Reactive oxygen species (ROS) are produced in enzymatic and non-enzymatic reactions and have important roles in cell signalling but also detrimental effects. ROS-induced damage has been implicated in a number of neurological diseases; however, antioxidant therapies targeting brain diseases have been unsuccessful. Such failure might be related to inhibition of ROS-induced signalling in the brain. Using direct kinetic measures of lipid peroxidation in astrocytes and measurements of lipid peroxidation products in brain tissue, we here show that phospholipase C (PLC) preferentially cleaves oxidised lipids. Because of this, an increase in the rate of lipid peroxidation leads to increased Ca2+ release from endoplasmic reticulum (ER) stores in response to physiological activation of purinoreceptors with ATP. Both vitamin E and its water-soluble analogue Trolox, potent ROS scavengers, were able to suppress PLC activity, therefore dampening intracellular Ca2+ signalling. This implies that antioxidants can compromise intracellular Ca2+ signalling through inhibition of PLC, and that PLC plays a dual role – signalling and antioxidant defence.
Reactive oxygen species (ROS) are produced in biological reactions and have important roles in cell signalling and homeostasis but also detrimental effects, ultimately leading to cell death, if ROS levels increase dramatically within the cell (Dröge, 2002; Martindale and Holbrook, 2002). Lipids are important targets of ROS. It is therefore not surprising that lipid-rich tissues, such as the brain, are affected by the detrimental effects of ROS, which are known to play a role both in acute brain disease, such as stroke, and also in more chronic neurodegenerative disease such as Alzheimer's or Parkinson's disease (Abramov et al., 2004; Abramov et al., 2007; Gandhi et al., 2009). Despite compelling evidence that ROS contribute to neuronal damage, antioxidants targeting ROS in neurodegenerative disease have failed (Barnes and Yaffe, 2005; Kamat et al., 2008; Sano et al., 1997). Moreover, particularly vitamin E supplementation has even been shown to be harmful in some cases (Bjelakovic et al., 2007).
Lipids are important substrates of cell signalling pathways, and phospholipase C (PLC)-induced signalling presents one of the most abundant lipid-dependent signalling pathways. PLC cleaves phospholipids to diacyl glycerol (DAG) and inositol trisphosphate (InsP3). InsP3-induced Ca2+ release from internal Ca2+ stores represents one of the most powerful intracellular signals within the eukaryotic cell. ROS can activate such release of Ca2+ from endoplasmic reticulum (ER) stores; a mechanism which has been thought to be mediated by ROS-mediated activation of RyR receptors on the ER (Camello-Almaraz et al., 2006). However, there is also strong evidence that ROS activates PLC, which ultimately contributes to Ca2+ release (Servitja et al., 2000; Vaarmann et al., 2010). These indirect, PLC-mediated, ROS-induced Ca2+ signals are likely to play an important role in the brain, where there are high levels of oxidisable polyunsaturated fatty acids and high ROS load due to high oxygen consumption (Halliwell, 2006).
Surprisingly, despite abundant evidence that ROS activate PLC, the mechanism of this activation is unclear. This is crucial, because vitamin-E-derived antioxidants are particularly effective in protecting lipids from oxidation. Therefore, such antioxidant therapies potentially might interfere with PLC signalling, contributing to failure of antioxidant therapies in neurological disease.
RESULTS AND DISCUSSION
ATP-induced Ca2+ signals in astrocytes are reduced by antioxidants
Ca2+ signalling is essential for various processes in the central nervous system (Rink and Merritt, 1990). Under physiological conditions, Ca2+ mobilization from the ER stores is mediated by a variety of agonists, which are known to trigger InsP3 synthesis (Rossi et al., 2012). InsP3-related Ca2+ signals are evoked through stimulation of P2Y receptors by ATP or inorganic polyphosphate (Burnstock, 2013; Holmström et al., 2013), or by direct activation of PLC by H2O2 produced during the utilisation of dopamine (Vaarmann et al., 2010). We here focus on purinoreceptor signalling in astrocytes, which is one important PLC- and InsP3-mediated signalling pathway in the brain (James and Butt, 2001). ATP (50 µM) in astrocytes caused a stereotypical response that consists of an initial transient increase in the cytosolic Ca2+ level ([Ca2+]c) followed by recovery to the basal Ca2+ level in 2–3 minutes (n = 78; Fig. 1A). Pre-incubation of the cells with vitamin E (10 µM, 30 minutes), an effective lipid-soluble antioxidant, significantly reduced the ATP-induced Ca2+ signal (0.53±0.03 versus 0.30±0.07 fura-2 ratio; n = 72; P<0.001; Fig. 1B,D). Similarly, incubation of astrocytes with the cell-permeable water-soluble derivate of vitamin E Trolox (100 µM, 30 minutes), decreased the ATP-induced Ca2+ signal even further (0.20±0.05 fura-2 ratio; n = 70; P<0.001; Fig. 1C,D). This attenuation of ATP-induced Ca2+ signals with Trolox was also observed if Ca2+ signals were triggered with higher (200 µM; n = 27; P<0.01; Fig. 1H) or lower (5 µM; n = 98; P<0.001; Fig. 1E–G) concentrations of ATP. Moreover, the size, that is the absolute amplitude reduction by Trolox of the ATP-induced Ca2+ signal, was independent of the ATP concentration (Fig. 1D,G,H), implying that all three ATP concentrations induced a similar activation of the downstream signalling cascades. It has been shown that 3 µM ATP or ADP induces maximal Ca2+ release from ER stores in astrocytes by activation of purinergic receptors (Doengi et al., 2008). In this situation, the amplitude of the Ca2+ signal in Fig. 1A,E is dependent on the basal Ca2+ level within the astrocyte rather than on the ATP concentration (Kirischuk et al., 1995).
