Lipid accumulation in hepatocytes can lead to non-alcoholic fatty liver disease (NAFLD), which can progress to non-alcoholic steatohepatitis (NASH) and Type 2 diabetes (T2D). Hormone-initiated release of Ca2+ from the endoplasmic reticulum (ER) stores and subsequent replenishment of these stores by Ca2+ entry through SOCs (store-operated Ca2+ channels; SOCE) plays a critical role in the regulation of liver metabolism. ER Ca2+ homoeostasis is known to be altered in steatotic hepatocytes. Whether store-operated Ca2+ entry is altered in steatotic hepatocytes and the mechanisms involved were investigated. Lipid accumulation in vitro was induced in cultured liver cells by amiodarone or palmitate and in vivo in hepatocytes isolated from obese Zucker rats. Rates of Ca2+ entry and release were substantially reduced in lipid-loaded cells. Inhibition of Ca2+ entry was associated with reduced hormone-initiated intracellular Ca2+ signalling and enhanced lipid accumulation. Impaired Ca2+ entry was not associated with altered expression of stromal interaction molecule 1 (STIM1) or Orai1. Inhibition of protein kinase C (PKC) reversed the impairment of Ca2+ entry in lipid-loaded cells. It is concluded that steatosis leads to a substantial inhibition of SOCE through a PKC-dependent mechanism. This enhances lipid accumulation by positive feedback and may contribute to the development of NASH and insulin resistance.
Chronic obesity is associated with the accumulation of lipid droplets containing di- and tri-acylglycerol (glyceride) in hepatocytes and the development of a fatty liver [non-alcoholic fatty liver disease (NAFLD)]. If uncontrolled, this can lead to hepatosteatosis and non-alcoholic steatohepatitis (NASH), insulin resistance and Type 2 diabetes (T2D) [1,2]. Moreover, hyperglycaemia associated with T2D can enhance lipid accumulation and steatosis . In normal hepatocytes, hormone-induced increases in the cytoplasmic free Ca2+ concentration ([Ca2+]cyt) play an essential role in the regulation of carbohydrate, lipid and nitrogen metabolism [4,5]. Steatosis leads to the generation of reactive oxygen species, the activation of protein kinase C (PKC) and the endoplasmic reticulum (ER) stress response [6–11]. The ER stress response is associated with altered ER Ca2+ homoeostasis involving decreased [Ca2+]ER (ER luminal Ca2+ concentration) and decreased expression and activity of sarco/endoplasmic reticulum (Ca2+ + Mg2+)ATPase (SERCA)2b [12,13].
Ca2+ entry through SOCs (store-operated Ca2+ channels; SOCE) in the hepatocyte plasma membrane and subsequent re-filling of the [Ca2+]ER stores is necessary for hormone-initiated intracellular Ca2+ signalling . SOCs in animal cells are highly selective for Ca2+ (reviewed in ). The SOC pore is composed of Orai polypeptides located in the plasma membrane that are activated by stromal interaction molecule (STIM) polypeptides located in the ER membrane. STIM senses the decrease in Ca2+ in the lumen of the ER, moves in the plane of the ER membrane, interacts with Orai and opens the channel [14,15]. Three isoforms of Orai (Orai1, 2 and 3) and two isoforms of STIM (STIM1 and 2) have been identified. Whereas many SOCs are comprised of Orai1 polypeptides and are activated by STIM1, the other isoforms of Orai and STIM are also thought to play important, but less prominent roles . Rat hepatocyte SOCs are highly selective for Ca2+ and are principally composed of Orai1 and STIM1 proteins .
The effects of lipid accumulation and steatosis on SOCE and hormone initiated intracellular Ca2+ signalling are not known. Since SOCE plays necessary roles in the hormonal regulation of liver metabolism and in the maintenance of adequate [Ca2+]ER, the aim of the present experiments was to determine whether SOCE is altered in lipid-loaded and in steatotic liver cells and to investigate the consequences for hormonal intracellular Ca2+ signalling and the mechanisms involved.
All experimental animals received humane care, and the experimental protocols were conducted according to the criteria outlined in the “Australian Code of Practice for the Care and Use of Animals for Scientific Purposes” (National Health and Medical Research Council of Australia). Further details of materials and methods are provided in Supplementary Material.
