Transient receptor potential melastatin 2 (TRPM2) channels are sensitive to oxidative stress, and their activation can lead to cell death. Although these channels have been extensively studied in expression systems, their role in the brain, particularly in the substantia nigra pars compacta (SNc), remains unknown. In this study, we assessed the expression and functional properties of TRPM2 channels in rat dopaminergic SNc neurons, using acute brain slices. RT-PCR analysis revealed TRPM2 mRNA expression in the SNc region. Immunohistochemistry demonstrated expression of TRPM2 protein in tyrosine hydroxylase-positive neurons. Channel function was tested with whole cell patch-clamp recordings and calcium (fura-2) imaging. Intracellular application of ADP-ribose (50–400 μM) evoked a dose-dependent, desensitizing inward current and intracellular free calcium concentration ([Ca2+]i) rise. These responses were strongly inhibited by the nonselective TRPM2 channel blockers clotrimazole and flufenamic acid. Exogenous application of H2O2 (1–5 mM) evoked a rise in [Ca2+]i and an outward current mainly due to activation of ATP-sensitive potassium (KATP) channels. Inhibition of K+ conductance with Cs+ and tetraethylammonium unmasked an inward current. The inward current and/or [Ca2+]i rise were partially blocked by clotrimazole and N-(p-amylcinnamoyl)anthranilic acid (ACA). The H2O2-induced [Ca2+]i rise was abolished in “zero” extracellular Ca2+ concentration and was enhanced at higher baseline [Ca2+]i, consistent with activation of TRPM2 channels in the cell membrane. These results provide evidence for the functional expression of TRPM2 channels in dopaminergic SNc neurons. Given the involvement of oxidative stress in degeneration of SNc neurons in Parkinson's disease, further studies are needed to determine the pathophysiological role of these channels in the disease process.
- oxidative stress
- hydrogen peroxide
- whole cell patch-clamp recording
oxidative stress is believed to be a major factor in the pathogenesis of a number of neurodegenerative disorders including Parkinson's disease (PD), and it is considered to play a role in both the initiation of the disease process and the secondary neural damage (see, e.g., Chinta and Andersen 2008; Jenner 2007; Reynolds et al. 2007; Zhou et al. 2008). Oxidative stress results from a loss of redox balance within the cell, when the production of reactive oxygen species (ROS) overwhelms the detoxification capability of the cell.
Cell damage resulting from oxidative stress is believed to be caused mainly by the direct toxic reaction of free radicals with DNA, proteins, carbohydrates, and lipids (Davies 2000). In addition, recent studies have identified a number of ion channels belonging to the transient receptor potential (TRP) channel superfamily that are sensitive to oxidative stress and may contribute to ROS-induced cell death (Ramsey et al. 2006). These include TRP melastatin 2 (TRPM2) (Eisfeld and Luckhoff 2007; Jiang et al. 2010; Sumoza-Toledo and Penner 2011), TRPC3/4 (Poteser et al. 2006), TRPP2 (Montalbetti et al. 2008), TRPV1 (Crouzin et al. 2010), and TRPA1 (Andersson et al. 2008) channels. Of particular interest are TRPM2 channels, as they are permeable to calcium and are highly expressed in the brain, including the cerebral cortex, cerebellum, striatum, and hippocampus (see, e.g., Bai and Lipski 2010; Hara et al. 2002). In addition, expression of these channels was detected in the human substantia nigra with Northern blotting (Nagamine et al. 1998; Uemura et al. 2005). Functional expression of native TRPM2 channels has been demonstrated in striatal (Bao et al. 2005; Fonfria et al. 2005; Hill et al. 2006), hippocampal (Olah et al. 2009), and cortical (Kaneko et al. 2006) neurons.
