SOCs in liver cells and on similar CRAC (Ca²⁺-release-activated Ca²⁺)-type SOCs in other cell lines are not known. Therefore, in the next set of experiments, we investigated the effect of AA on I_SOC in H4IIE cells. We have characterized I_SOC in H4IIE cells previously and have shown that it has many of the properties of I_CRAC in lymphocytes and mast cells [18]. In the absence of AA, depletion of intracellular Ca²⁺ stores in H4IIE cells activated an inward current (I_SOC) with a magnitude of approx. −2.5 pA/pF, measured at −118 mV, as described previously [18] (Figure 2A). The current reached its maximum within 60 s after achieving the whole-cell configuration with the pipette containing 20 μM Ins(1,4,5)P₃, and then slowly decayed to a level of 60–70% of the maximal amplitude. Addition of 10 μM AA to the bath solution after full development of the current produced an almost complete block of I_SOC within 150 s. At lower concentrations, the block was slower in onset and less complete. The lowest concentration of AA that produced a significant effect within 200 s of application was 1 μM (Figure 2A). Higher concentrations of AA (20–50 μM) blocked I_SOC faster; however, they frequently caused development of a non-specific leakage and electrical breakdown of the membrane during steps to negative potentials (results not shown). The apparent IC₅₀ for AA measured 200 s after AA application was 2.4±0.1 μM (n=4) (Figure 2B).

Figure 2. Effects of AA on I_SOC in H4IIE cells.

(A) Time course of I_SOC inhibition by AA. Amplitude of I_SOC measured at −118 mV is plotted against time. Application of AA is indicated by a horizontal bar (n=4–9). (B) Dose-dependent inhibition of I_SOC by AA. Each point on the graph is the normalized average amplitude of I_SOC at −118 mV measured after 200 s of application of the corresponding concentration of AA in the bath (n=4–9). (C) Effect of AA on I_SOC development. Cells were either pre-incubated with AA in the bath for at least 2 min before achieving the whole-cell configuration or AA was added to the pipette solution (n=6).

Pre-incubation of H4IIE cells with 10 μM AA in the bath for at least 2 min before achieving whole-cell configuration did not prevent development of I_SOC (Figure 2C). However, the maximal amplitude of the current was significantly smaller and it was completely inactivated within 200 s of recording. The time constants of I_SOC development were 24±2 s (n=6) for the control cells and 20±3 s (n=6) for the cells pre-incubated with AA. These results indicate that it is unlikely that AA interferes with the mechanism of I_SOC activation, or that it inhibits I_SOC by binding to the closed channel. It is more likely that AA either inhibits the open channel directly or modulates the mechanism by which I_SOC is slowly inactivated [31]. Addition of 10 μM AA to the internal solution containing 20 μM Ins(1,4,5)P₃ had no effect on the development or amplitude of I_SOC (Figure 2C).

Under conditions used in the experiments with H4IIE cells described above (Figure 1), AA did not seem to activate any specific conductance. However, in other cell types, AA has been shown to activate highly selective Ca²⁺ channels with properties distinct from that of the SOCs in those cells [16]. One of the main differences between I_SOC and the current activated by AA (I_ARC) in HEK-293 cells is the lack of the fast inactivation in the latter [16]. I_SOC in H4IIE cells also shows a significant fast Ca²⁺-dependent inactivation at negative potentials [32], and therefore can be distinguished easily from a current that shows no fast inactivation. While there was no evidence that AA activated any current in H4IIE cells, the possibility of a transient activation of such a current when I_SOC is blocked could not be excluded. Therefore we compared current traces obtained in response to −138 mV voltage steps before, and 60 s after, application of AA. The kinetics of the I_SOC inactivation and the relative amplitude of the non-inactivating component at negative potentials remained unaffected in the presence of AA (Figures 3A and 3B). This argues against the presence of any non-inactivating Ca²⁺ current additional to I_SOC. Moreover, the I–V plot remained inwardly rectifying in the presence of AA, with no evidence for the development of any outward current or a shift in the reversal potential (which would be expected if any other conductance had developed) (Figure 3C).