Preparation of isolated hepatocytes and culture of hepatocytes and rat liver cells
Induction and measurement of lipid accumulation
Lipid-accumulation in H4IIE cells and in isolated hepatocytes was induced by 24 h incubation with amiodarone (20 μM, except where indicated otherwise) [19,20] or palmitate (500 μM). Before measurement of [Ca2+]cyt or other parameters, cells were washed to remove amiodarone or palmitate. Fluorescent Nile Red staining was used to measure intracellular lipid accumulation .
Patch clamp recording and measurement of Ca2+ entry using fluorescence imaging microscopy
Real-time quantitative PCR
Total RNA was prepared from freshly isolated hepatocytes or cultured H4IIE cells using TRIzol reagent (Invitrogen) and cDNA was synthesized as described . mRNA was quantified by qPCR (real-time quantitative PCR) using TaqMan™ probe technology, a Rotor-Gene 3000 (Qiagen) and the 2−ΔΔCT method.
Western blot analysis
Protein extraction, SDS-PAGE, Western blotting and densitometry was performed as previously described  using primary antibodies for STIM1, STIM2, Orai1, Orai2, Orai3, and SERCA with β-actin as a loading control.
Unless stated otherwise, data values are expressed as means ± S.E.M. Statistical significance (P<0.05) was determined between groups using one-way ANOVA with Bonferroni post-hoc test, ANOVA followed by Newman–Keuls multiple comparison post-hoc test or Student's t-test (two tailed, unpaired).
Lipid accumulation is associated with a substantial inhibition of Ca2+ entry through SOCs
In order to assess the effects of lipid accumulation on Ca2+ entry, three models of lipid accumulation and steatosis were employed. These were two in vitro models in which H4IIE rat liver cells or rat hepatocytes were incubated with amiodarone or palmitate to induce lipid accumulation [9,24] and the genetically obese Zucker rat in which the liver is steatotic .
In vitro lipid accumulation was induced in H4IIE cells by pre-treatment with amiodarone (20 μM, 24 h) (Figure 1A), as shown previously . To determine whether lipid accumulation is associated with altered SOCE, cells were loaded with fura-2, incubated in the absence of added extracellular Ca2+ ([Ca2+]ext), exposed to the SERCA inhibitor DBHQ [2,5-di-(tert-butyl)-1,4-benzohydro-quinone] to release [Ca2+]ER, then [Ca2+]ext was added to permit Ca2+ entry. Amiodarone pre-treated cells exhibited a substantial reduction in the rate of Ca2+ entry (the initial rate of increase in [Ca2+]cyt following [Ca2+]ext addition) compared with controls (Figures 1B and 1D) and exhibited a substantial decrease in the height of the peak of the DBHQ-induced increase in [Ca2+]cyt in the absence of added [Ca2+]ext, (Figures 1B and 1C). No change in the basal value of [Ca2+]cyt was observed in amiodarone pre-treated cells compared with controls (results not shown). When DBHQ was added to H4IIE cells loaded with lipid by pre-treatment with amiodarone and incubated initially in the presence of [Ca2+]ext, the height of the peak and value of the plateau induced by DBHQ were greatly reduced in lipid-loaded cells compared with controls (Supplementary Figures S1A—S1C). Pre-treatment of freshly-isolated rat hepatocytes with amiodarone also inhibited DBHQ-initiated Ca2+ entry (Supplementary Figure S2). Results similar to those shown in Figures 1A–1D were observed in H4IIE cells pre-treated with palmitate (500 μM, 24 h) instead of amiodarone (Figure 2A–2D).
Lipid accumulation induced by pre-treatment with amiodarone leads to substantial decreases in Ca2+ release from the ER and in Ca2+ entry initiated by the SERCA inhibitor DBHQ
Lipid accumulation in liver cells induced by pre-treatment with palmitate leads to substantial decreases in Ca2+ release from the ER and in Ca2+ entry initiated by the SERCA inhibitor DBHQ
These results indicate that lipid accumulation is associated with a substantial decrease in SOCE and in [Ca2+]ER in liver cells. Patch-clamping confirmed the involvement of SOCE. The characteristics of the current (ISOC) induced by addition of inositol 1,4,5 trisphosphate (InsP3) through the patch pipette, in both lipid-loaded (5 μM amiodarone, 24 h) and control H4IIE cells, were similar to those observed previously for SOCE in H4IIE cells and rat hepatocytes  (result not shown). This indicates that the current induced by InsP3 is highly selective for Ca2+ and principally mediated by Orai1 and STIM1 . In lipid-loaded cells, the maximal amplitude of ISOC was reduced by 70% compared with controls (Figure 1E).