TRPM2 channels contain an enzymatic domain located near the COOH terminus that shares significant homology to NUDT9 (ADP-ribose pyrophosphatase) containing the Nudix sequence motif (Perraud et al. 2001). These channels are strongly activated by H2O2 and ADP-ribose (ADPR) and are modulated by other ROS, reactive nitrogen species, ADPR-related compounds (NAD and cyclic ADPR), TNF-α, arachidonic acid, and temperature changes (Hara et al. 2002; Perraud et al. 2001; Sano et al. 2001; Togashi et al. 2006; Toth and Csanady 2010; Wehage et al. 2002). TRPM2 protein forms a voltage-independent nonselective cationic channel (PCa/PNa ∼ 0.7; Perraud et al. 2001; Sano et al. 2001; Xia et al. 2008). Intracellular calcium is required for channel activation and can markedly potentiate the effects of H2O2 or ADPR (Csanady and Torocsik 2009; Du et al. 2009; McHugh et al. 2003; Starkus et al. 2007).
TRPM2 channels have been demonstrated to mediate cell death in heterologous expression systems and native nonneuronal cells by increasing susceptibility to ROS (including H2O2), TNF-α, and amyloid β-peptide (Fonfria et al. 2005; Hara et al. 2002; Ishii et al. 2007; Kaneko et al. 2006; Yang et al. 2006; Zhang et al. 2003, 2006). In contrast, reduction of channel expression or inhibition of its function increased cellular resistance to damage. However, it is unclear whether these channels affect the function of dopaminergic neurons of the substantia nigra pars compacta (SNc) and contribute to the damaging effects of oxidative stress observed in PD. Our recent study suggested the involvement of TRPM2 channels in the response of SNc neurons to rotenone, a neurotoxin used to model PD (Freestone et al. 2009). In the present study, we investigated the expression of these channels in SNc neurons at the molecular and functional levels. We hypothesized that TRPM2 channels mediate the specific effects of exogenously applied ADPR and H2O2, leading to changes in neuronal excitability and calcium homeostasis, and potentially to damage of these neurons.
Brain tissue was obtained from Wistar rats (P15–P28). Live brain slices were used for mRNA extraction, patch-clamp recordings, and Ca2+ imaging, whereas fixed tissue was used for immunohistochemistry. All procedures were approved by the University of Auckland Animal Ethics Committee in accordance with the New Zealand Animal Welfare Act 1999.
Rats were anesthetized with 100% CO2 or chloral hydrate (450 mg/kg ip) and decapitated. The brain was quickly removed, and coronal midbrain slices (250 μm) containing the SNc were cut with a vibratome in artificial cerebral spinal fluid (standard aCSF, in mM: 115 NaCl, 2.5 KCl, 2.4 CaCl2, 1.3 MgCl2, 1.2 NaH2PO4, 24 NaHCO3, 10 glucose; 4–8°C) equilibrated with carbogen (95% O2-5% CO2). After cutting, slices were allowed to recover in a holding chamber containing carbogenated aCSF for at least 40 min at 34°C.
For mRNA extraction, aCSF was substituted with sterile phosphate-buffered saline (PBS, in mM: 137 NaCl, 3 KCl, 10 Na2HPO4, 2 KH2PO4) and slices were processed immediately after cutting. The SNc region was cut from midbrain slices transilluminated under a dissecting microscope with a scalpel blade or a modified hypodermic needle as described by us previously (Comer et al. 1998).
Reverse transcription-polymerase chain reaction.
Total RNA was extracted from the dissected tissue with the RNeasy Mini Kit (Qiagen). DNase digestion was performed on-column. Approximately 1 μg of total RNA was extracted per milligram of tissue. First-strand cDNA was synthesized with SuperScript III Reverse Transcriptase and oligo(dT) primers (Invitrogen). Platinum Taq DNA Polymerase (Invitrogen) was used for PCR amplification. Primers for tyrosine hydroxylase (TH; Comer et al. 1998) were used for positive controls. PCR products were analyzed by electrophoresis on 1.5% agarose gels in 0.5× Tris-borate-EDTA buffer (in mM: 90 Tris base, 90 boric acid, 2 EDTA) and visualized with ethidium bromide (2 μg/ml). Amplicons were either purified with the QIAquick PCR Purification Kit (Qiagen) or extracted from gels with the QIAEX II Gel Extraction Kit (Qiagen). Purified products were sequenced with the ABI PRISM 377 DNA Sequencer (Applied Biosystems).