Figure 3. AA does not alter the kinetics of fast inactivation or the inward rectification of I_SOC in H4IIE cells.

(A) Current traces were obtained in response to voltage steps to −138 mV before and after application of 10 μM AA in the bath for 60 s. (B) Current traces shown on (A) were normalized to the peak value and superimposed. (C) Current–voltage plots in the absence and presence of AA in the bath.

One of the methods of distinguishing between Ca²⁺ influxes carried by I_SOC and I_ARC in m3HEK293 cells, where I_ARC was first described, was to use iso-tetrandrine. Iso-tetrandrine is known to inhibit the activation of phospholipase A₂ by receptors [21], and has therefore been used to inhibit agonist-induced AA release [6,15]. Iso-tetrandrine is also one of the stereoisomers of tetrandrine that blocks L-type and T-type Ca²⁺ channels [33]. The effects of either compound on store-operated Ca²⁺ entry have not been investigated previously in most cell types. However, in avian nasal gland cells, iso-tetrandrine had no effect on thapsigargin-activated Ca²⁺ inflow [6]. In H4IIE cells, both tetrandrine and iso-tetrandrine caused a dose-dependent inhibition of I_SOC (Figure 4, only tetrandrine is shown). The apparent IC₅₀ for tetrandrine was 8.6±0.3 μM (n=4–12) (Figure 4B). At concentrations of 10 and 100 μM, iso-tetrandrine showed potency similar to that of tetrandrine in inhibiting I_SOC (results not shown). The onset of the block induced by tetrandrine and iso-tetrandrine was as rapid as that observed with La³⁺ (T. Litjens and G. Rychkov, unpublished work) and was easily reversible upon washout (results not shown). Both tetrandrine and iso-tetrandrine had the same effect when I_SOC was activated by either thapsigargin or Ins(1,4,5)P₃.

Figure 4. Inhibition of I_SOC by tetrandrine in H4IIE cells.

(A) Time course of I_SOC inhibition by 100 μM tetrandrine (application is indicated by the horizontal bar). (B) Dose-dependent inhibition of I_SOC by tetrandrine (n=4–12).

DISCUSSION

The existence of AA-activated Ca²⁺ entry has been shown in a variety of cell types [5,6,34,35]. In some of these cell types, AA-activated Ca²⁺ channels provide a major pathway for Ca²⁺ entry during Ca²⁺ oscillations triggered by receptor stimulation [5,17,29]. In contrast, in astrocytes, AA inhibits Ca²⁺ oscillation activated by ATP and activates sustained Ca²⁺ influx [30]. The results of the present study demonstrate that AA, in the predicted physiological range of concentrations [36], does not activate Ca²⁺-permeable channels in H4IIE liver cells, but strongly inhibits Ca²⁺ entry through SOCs (cf. [15,28–30]). The observation that AA also did not activate Ca²⁺-permeable channels in primary rat hepatocytes indicates that the failure to observe an AA-activated current in H4IIE liver cells is unlikely to be due to the absence of receptors for agonists in this immortalized cell line.

These results imply that, unlike some other cell types, including vascular smooth muscle cells and m3HEK293 cells, liver cells do not rely on AA-activated Ca²⁺ entry to maintain Ca²⁺ oscillations generated in response to phospholipase-C-coupled hormones. Moreover, the idea that SOCs are responsible for Ca²⁺ inflow that maintains Ca²⁺ oscillations in liver cells [19] is consistent with the present observations of the absence of AA-activated Ca²⁺-permeable channels in liver cells.