To determine if SOCE and Ca2+ release are decreased by lipid accumulation in vivo, hepatocytes were isolated from obese and lean Zucker rats. As shown previously , obese rats exhibited high blood concentrations of glucose, triacylglycerol (triglyceride) and cholesterol (Supplementary Figure S3). Hepatocytes freshly isolated from obese Zucker rats exhibited substantial decreases in the initial rate of Ca2+ entry and in DBHQ-induced Ca2+ release, compared with those from lean rats (Figures 3A–3C) (cf ). Patch-clamping confirmed SOCE inhibition in steatotic hepatocytes from obese Zucker rats. The amplitude of ISOC, activated by ATP (25 μM), was 60% smaller than that of Hooded Wistar rats (lean controls) (Figures 3D and 3E). To confirm that the current activated in response to extracellular application of ATP is ISOC we replaced Ca2+ and Na+ in the bath solution with 100 mM Ba2+. Current activated by ATP in hepatocytes exhibited an instantaneous increase in the amplitude when Ba2+ was added to the bath, followed by a decline (Figure 3E), which is characteristic of ISOC mediated by Orai1 and STIM1 proteins in different cell types [29–31].
Lipid accumulation in hepatocytes from obese Zucker rats leads to substantial decreases in Ca2+ release from the ER and in Ca2+ entry initiated by the SERCA inhibitor DBHQ or ATP
The possibility that the observed inhibition of SOCE caused by pre-treatment with amiodarone is due to a direct effect of this agent on SOCs in the plasma membrane (cf [32,33]) was tested. H4IIE cells incubated with amiodarone (20 μM) for 10 min, then washed using the same protocol as for 24 h treatment, exhibited no change in DBHQ-induced Ca2+ release or entry compared with controls (Supplementary Figure S4). This observation indicates that the observed inhibition of SOCE induced by amiodarone pre-treatment for 24 h is not due to direct effects of this agent on SOCs.
Lipid-induced inhibition of SOCE substantially decreases hormone-initiated intracellular Ca2+ signalling whereas pharmacological SOCE inhibition enhances lipid accumulation
To evaluate the effects of hormones on [Ca2+]cyt in lipid-loaded cells, H4IIE cells were differentiated by incubation with insulin and dexamethasone to induce expression of plasma membrane hormone receptors . In differentiated cells incubated in the presence of [Ca2+]ext, both the P2Y purinergic agonist, ATP and the α-adrenergic agonist phenylephrine, induced an increase in [Ca2+]cyt, which was characterized by an initial peak followed by a plateau (Figures 4A and 4C). These are typical of hormone-induced responses previously observed in freshly-isolated hepatocytes . However, for differentiated cells pre-treated with amiodarone, both the initial peak and the subsequent plateau were greatly reduced (Figures 4B and 4D). A similar pattern was observed for the actions of ATP and phenylephrine on hepatocytes isolated from obese Zucker rats compared with the response for hepatoctyes from lean rats (Figures 4E–4H).
Inhibition of SOCE by intracellular lipid accumulation decreases hormone-initiated intracellular Ca2+ signalling
Since decreased SOCE may lead to decreased [Ca2+]ER which may affect intracellular lipid accumulation [3,13], we tested the effect of pharmacologically inhibiting SOCE, using 2-aminoethoxydiphenyl borate (2-APB) , on the accumulation of lipids. The results show that 2-APB caused a doubling of the lipids accumulated in 24 h by pre-treatment with amiodarone (Figure 5A) or palmitate (Figure 5B) or with amiodarone in differentiated H4IIE cells (Figure 5C). Similar results were obtained using the more selective SOCE inhibitor, YM-58483/BTP-2 (3,5-bis(trifluoromethyl)pyrazole-2)  in H4IIE cells incubated in the presence of amiodarone or palmitate (Figures 5D and 5E) When the expression of STIM1 and Orai1 in H4IIE cells was knocked down using siRNA an enhancement of lipid accumulation was also observed (Figure 5F).