Two primer sets were designed based on the rat TRPM2 mRNA sequence (NCBI Reference Sequence NM_001011559): 1) forward 5′-AGGGATCCAGGCTTTCCTAA-3′, reverse 5′-GCCATCGGGATCGTAGAATA-3′, amplicon 358 bp, spanning residues 2196–2553; 2) forward 5′-TGGGGAAGTATGTCCGACTC-3′, reverse 5′-CAGGATGAAGTGGGAGTGGT-3′, amplicon 443 bp, spanning residues 413–855. In addition, a set of primers targeting residues 2074–2445 (Bai and Lipski 2010) was used. All oligonucleotides were obtained from Invitrogen.
Rats were anesthetized with chloral hydrate and perfused transcardially with 0.15 M NaCl and 1 U/ml heparin, followed by 4% formaldehyde in PBS. Tissue blocks containing the midbrain were dissected and postfixed overnight at 4°C. Tissue was cryoprotected in PBS containing 30% sucrose prior to cryostat sectioning (30–50 μm). Sections were incubated overnight at 4°C with primary antibody diluted in PBS containing 0.4% Triton X-100 and 1% bovine serum albumin or 5% goat serum, followed by incubation with secondary antibody for 1–2 h at room temperature. Processed sections were mounted with Prolong Gold Antifade Reagent (Invitrogen).
Anti-TH antibodies were used for immunolabeling of dopaminergic neurons (polyclonal AB152 or monoclonal MAB5280; Millipore). Two TRPM2 antibodies were used: 1) rabbit anti-human TRPM2 (NB500-242; Novus Biologicals) and 2) rabbit anti-rat TRPM2 (NB110-81601; Novus Biologicals). Alexa Fluor 488-, 568-, or 594-conjugated secondary antibodies (Invitrogen) were used for fluorescence detection. Images were obtained with a confocal microscope (LSM 710; Zeiss). All sections incubated without primary antibodies were blank (negative controls; not shown). Preabsorption of the rabbit anti-human TRPM2 antibody with immunizing peptide resulted in no TRPM2 immunostaining (Bai and Lipski 2010).
Whole cell patch-clamp recording.
Slices were placed in a custom-built recording chamber (volume ≈ 0.6 ml) on the stage of an upright microscope (Eclipse E600FN; Nikon) and submerged in carbogenated aCSF (34°C; 2.5–3.0 ml/min). Patch pipettes were pulled from thin-wall glass capillaries and filled with either a K+-based (in mM: 135 K-methanesulfonate, 10 HEPES, 0.75 EGTA, 2 MgATP, 0.3 NaGTP, 2 MgCl2, 0.1 CaCl2) or a Cs+-based [in mM: 118 Cs-methanesulfonate, 30 tetraethylammonium (TEA), 10 HEPES, 0.75 EGTA, 2 MgATP, 0.3 NaGTP, 2 MgCl2, 0.1 CaCl2] solution. Neurons were patched after visualization by infrared differential interference contrast (IR-DIC) under a ×40 objective. Voltage-clamp recordings were performed with a Multiclamp 700A amplifier (Molecular Devices) using a 10-kHz low-pass filter. Pipette and cell capacitance were compensated for each cell. Access resistance was measured routinely, and recordings were discarded if the resistance exceeded 25 MΩ. In all recordings, cells were clamped at −60 mV (after adjusting for junction potential).