The observation that the intracellular application of AA had no effect on I_SOC development and amplitude suggests an extracellular or membrane-delimited site of action for AA. However, an intracellular site cannot be unequivocally ruled out, since application of a membrane-permeable substance, such as AA, through a patch pipette may not be as effective in maintaining the exogenous AA concentration, since the AA will diffuse into the infinitely large (compared with the cell volume) bath faster than it diffuses through a patch pipette. AA may also be metabolized faster than it is incorporated into the plasma membrane, if the site of action is membrane-delimited. This appears to be the case in studies of the regulation of I_CRAC by sphingosine, another membrane-permeable compound. Sphingosine that accumulates in the membrane, as assessed by the changes in the membrane capacitance, and inhibits I_CRAC in RBL (rat basophil leukaemia) cells when applied in the bath has no effect on either capacitance or I_CRAC when applied through the pipette [37]. Regardless of the site of action in liver cells, AA may be an important modulator of Ca²⁺ entry through SOCs when these are activated by physiological concentrations of hormones that induce the formation of AA. Under physiological conditions, AA is produced in the membrane [8], and, in vivo, when cells are packed close to each other in an intact organ, the concentration of AA is likely to rise both inside and outside the cells.

In m3HEK293 cells, Ca²⁺ currents activated by AA and depletion of Ca²⁺ stores were additive [16]. Thus addition of AA after I_SOC development in m3HEK293 cells increased membrane current further. In contrast, in the present study with H4IIE cells, I_SOC was inhibited completely by similar concentrations of AA. This suggests that I_SOC in HEK-293 cells and I_SOC in H4IIE cells are likely to be mediated by different types of SOC. Moreover, these results emphasize that conclusions about the nature of Ca²⁺-permeable pathways in the plasma membrane for one cell type cannot necessarily be applied to another cell type.

Another important observation of the present study is that iso-tetrandrine and tetrandrine are potent inhibitors of I_SOC in liver cells. The mechanism of inhibition is unlikely to involve an effect of these agents on phospholipase A₂, as the block is very rapid and readily reversible. A direct effect of tetrandrine and iso-tetrandrine on the SOC is consistent with these observations. In addition, if iso-tetrandrine and tetrandrine were acting via phospholipase A₂, they would be expected to enhance I_SOC, not to inhibit it, since the present results show that AA, a product of phospholipase A₂ activity, inhibits SOCs in liver cells. In avian nasal gland cells, iso-tetrandrine has been shown to inhibit Ca²⁺ oscillations activated by a low concentration (0.5 μM) of carbachol, but was found to be ineffective in inhibiting Ca²⁺ entry activated by a high concentration (10 μM) of carbachol or by thapsigargin [6]. It was concluded from these experiments that, in avian nasal gland cells, iso-tetrandrine does not affect SOCs, that SOCs are not activated by low concentrations of carbachol, and therefore that the major physiological pathway for Ca²⁺ entry in these cells is through Ca²⁺ channels activated by AA. In view of the present study, caution should be exercised in interpretation of the results obtained using these compounds as inhibitors of receptor-activation of phospholipase A₂ in studies involving the measurement of SOCs, especially when employing Ca²⁺ fluorescent dyes to measure Ca²⁺ inflow, and when the nature of the channels underlying that inflow is not entirely certain.

The observations made by several research groups with different cell types on the effects of AA on plasma membrane Ca²⁺-permeable channels can be summarized as follows: in different cell types, AA has been shown either to induce Ca²⁺ oscillations [6], or to inhibit them [30]; to inhibit SOC-mediated Ca²⁺ entry [15,28–30], or to have no effect on SOCs [16]; to activate Ca²⁺ channels with a high selectivity for Ca²⁺ [16], or to activate Ca²⁺-permeable non-selective cation channels [38]; and to activate Ca²⁺-permeable channels which are very sensitive to inhibition by Gd³⁺ [34], or channels which are insensitive to Gd³⁺ [5,39,40]. On the basis of all these results, it can be suggested that AA plays different physiological roles in regulating Ca²⁺ entry, and may activate or inhibit different types of Ca²⁺-permeable channels in different cell types. The results also indicate that SOCs are likely to be different in different cell types, and no single model of the mechanism of Ca²⁺ oscillations can be applied to all cell types.