Pharmacological inhibition of SOCE or knockdown of STIM1 and Orai1 enhances lipid accumulation
Lipid accumulation is associated with a decrease in Ca2+ in the ER
To confirm that lipid-loading reduces the amount of Ca2+ in intracellular stores, cells loaded with fura-2 and incubated in the absence of added [Ca2+]ext were exposed to the Ca2+ ionophore, ionomycin. The height of the peak observed following addition of 15 μM ionomycin and the rate of increase in [Ca2+]cyt following subsequent [Ca2+]ext addition were both reduced in lipid-loaded cells, compared with controls (Supplementary Figures S5A–S5D). Similar results were observed when cells where treated with a higher (50 μM) concentration of ionomycin to ensure maximum permeability of the [Ca2+]ER store (Supplementary Figures S5E–S5H). These results indicate that lipid accumulation is associated with a decrease in the amount of [Ca2+]ER.
Expression of STIM and Orai
To test whether the lipid-induced inhibition of SOCE is due to altered expression of STIM1 and/or Orai1 proteins (SOCE components), STIM and Orai mRNA and protein were measured. Following in vitro induced lipid accumulation in H4IIE cells, qPCR and Western blot analysis revealed there to be no change in mRNA levels encoding STIM1 and 2 and Orai1, 2 and 3 or in protein levels of STIM1 and Orai1 (Figures 6A and 6B). qPCR analysis revealed that lean and obese Zucker hepatocytes express STIM and Orai mRNA in a similar expression pattern to that of H4IIE cells, with no difference in mRNA levels between lean and obese (Figure 6C). The mRNA expression profile of STIM and Orai in lean Hooded Wistar rat hepatocytes was similar to that of lean Zucker rat hepatocyes (Figure 6D cf6C). Western blot detected the STIM1, STIM2 and Orai1, Orai2 (at very low levels) and Orai3 proteins in lean and obese Zucker rat hepatocytes (Figure 6E). Expression of STIM2, normalized to β-actin, was decreased by a small, but significant amount, in steatotic hepatocytes (Figure 6E).
Expression of the isoforms of STIM and Orai in lipid-loaded and control H4IIE liver cells and in rat hepatocytes
Role of protein kinase C in the mechanism of lipid-induced SOCE inhibition
Lipid accumulation in hepatocytes is known to be associated with the activation of several isoforms of PKC [8,9,37] whereas other results indicate that PKC can inhibit SOCE via phosphorylation of Orai1 [38,39]. To confirm that activation of PKC leads to the inhibition of SOCE in liver cells, control (non lipid-loaded) H4IIE cells were incubated with PMA (4 μM) for 15 min to activate PKC . This caused a large inhibition of Ca2+ entry (Figure 7A), comparable with that observed in lipid-loaded cells Figure 1B), thus providing evidence that the activation of PKC can inhibit SOCE. It has previously been shown that the pre-treatment of cells with PMA for 24 h leads to the degradation of PKC isoforms . This strategy was used to test the role of PKC in the lipid-induced inhibition of SOCE. Lipid-loaded cells pre-treated with PMA (4 μM for 24 h) exhibited much greater rates of Ca2+ entry than lipid-loaded cells pre-treated with vehicle, whereas pre-treatment of control cells with PMA for 24 h caused little change in Ca2+ entry (Figures 7B and 7C). The rate of Ca2+ entry in lipid-loaded cells treated for 24 h with PMA was similar to the rates of Ca2+ entry in untreated control cells and in control cells pre-treated for 24 h with PMA (Figure 6B cfFigure 6C) indicating that the down-regulation of PKC can completely reverse the lipid-induced inhibition of SOCE.
Acute treatment of control (non lipid-loaded) liver cells with PMA inhibits SOCE whereas pre-treatment (24 h) of lipid-loaded liver cells with PMA reverses the lipid-induced inhibition of SOCE
To further test whether PKC contributes to SOCE inhibition in lipid-loaded cells, Ca2+ entry was measured in the presence of the PKC inhibitor GF109203X (GFX) . When GFX (20 μM) was added at the beginning of the period over which [Ca2+]cyt was measured (5 min before DBHQ), GFX increased the rate of Ca2+ entry in lipid-loaded (Figure 8A), but not control (Figure 8B), cells. GFX also reduced the height of the peak of the DBHQ-induced increase in [Ca2+]cyt in the absence of added [Ca2+]ext in both lipid-loaded and control cells (Figures 8A and 8B), as shown previously for another cell type . The addition of GFX to control (non lipid-loaded) cells in the absence of DBHQ caused no increase in [Ca2+]cyt, suggesting that if GFX reduces the amount of Ca2+ in intracellular stores, it does so very slowly.