Intracellular free calcium concentration ([Ca2+]i) was measured with fura-2 (pentapotassium salt, 0.25 mM; Invitrogen) included in the pipette solution concurrently with whole cell currents recorded in voltage clamp. Except for experiments in which cells were loaded with ADPR, [Ca2+]i measurements were delayed at least 10 min after entering the whole cell configuration to allow diffusion of the calcium indicator into the cell. Cells were illuminated with a xenon arc lamp (175 W, Lambda LS; Sutter Instrument) and a filter wheel (Optospin; Cairn Research) that allowed switching of the excitation between 340 ± 15 and 380 ± 5 nm. A band-pass emission filter (540 ± 25 nm) was used. Images were captured with a cooled CCD camera (12-bit, ORCA-ER; Hamamatsu) with 2 × 2 binning and 200- to 400-ms exposure. A pair of images (340- and 380-nm excitation) was captured every 5 or 10 s. Images were acquired and analyzed with Imaging Workbench (v. 5.2; INDEC BioSystems). Fluorescence ratios were converted to concentrations according to the procedure described by Grynkiewicz et al. (1985).
ADPR and 3,4-dihydro-5-[4-(1-piperidinyl)butoxyl]-1(2H)-isoquinolinone (DPQ) (both from Sigma-Aldrich) were added to the pipette solution. H2O2 (Scharlau) was slowly infused (0.73 ml/h) at high concentration into the inflow port of the recording chamber with a syringe pump (SP100iZ; WPI) to achieve a final concentration of 1–5 mM in the recording chamber. The following drugs were dissolved in aCSF: 2-amino-5-phosphonopentanoic acid (AP5), clotrimazole, flufenamic acid (all from Sigma-Aldrich) and 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX; Tocris). Stock solutions of N-(p-amylcinnamoyl)anthranilic acid (ACA; Calbiochem), glibenclamide (Tocris), and tolbutamide (Sigma-Aldrich) were prepared in DMSO; the final DMSO concentration in aCSF did not exceed 0.1%.
Electrophysiological and calcium imaging data were analyzed with Clampfit (v. 9; Molecular Devices), Excel (v. 11; Microsoft), and Origin (v. 7.5; OriginLab). Statistical tests were performed with Statistica (v. 8; StatSoft). Student's t-test or split-plot ANOVA followed by planned comparisons was used to determine statistical significance. Data are presented as means ± SE. Differences were considered statistically significant at P < 0.05.
TRPM2 mRNA and protein expression.
To assess the expression of TRPM2 mRNA in the SNc region, RT-PCR was performed on freshly dissected SNc tissue (Fig. 1A). Two primer sets targeting different regions of the rat TRPM2 mRNA sequence were used (see methods). Gel electrophoresis detected PCR products of the expected size for each primer set [443 and 358 bp, n (animals) = 2; Fig. 1B]. The amplicons were sequenced and found to be identical to the reported cDNA sequence. In addition, RT-PCR performed with previously described TRPM2 primers (Bai and Lipski 2010) resulted in amplicons of predicted size (not shown). Expression of TH mRNA was tested and confirmed that the collected tissue included dopaminergic neurons.
To examine TRPM2 protein expression in the SNc region, fluorescence immunohistochemistry was performed on midbrain sections. Dopaminergic neurons were immunolabeled with an anti-TH antibody that allowed visualization of both their cell bodies and proximal dendrites. As expected, TH-immunoreactive neurons were found in the SNc, with fewer TH-positive cells scattered in the substantia nigra pars reticulata (SNr; Fig. 2). TRPM2 protein was labeled with either the rabbit anti-rat TRPM2 antibody (Fig. 2) or the rabbit anti-human TRPM2 antibody (data not shown). The pattern of immunostaining was identical between sections incubated with the two antibodies. TRPM2 immunoreactivity was observed in cells throughout midbrain slices, including the SNc and SNr regions. At the cellular level, TRPM2 immunoreactivity was punctate, with staining in the cytoplasm of the soma and proximal dendrites. Double-labeling with anti-TH and anti-TRPM2 antibodies revealed TRPM2 protein expression in dopaminergic neurons (Fig. 2). In addition, TRPM2 immunoreactivity was present in non-TH-immunoreactive cells.
Identification of dopaminergic SNc neurons.