We hypothesized that lipid peroxidation, i.e. oxidized lipids are activators of the PLC-InsP3 pathway, leading to Ca2+ release. To test this hypothesis, we examined the effect of H2O2 on ATP-induced PLC activation resulting in ATP-induced Ca2+ signals.
ATP-induced Ca2+ signals in astrocytes are increased by oxidative stress
A 20-minute pre-incubation of astrocytes with H2O2 (100 µM) induced a significant decrease in the ATP-induced [Ca2+]c rise in the majority of astrocytes (n = 58; from 0.53±0.03 fura-2 ratio to 0.34±0.03; P<0.001; Fig. 2A). Given that this result is in striking contrast to what is expected considering the previous results (Fig. 1), we tested the acute effects of H2O2 on [Ca2+]c levels in astrocytes. Application of H2O2 stimulated Ca2+ oscillations in 44.2±3.1% of astrocytes (Fig. 2B, blank trace). This signal was initiated and maintained from intracellular Ca2+ stores by activation of PLC, because inhibition of this enzyme with U73122 (5 µM) completely prevented H2O2-induced [Ca2+]c changes in astrocytes (n = 48; Fig. 2B, U73122 trace). Therefore, H2O2 itself induced a Ca2+ signal in ∼50% of astrocytes that was dependent on intracellular Ca2+ stores. A subsequent ATP-induced Ca2+ signal might therefore be smaller in astrocytes, as the intracellular Ca2+ stores are already depleted. Therefore, we performed an analysis of the ATP-induced Ca2+ signal in H2O2-treated astrocytes separating astrocytes into two groups depending on the presence of the H2O2-induced Ca2+ signal. We found that in the presence of an H2O2-induced Ca2+ signal (which was independent of shape and type of astrocytes), and therefore depleted ER stores, subsequent ATP-stimulated astrocytic Ca2+ signals were small when compared to astrocytic Ca2+ signals evoked by ATP stimulation only (0.29±0.02 versus 0.53±0.03 fura-2 ratio; P<0.001; Fig. 2D). In H2O2-treated cells without a Ca2+ signal, ATP (50 µM) induced a significantly higher [Ca2+]c increase compare to control treatment (0.92±0.04 versus 0.53±0.03 fura-2 ratio; P<0.001; Fig. 2C,D). Here, we describe the interaction of two very important processes: Ca2+ signalling and antioxidant defence. Inhibition of ROS production by vitamin E and Trolox led to a decrease in the Ca2+ response in astrocytes after activation with ATP, which was dependent on intracellular ER stores. Some other studies have shown that ROS, like H2O2 can affect PLC function by pre-depleting internal Ca2+ stores (Gerich et al., 2009; Sun et al., 2005; Volk et al., 1997). However, we found a subgroup of astrocytes that did not exhibit H2O2-induced Ca2+ signals. Consequently in these cells ER stores were not depleted. This subgroup of astrocytes showed an increased response following ATP application (Fig. 2C), supporting our hypothesis that ROS and lipid peroxidation increase the PLC-mediated signal. To assess whether antioxidants influence intracellular Ca2+ stores or store-operated Ca2+ entry, we pre-incubated astrocytes with Trolox (100 µM, 30 minutes) and the size of the ER Ca2+ store was estimated as an increase in fura-2 ratio after stimulation with thapsigargin (1 µM) in Ca2+ free (+0.5 mM EGTA) external medium (Fig. 2E). Store-operated Ca2+ entry was assessed thereafter, replacing the external Ca2+-free medium with medium containing Ca2+ (2 mM). This led to a massive Ca2+ influx into astrocytes, as measured with fura-2 ratios (Fig. 2E). Both intracellular Ca2+ stores and store-operated Ca2+ entry, as measured with fura-2 ratios, did not differ significantly between astrocytes pre-treated with Trolox (n = 53) and astrocytes which did not undergo antioxidant treatment (n = 49; Fig. 2F,G; both P>0.05) suggesting that the effect of antioxidants such as Trolox and vitamin E on PLC are independent of downstream signalling cascades and store-operated Ca2+ entry.