The PKC inhibitor GFX reverses the inhibition of SOCE induced by lipid accumulation in H4IIE liver cells
The addition of GFX to hepatocytes isolated from obese Zucker rats also increased the initial rate of Ca2+ entry and reduced the height of the peak of the DBHQ-induced increase in [Ca2+]cyt (Figure 9A). In hepatocytes isolated from lean Zucker rats, GFX did not alter the initial rate of Ca2+ entry, but did reduce the height of the peak of the DBHQ-induced increase in [Ca2+]cyt (Figure 9B). To confirm the action of GFX on Zucker hepatocytes, cells from lean rats were loaded with lipid by incubation with amiodarone and the effect of GFX tested. GFX increased the initial rate of Ca2+ entry in lipid-loaded cells and also decreased the height of the peak of the DBHQ-induced increase in [Ca2+]cyt (Figure 9C).
The PKC inhibitor GFX reverses the inhibition of SOCE induced by lipid accumulation in lean rat hepatocytes and in hepatocytes isolated from obese Zucker rats
Calphostin C, another PKC inhibitor , also increased the rate of Ca2+ entry in lipid-loaded cells (Supplementary Figure S6). The initial rates of Ca2+ entry in H4IIE cells, loaded with lipid by pre-incubation with amiodarone and incubated in the absence and presence of calphostin C were 8.0±2.1 nM/min and 24.8±2.9 nM/min (means ± S.E.M., n=3) respectively. This action of calphostin C was also associated with a decrease in the height of the peak of the DBHQ-induced increase in [Ca2+]cyt (Supplementary Figure S6).
In case the time of exposure of the cells to GFX when the PKC inhibitor was added 5 min before DBHQ, as in Figure 8(A), was insufficient to ensure maximal inhibition of PKC, GFX (20 μM) was added to lipid-loaded H4IIE cells 30 min before beginning the measurement of [Ca2+]cyt. This also resulted in an increased rate of Ca2+ entry in lipid-loaded compared with control cells (Figure 8C). The rate of Ca2+ entry in lipid-loaded cells pre-treated for 30 min with GFX was 70% of that in control cells pre-treated for 5 min with GFX (Figure 8C cfFigure 8B). This comparison indicates that GFX can almost completely restore the inhibition of Ca2+ entry induced by lipid and hence suggests that the action of PKC is responsible for a large component of the observed lipid-induced inhibition of SOCE. When added 30 min before DBHQ, GFX reduced the height of the peak of the DBHQ-induced increase in [Ca2+]cyt (Figure 8C) to about the same degree as that observed for cells where GFX was added 5 min before DBHQ.
In order to test whether GFX can reverse the lipid-induced inhibition of SOCE independently of the DBHQ-induced release of Ca2+, GFX was added to lipid-loaded cells after the addition of DBHQ. Under these conditions, GFX also caused a substantial increase in the rate of Ca2+ entry (Figure 8D). In this experiment, there was no change in the height of the peak of the DBHQ-induced increase in [Ca2+]cyt in the absence of added [Ca2+]ext, as expected, since no GFX was present at the time of DBHQ addition. The rate of Ca2+ entry in lipid-loaded cells to which GFX was added after DBHQ was 84% of the rate when GFX was added to lipid-loaded cells 30 min before DBHQ addition (Figure 8D cfFigure 8C). This comparison suggests that the action of GFX in enhancing Ca2+ entry is independent of the effects of lipid accumulation on the amount of Ca2+ in intracellular stores.
The results described above indicate that, in addition to reversing the lipid-induced inhibition of Ca2+ entry, GFX decreases the height of the peak of the DBHQ-induced increase in [Ca2+]cyt. To further investigate the action of GFX on intracellular Ca2+ stores, the effect of GFX on ionomycin-releasable Ca2+ was investigated. Control (non lipid-loaded) H4IIE cells loaded with fura-2 were incubated with vehicle or with GFX (20 μM) added 1 min before or 30 min before ionomycin (50 μM). [Ca2+]cyt was measured as a function of time and the peak height of the ionomycin-induced increase in [Ca2+]cyt determined. The peak height values for vehicle and GFX added at 1 and 30 min before ionomycin were 157±12, 126±15 and 37±5 nM (means ± S.E.M., n=3). These results indicate that GFX does, over time, decrease the amount of Ca2+ in intracellular stores.