The following criteria were used to identify dopaminergic SNc neurons during whole cell patch-clamp recording (Freestone et al. 2009; Grace and Bunney 2000; Guatteo et al. 2005): 1) location near the ventral border of the medial lemniscus; 2) large fusiform shaped soma (>20 μm); 3) in voltage clamp, the presence of Ih in response to hyperpolarizing voltage commands (−80 and −100 mV; 800 ms); 4) in current clamp, the presence of a depolarizing “sag” in response to hyperpolarizing current steps (up to 200 pA; 800 ms); 5) in current clamp, response to depolarizing current steps (up to 100 pA, 800 ms) in the form of a burst of action potentials exhibiting frequency adaptation; and 6) in current clamp, slow and regular firing (<5 Hz) with zero current. With a K+-based pipette solution, membrane resistance was 97 ± 6 MΩ, membrane capacitance was 150 ± 7 pF, membrane potential was −57 ± 2 mV, and resting [Ca2+]i was 93 ± 11 nM (n = 13).
Response to ADPR.
TRPM2 channels are gated by ADPR, and intracellular application of this compound was previously found to selectively activate these channels (see introduction). To examine functional expression of TRPM2 channels in dopaminergic SNc neurons, the effects of increased cytoplasmic levels of ADPR were tested. Whole cell patch-clamp recordings and fura-2 imaging were employed to monitor whole cell currents and [Ca2+]i concurrently. Data acquisition commenced immediately upon entering the whole cell configuration. Recordings with patch pipettes containing 50–400 μM ADPR, or no drug (controls), were performed (Fig. 3, A and B). At 400 μM, ADPR evoked a desensitizing inward current and [Ca2+]i rise [607 ± 235 pA, P = 0.044; Δ 213.4 ± 82.5 nM, P = 0.048; n(control) = 6, n(ADPR) = 8]. Smaller responses were evoked with 200 μM [228 ± 49 pA, P < 0.01; Δ 94 ± 27 nM, P < 0.01; n(control) = 10, n(ADPR) = 10], while no statistically significant effects were detected with 50 μM [62 ± 74 pA, P = 0.846; Δ 2 ± 12 nM, P = 0.416; n(control) = 10, n(ADPR) = 9]. The peak amplitudes of the ADPR-evoked current and calcium responses were dose dependent (Fig. 3, A and B). These data are consistent with the activation of TRPM2 channels.
The ADPR-induced changes in whole cell current and [Ca2+]i were further examined by using two nonselective TRPM2 channel blockers, clotrimazole and flufenamic acid. Preincubation of slices with clotrimazole (5 μM) reduced both the ADPR (200 μM)-induced inward current (by 85%, P = 0.037) and [Ca2+]i rise (by 65%, P = 0.029, n = 12) (Fig. 3, C and D). Similarly, flufenamic acid (50 μM; n = 12; Fig. 3, C and D) abolished the inward current (reduction by 101%, P < 0.01) and significantly reduced the [Ca2+]i rise (by 80%, P = 0.016).
Response to exogenous H2O2.
Exogenous H2O2, an agent commonly used for inducing oxidative stress because of its relative stability and membrane permeability, was applied to the slice. H2O2 is known to activate TRPM2 channels (see introduction) and, in SNc neurons, ATP-sensitive potassium (KATP) channels (Avshalumov et al. 2007). We assessed the effects of H2O2 on SNc neurons before and after blocking KATP channels.
Bath application of 5 mM H2O2 induced an outward current (232 ± 36 pA, P < 0.01, n = 7; measured 10 min after onset of application; Fig. 4, A and B), consistent with previous studies (Avshalumov et al. 2005; Geracitano et al. 2005), and an increase in [Ca2+]i (Δ 28.6 ± 7.9 nM, P = 0.017; Fig. 4, C and D). During prolonged exposure (>15 min), the integrity of the cell membrane and/or the seal often deteriorated, as indicated by the development of a sudden inward current (>1 nA) and a large rise in [Ca2+]i (not shown). Therefore analysis was restricted to ≤15 min of H2O2 application. To eliminate the KATP channel-mediated response, slices were preincubated with tolbutamide (100 μM) or glibenclamide (10 μM). In the presence of these blockers, the H2O2 (5 mM)-evoked current was significantly reduced (69 ± 13 pA, P = 0.042, n = 6; reduced by 70%, P < 0.01; Fig. 4, A and B). The H2O2-induced [Ca2+]i rise was not attenuated by the KATP channel blockers (Δ 43.7 ± 13.3 nM, P < 0.01; no change, P = 0.333; Fig. 4, C and D). These results show that the H2O2-induced [Ca2+]i rise and a component of the whole cell current evoked by this oxidant are independent of KATP channel activation.