Stimulation of astrocytes with ATP increases the rate of ROS production and induces a transient decrease in lipid peroxidation
In agreement with our and others’ observations (Abramov et al., 2005), ATP (50 µM) in astrocytes induced a 2.5-fold increase in the rate of ROS production (n = 78; P<0.05; Fig. 3A). ATP-induced ROS production was completely blocked by the NADPH oxidase inhibitor DPI (0.5 µM; n = 37; Fig. 3A).
In astrocytes treated with ATP, we monitored the rate of lipid peroxidation using the indicator C11-BODIPY (581/591). ATP (50 µM) induced a significant decrease in the rate of lipid peroxidation (to 37.6±2.3% of the basal rate; n = 128; P<0.001) for 3–5 minutes, followed by a return to basal rates of lipid peroxidation (Fig. 3B). Pre-treatment of astrocytes with vitamin E (10 µM, 15 minutes) significantly reduced basal rates of lipid peroxidation (to 88.6±9.7% of control; n = 86; P<0.05; Fig. 3C) but did not completely abolish the transient decrease in the rate of lipid peroxidation in response to 50 µM ATP (to 41.7±3.1% of control basal rate; Fig. 3C).
Pre-incubation of the cells with PLC inhibitors U73122 (5 µM, 10 minutes; n = 45) and edelfosine (ET-18-OCH3; 10 µM, 10 minutes; n = 63) completely prevented the ATP-induced decrease in the rate of lipid peroxidation (Fig. 3D,E). In contrast, xestospongin C (5 µM; n = 58; Fig. 3E,F), a potent antagonist of InsP3-mediated Ca2+ release, did not block the ATP-induced decrease in the rate of lipid peroxidation, suggesting that downstream targets of PLC signalling, or more specifically InsP3 signalling, do not affect the ATP-induced reduction in the rate of lipid peroxidation. These findings provide direct evidence for the fact that the ATP-induced decrease of lipid peroxidation in astrocytes is independent of InsP3-induced Ca2+ signalling (Fig. 3E,F). Taken together, this data imply that oxidized lipids are utilised by PLC.
H2O2 induces lipid peroxidation, activates PLC and increases the InsP3 level in rat brain homogenate
In rat brain homogenates treated with H2O2, we assessed the malondialdehyde (MDA) level [as the most abundant end-product of polyunsaturated fatty acids (PUFAs) peroxidation and a well-established marker of lipid peroxidation] together with PLC activity and InsP3 level. H2O2 (100 µM) induced an increase in the MDA level of 135±5.2% compared to control (set at 100% lipid peroxidation; P<0.05; Fig. 4A). As expected, in samples co-treated with H2O2 and vitamin E or Trolox the level of MDA was reduced to around the control value (Fig. 4A); vitamin E alone reduced lipid peroxidation by 21%, and Trolox alone reduced it by 40% (data not shown).
In the same supernatant samples of brain homogenate treated with H2O2, the activity of PLC and InsP3 concentration were determined. In samples treated with H2O2 the activity of PLC increased by 28% compared to control (set at 100% of PLC activity; P<0.05; Fig. 4B) and the level of InsP3 by 14.3% (control set at 100% of InsP3 level; P<0.05; Fig. 4C) confirming the fact that ROS can activate PLC (Servitja et al., 2000; Vaarmann et al., 2010; Wang et al., 2001). The PLC activity and InsP3 level in the brain supernatant samples co-treated with H2O2 and vitamin E or Trolox were similar to the control samples (around 100%), indicating that vitamin E and Trolox, by preventing lipid peroxidation, prevent PLC activation and generation of InsP3 (Fig. 4B,C). These results confirm that oxidised lipids are PLC substrates.