The results of the present study have shown that lipid-loaded and steatotic liver cells exhibit substantial decreases in SOCE and Ca2+ release from the ER and impaired hormone-initiated intracellular Ca2+ signalling. Pharmacological inhibition of SOCE or knockdown of STIM1 and Orai1 were shown to be associated with enhanced lipid accumulation under conditions (amiodarone or palmitate) which promote the synthesis of triacylglycerols and other lipids and the formation of cytoplasmic lipid droplets. It has been reported that decreased activity and/or expression of SERCA2b in the ER in hepatocytes from obese mice is associated with a decreased [Ca2+]ER leading to the ER stress response, decreased protein folding, decreased protein chaperone activity and glucose intolerance[12,13]. [Ca2+]ER homoeostasis is closely linked to the synthesis of triacylglycerols in hepatocytes and a low [Ca2+]ER concentration is associated with increased triacylglycerol synthesis . Moreover, ER membranes are the site from which cytoplasmic lipid droplets containing triacylglycerol and other neutral lipids originate in response to an excess of palmitate and other fatty acids .
Since the main function of SOCE is to replenish [Ca2+]ER stores , the present observation that SOCE is greatly reduced in lipid-loaded and steatotic liver cells indicates that inhibited SOCE acts in concert with inhibition of SERCA2b to create a low [Ca2+]ER in steatotic liver cells. This, in part, leads to an enhancement of cytoplasmic lipid droplet accumulation. Hence, it is proposed that the lipid-induced inhibition of SOCE and the associated low [Ca2+]ER act as a positive feedback loop to enhance further lipid accumulation. The conclusion that inhibition of SOCE enhances lipid accumulation is consistent with a recent report showing that knockout of STIM1 in Drosophila leads to lipid accumulation in fat storage cells . Recently we have shown that the lipid-induced inhibition of SOCE is reversed by the anti-diabetic glucagon-like peptide-1 agonist exendin-4 (Ali, E., Wilson, C., Rychkov, G., Tallis G. and Barritt G, unpublished results), which is known to be able to reduce liver steatosis (reviewed in ). Thus the present results may point to a new site of action for this anti-diabetic drug.
Evidence that PKC mediates the lipid-initiated inhibition of SOCE is provided by the results of experiments employing GFX and calphostin C, inhibitors of PKC, and PMA, an activator of PKC. Moreover, the results obtained from the experiments where PKC was down-regulated by 24 h pre-incubation with PMA or inhibited by 30 min treatment with GFX, indicate that activation of PKC accounts for most of the observed lipid-induced inhibition of SOCE. Although Ca2+ release from all or part of the ER is an essential step in SOCE activation, two observations suggest that the decreased capacity of the ER to store Ca2+ is unlikely to play a major role in the mechanism(s) by which lipid accumulation inhibits SOCE. First, GFX, when added after DBHQ (to induce [Ca2+]ER release) causes about the same degree of reversal of SOCE inhibition as is observed when GFX is added before DBHQ, suggesting that the step(s) involved in SOCE inhibition is downstream of [Ca2+]ER release. Secondly, in liver cells a small sub-region of the ER close to the plasma membrane, rather than the whole ER, seems to be involved in the activation of SOCE .
Hepatocyte steatosis is associated with the activation of PKCε, PKCδ and PKCβ [7–9,11,47]. PKCβ can phosphorylate Orai1 at N-terminal residues Ser27 and Ser30 leading to inhibition of SOCE . Since the activity of STIM1 is also regulated by phosphorylation , PKC-mediated phosphorylation of STIM1 could also account for the observed SOCE inhibition, although to our knowledge no PKC-mediated phosphorylation of STIM1 has been reported .
The activation of PKC inhibits SOCE in other cell types [39,50,51] and this was confirmed for hepatocyte SOCE in the present results employing PMA. Thus the mechanism by which lipid accumulation inhibits SOCE in liver cells probably involves the activation of one or more isoforms of PKC by diacylglycerol present in lipid droplets, altered phosphorylation of Orai1 or STIM1 and inhibition of Ca2+ entry through Orai1. The observation that GFX, when added after DBHQ, had induced the release of Ca2+ from the ER, reversed the lipid-induced inhibition of SOCE suggests that the site of action of PKC is Orai1 rather than STIM1. Thus, in liver cells, the results of patch clamp experiments (e.g. the action of ATP, Figure 3D of the present study) indicate that SOCE can be fully activated at about 3 min after initiation of the release of Ca2+ from the ER. At this time, STIM1 would have moved to the ER-plasma membrane junction and interacted with Orai1. In the experiment described in Figure 8A, GFX was added at 10 min after DBHQ at which time STIM1 would have moved fully to the plasma membrane, suggesting that the actions of GFX and hence PKC are on the phosphorylation of Orai1.