The persistence of outward current in the presence of KATP channel blockers suggests an incomplete inhibition of KATP channels or involvement of another K+ conductance. To determine whether this residual current was masking activation of other (non-K+) channels, recordings were performed with a Cs+-based patch pipette solution containing TEA (see methods). In this condition, the H2O2-evoked outward current was abolished, and instead an inward current was revealed (40 ± 3 pA, P < 0.01, n = 7; reduced by 114% compared with K+-based solution, P < 0.01; Fig. 5, A and B). During the block of K+ conductance, the H2O2-induced [Ca2+]i rise was enhanced (Δ 516.6 ± 93.3 nM, P < 0.01; increased by 266% compared with K+-based solution, P = 0.035; Fig. 5, C and D).
To assess the effects of smaller H2O2 concentrations, additional experiments were performed with 1 and 3 mM H2O2 and the Cs+-based patch pipette solution (Fig. 6). No statistically significant changes in whole cell current and [Ca2+]i were recorded with 1 mM, whereas a significant and reversible change in [Ca2+]i was observed with 3 mM (Δ 275.9 ± 50.3 nM, P < 0.01). The H2O2-induced [Ca2+]i rise clearly demonstrated a dose dependence (P < 0.01).
Effects of TRP channel inhibitors and DPQ on H2O2-evoked responses.
To assess the contribution of TRPM2 channels to the H2O2-evoked responses, two nonselective blockers of TRPM2 channels were used: clotrimazole and ACA. Recordings were performed with the Cs+-based pipette solution to eliminate the effects of KATP and other K+ channels. Slices were treated with the channel blockers for >10 min before and during H2O2 application. In the presence of clotrimazole (5 μM), the H2O2-evoked inward current was reduced by 50% and the [Ca2+]i rise by 64% (P = 0.013 and P < 0.01, n = 4; Fig. 7, A and B). ACA (20 μM) application also resulted in a smaller [Ca2+]i rise (48% reduction, P = 0.022, n = 6; Fig. 7, A and B), but its effect on the inward current was not statistically significant (P = 0.276). Baseline values of whole cell current and [Ca2+]i were not changed by clotrimazole or ACA preincubation.
It has been suggested that TRPM2 channel activation by oxidative stress is mediated through an increase in intracellular ADPR concentration (Buelow et al. 2008; Fonfria et al. 2004; Perraud et al. 2005). This process was demonstrated to be dependent on PARP-1 (polyADPR polymerase), an enzyme upstream of ADPR synthesis. To test whether PARP-1 activity is required for the H2O2-induced responses, DPQ (50 μM; PARP inhibitor), was added to the pipette solution. Baseline whole cell current was larger (i.e., more negative) in neurons loaded with DPQ. However, the effects of the drug on H2O2-evoked changes in whole cell current and [Ca2+]i were not statistically significant (P = 0.374 and P = 0.414, n = 9; Fig. 7, C and D).
Calcium dependence of H2O2-evoked responses.