Our study, for the first time, shows that lipid peroxidation potentiates intracellular Ca2+ signalling through activation of PLC. Using direct kinetic measures of lipid peroxidation, we show that activation of PLC induces a transient reduction in the rate of lipid peroxidation, implying that the substrates of PLC are oxidised lipids. This adds a new layer of complexity to the functions of PLC, as PLC seems to not only to play a role in signalling pathways but also antioxidant homeostasis. By contrast, our data also implies that dampening of ROS levels in cells with high doses of vitamin E, or other antioxidants, inhibits PLC-dependent signalling. This involves multiple physiological processes where InsP3 plays a key role, and might also partly explain the failure of antioxidant therapies in brain disease (Barnes and Yaffe, 2005; Kamat et al., 2008; Sano et al., 1997).
MATERIALS AND METHODS
Primary co-culture preparation
Mixed cultures of hippocampal, cortical or midbrain neurones and glial cells were prepared as described previously (Domijan and Abramov, 2011). Cultures were maintained at 37°C in a humidified atmosphere of 5% CO2 and 95% air for a minimum of 12 days before experimental use to ensure the expression of receptors. All animal experiments were performed according to approved guidelines.
Measurement of [Ca2+]c and ROS
For measurement of the cytosolic Ca2+ level ([Ca2+]c), cells were loaded for 30 minutes at room temperature with 5 µM fura-2 AM (Molecular Probes, Eugene, OR) and 0.005% Pluronic acid in HBSS buffer salt solution (GIBCO® HBSS, Invitrogen). For measurement of ROS production, dihydroethidium (HEt; 2 µM; Molecular Probes, Eugene, OR) was used.
[Ca2+]c was monitored in single cells using excitation light provided by a Xenon arc lamp, the beam passing through a monochromator at 340 and 380 nm. Emitted fluorescence light was reflected through a 515-nm long-pass filter to a cooled CCD camera. For HEt measurements the excitation wavelength (λex) was 360 nm and emission wavelength (λem) was 430 nm for non-oxidised DHE, and λex 530 nm and λem 560 nm for oxidased HEt.
All imaging data were collected and analysed using software from Andor (Belfast, UK).
Measurement of lipid peroxides
To estimate the rate of lipid peroxidation in the cells, cells were pre-incubated for 20 minutes with C11-BODIPY (5 µM; Molecular Probes, Eugene, OR). C11-BODIPY fluorescence was measured using a Zeiss 710 CLSM confocal microscope. C11-BODIPY (581/591) was excited using the 488 and 563 nm laser line, and fluorescence measured from 505 to 550 nm and 570 and 630 nm. All data presented were obtained from at least five experiments.
Measurement of lipid peroxidation level, PLC activity and InsP3 level
In the supernatants of brain homogenates of adult, male Sprague-Dawley rats (UCL breeding colony) level of malondialdehyde (MDA), a well-known marker of lipid peroxidation, PLC activity and InsP3 level were assessed. The brain homogenate (0.9% NaCl) was centrifuged (1000 g, 10 minutes) and afterwards the collected supernatant was treated with H2O2 (100 µM), and with H2O2 (100 µM) plus vitamin E (100 µM) or Trolox (50 µM). Three control samples were included: a control sample treated with equal volume of distilled water, and control samples treated with only vitamin E (100 µM) or Trolox (50 µM). The samples were incubated (37°C for 30 minutes), and after incubation in supernatants MDA concentration, activity of PLC and InsP3 concentration were determined. MDA was measured according to the thiobarbituric acid assay, using HPLC with a fluorescence detector set at λex 527 nm and λem 551 nm (Drury et al., 1997; Gajski et al., 2012). PLC activity and InsP3 concentration were assessed by use of the commercial assay kits ‘EnzChek Direct Phospholipase C’ (Molecular Probes, Eugene, OR, USA) and ‘Rat Inositol 1,4,5,-Trisphosphate (InsP3) ELISA Kit’ (Cusabio, Wuhan, China), respectively, according to manufactures' recommendations with the aid of a plate reader (Victor, PerkinElmer, Waltham, MA, USA).
Results for cytosolic Ca2+ level, ROS and lipid peroxides are expressed as means ± the standard error of the mean (s.e.m.). Results for MDA, PLC and InsP3 are presented as mean percentage ± the standard deviation (s.d.) from the control value (the untreated supernatant sample, set to 100%). Statistical analysis was performed with the aid of Origin 8 (Microcal Software Inc., Northampton, MA, USA) software.
A.-M.D., S.K. and A.Y.A. produced experiments, analyzed data and wrote the manuscript.
The work was funded by The Liverhulme Trust.
The authors declare no competing interests.