The observed decrease in Ca2+ in the DBHQ- and ionomycin-releasable stores in lipid-loaded liver cells is most probably due to both the inhibition of Ca2+ uptake into the ER as a result of impaired SOCE, as shown in the present study, and to the lipid-induced inhibition of SERCA2b as shown by others [12,52]. It has also been shown that in some other cell types the activation of PKC leads to reduction of the amount of [Ca2+]ER [53,54]. This effect of PKC may be mediated by the inhibition of SERCA2b or by other mechanisms. In the present study, it was found that when liver cells were exposed to the PKC inhibitor GFX for a period of 5 or 30 min, GFX reduced the amount of Ca2+ in the DBHQ- and ionomycin-releasable stores. Such an effect of GFX has previously been reported and may be due to an inhibition by GFX of Ca2+ uptake into the ER, possibly independent of PKC .
There was little change in the expression of Orai1 and STIM1, the main isoforms of Orai and STIM, which constitute SOCs in liver cells  indicating that the lipid-induced inhibition of SOCE is unlikely to be due to decreased expression of Orai1 and STIM1. The observation that Orai1 protein expression is decreased with no observed change in expression of mRNA may represent post transcriptional regulation [55,56]. A decrease in STIM2 protein was observed in hepatocytes from obese Zucker rats. However, this was not associated with a significant decrease in STIM2 mRNA in either obese Zucker rat hepatocytes or in lipid loaded H4IIE cells. Since previous studies have shown that the main components of SOCE in liver cells are Orai1 and STIM1 , it is considered unlikely that the decrease in STIM2 protein contributes to the observed inhibition of SOCE but may be involved in alterations of other Ca2+ entry pathways.
Taken together, the present results show that SOCE is substantially inhibited in lipid-loaded liver cells. Evidence is provided to indicate that the inhibition is mediated by the activation of one or more isoforms of PKC. The inhibition of SOCE greatly reduces the intracellular Ca2+ signals generated by Ca2+-dependent hormones and, more importantly, contributes to a steady-state lowered [Ca2+]ER and the ER stress response. This, in turn, enhances the accumulation of intracellular lipids in a positive feedback mechanism. Thus the inhibition of SOCE may lead to greatly enhanced steatosis and subsequent insulin insensitivity in patients with NASH. Agents which can increase SOCE in steatotic liver cells may lead to a reduction in lipid accumulation and to slowed development of insulin resistance.
cytoplasmic free Ca2+ concentration
ER luminal Ca2+ concentration
inositol 1,4,5 trisphosphate
current through SOCs
protein kinase C
real-time quantitative PCR
sarco/endoplasmic reticulum (Ca2++Mg2+)ATPase
store-operated Ca2+ channel
store-operated Ca2+ entry
stromal interaction molecule
Type 2 diabetes
Grigori Rychkov, Greg Barritt and Claire Wilson conceived the ideas for the initial study. Together with Eunus Ali and Nathan Scrimgeour, they also devised and designed the experimental plans. Claire Wilson and Eunus Ali performed the majority of the experiments. Nathan Scrimgeour performed the patch clamp experiments with Zucker rat hepatocytes and lipid-loaded H4IIE cells. Alyce Martin conducted the western blot experiments on expression of STIM and Orai proteins in lipid-loaded H4IIE cells. Jin Hua prepared isolated hepatocytes and conducted qPCR experiments to determine the degree of knockdown of STIM1 and Orai1 in H4IIE cells treated with siRNA. George Tallis provided a clinical interpretation of the data. Claire Wilson, Greg Barritt, Eunus Ali and Grigori Rychkov wrote the manuscript. All authors analysed the data, read, and approved, the manuscript.
Advice provided by Dr Yabin Zhou is gratefully acknowledged.
This study was supported by the Diabetes Australia Research Trust, Flinders University Adelaide [grant number Y11-BARG]; and the Flinders Medical Centre Foundation.
These authors contributed equally to the work.