To determine the source of the H2O2-evoked [Ca2+]i rise, 5 mM H2O2 was applied at least 10 min after standard aCSF (2.4 mM Ca2+) was replaced with “zero” calcium aCSF (same composition as standard aCSF except 0 mM Ca2+ and 1 mM EGTA) (Fig. 8, A and C). As expected, removing extracellular calcium resulted in a reduction of baseline [Ca2+]i (Δ 68.8 ± 10.8 nM; P < 0.01, n = 5). In the absence of extracellular calcium, H2O2 did not evoke statistically significant changes in [Ca2+]i and whole cell current (P = 0.171 and 0.838). During reperfusion with standard aCSF (in the presence of H2O2), a substantial [Ca2+]i rise and inward current were observed (Δ 218.9 ± 55.2 nM, P = 0.017; 31 ± 11 pA, P = 0.049; measured 5 min after the onset of reperfusion; Fig. 8, C and D). These H2O2-induced responses were similar in magnitude to the responses recorded in the experiment in which the extracellular calcium level was maintained at 2.4 mM (Δ 206.9 ± 36.8 nM; 22 ± 3 pA; measured 5 min after the onset of H2O2 application; Fig. 6, A and C). Cs+-based pipette solution was used in these experiments. This indicates that extracellular calcium is the source of the [Ca2+]i rise during H2O2 application. Additional experiments in which extracellular Ca2+ was removed during H2O2 application (K+-based pipette solution) also demonstrated complete inhibition of [Ca2+]i rise (data not shown).
TRPM2 channel activation has been shown to be dependent on resting [Ca2+]i, and only a minimal response was elicited when [Ca2+]i was buffered to zero (Csanady and Torocsik 2009; Starkus et al. 2007). To test the effect of baseline [Ca2+]i on H2O2-induced responses, recordings were made with a “high Ca2+” pipette solution (total [Ca2+] increased to 0.89 mM, calculated free [Ca2+]pipette 750 nM; MAXC, http://maxchelator.stanford.edu) in the presence of tolbutamide (100 μM) or glibenclamide (10 μM). Baseline [Ca2+]i was 71 ± 10 nM in cells patched with the pipettes containing standard K+-based pipette solution (n = 6) and 115 ± 18 nM in neurons patched with the “high Ca2+” solution (n = 7). The discrepancy between the measured and expected free [Ca2+] most likely resulted from endogenous buffering and active Ca2+ extrusion. The H2O2-induced [Ca2+]i rise was significantly higher in cells patched with the “high Ca2+” pipette solution (control: Δ 43.7 ± 13.3 nM, n = 6; high [Ca2+]pipette: Δ 324.5 ± 54.6 nM, n = 6; P < 0.01; Fig. 9, C and D). These results demonstrate a strong dependence of the H2O2-induced [Ca2+]i rise on baseline [Ca2+]i. The effect of higher baseline [Ca2+]i on the H2O2-evoked whole cell current was not statistically significant (P = 0.641; Fig. 9, A and B).
Effects of ionotropic glutamate receptor blockers on H2O2-induced responses.
Previous studies have shown that H2O2 and other ROS can inhibit glutamate transporters (Trotti et al. 1998) and that the effects of exogenous H2O2 are partially mediated by the activation of ionotropic glutamate receptors in hippocampal slices (Avshalumov and Rice 2002). To test the involvement of these receptors in the H2O2-induced responses, a combination of AP5 (100 μM) and CNQX (20 μM) was bath applied 5 min after the onset of H2O2 application. Neither the H2O2-induced current nor the calcium response was affected in the presence of these blockers (data not shown).
Redox-sensitive channels are activated by oxidative stress, an important factor in the pathogenesis of PD and other neurodegenerative disorders (see, e.g., Barnham et al. 2004). Potentially, such channels can transduce a nonspecific pathological increase in ROS levels into specific cellular responses, including activation of cell death pathways. While a number of redox-sensitive TRP channels have been identified (TRPC3, TRPP2, TRPV1, and TRPA1; see introduction), TRPM2 are most likely to act as sensors of oxidative stress in dopaminergic SNc neurons. There are no data supporting the expression of TRPP2 and TRPA1 channels in ventral midbrain. TRPV1 channels have been detected in SNc neurons (Marinelli et al. 2007). However, at least in hippocampal neurons, TRPV1 channels appear to mediate the effect of oxidative stress by modulation of glutamate neurotransmission rather than a direct action on the neurons (Crouzin et al. 2010). TRPC3 channels are expressed in ventral midbrain mainly in oligodendrocytes (Fusco et al. 2004) and only in a small subset of SNc neurons (<10%; Tozzi et al. 2003). Given that all these other channels are unlikely to contribute significantly to the responses of SNc neurons during oxidative stress, we investigated the expression and properties of TRPM2 channels.
TRPM2 mRNA and protein expression in the SNc.
Previous studies reported TRPM2 mRNA expression in the human substantia nigra (Nagamine et al. 1998; Uemura et al. 2005). Consistent with this finding, RT-PCR analysis performed in this study detected TRPM2 mRNA in the rat SNc region. Although a number of human TRPM2 mRNA splice variants have been identified (e.g., Uemura et al. 2005; Wehage et al. 2002; Zhang et al. 2003), no specific splice variants have so far been reported in the rat. In our study, two primer sets were designed based on the rat sequence, which resembles the human full-length variant (TRPM2-L). The results obtained with both sets of primers were consistent, confirming the expression of TRPM2 mRNA in the SNc.
As the tissue collected for mRNA analysis contained a variety of cell types, we used immunohistochemistry to analyze TRPM2 protein expression in individual SNc neurons. Two polyclonal antibodies were used. One was produced by immunization against a human TRPM2 fragment in the channel domain (amino acids 650–700) and the second against a rat TRPM2 fragment located near the COOH terminus in the enzyme domain (amino acids 1430–1508). We have previously tested these antibodies in a study of TRPM2 channel expression in the hippocampus (Bai and Lipski 2010). In the present study, both antibodies produced a similar pattern of staining throughout the midbrain slice. Double-labeling experiments revealed colocalization of TRPM2 and TH, a marker of dopaminergic neurons, providing the first evidence for TRPM2 protein expression in dopaminergic SNc neurons. TRPM2 immunoreactivity was also observed in midbrain slices in regions outside the SNc. This is consistent with reports of TRPM2 channel expression in other cell types (see introduction). A recent study demonstrated that the H2O2-induced increase in spontaneous firing frequency of SNr GABAergic neurons was inhibited by flufenamic acid, consistent with TRPM2 channel expression in these neurons (Lee et al. 2011). The widespread expression of TRPM2 channels suggests that they mediate responses to oxidative stress in many types of cells, and that the high susceptibility of SNc neurons to oxidative damage is not solely due to these channels.
Similar to cultured cortical and hippocampal neurons (Bai and Lipski 2010; Kaneko et al. 2006; Olah et al. 2009), TRPM2 immunoreactivity was not limited to the plasma membrane and a punctuate pattern of staining was observed in the cytoplasm. This could be due to a high level of protein trafficking or to expression in organelles. The latter possibility is supported by a study demonstrating that TRPM2 can also function as a lysosomal Ca2+-release channel (Lange et al. 2009).
Responses of SNc neurons to ADPR.
ADPR was used as an activator of TRPM2 channels, to test for their functional expression in SNc neurons. Whereas oxidative stress activates not only these channels but also a large number of other redox-sensitive processes, intracellular ADPR selectively binds to the NUDT9 domain of TRPM2 protein. The ADPR concentrations required for channel activation have been reported to be in the range of 50 μM to 1 mM, depending on the experimental model. These concentrations lead to inward current and [Ca2+]i rise in heterologous expression systems and in cultured neurons (Hill et al. 2006; Kaneko et al. 2006; Olah et al. 2009; Sano et al. 2001). In the present study, a similar response was observed in SNc neurons in acute brain slice. Both the ADPR-evoked whole cell current and [Ca2+]i rise were dose dependent and desensitized over time. Rundown of TRPM2 channel response has been previously reported in whole cell recordings (Sano et al. 2001) and in excised membrane patches (Csanady and Torocsik 2009; Hill et al. 2006). While the mechanism underlying the desensitization is unknown, it was suggested to be due to a reduction in the number of active channels (Csanady and Torocsik 2009). Alternatively, the rundown can be due to enzymatic degradation of ADPR or insufficient levels of free [Ca2+]i needed to sustain the response. Previous studies have demonstrated that intracellular calcium is required both for TRPM2 channel activation and to maintain their open state (Csanady and Torocsik 2009; Starkus et al. 2007). Therefore, it is possible that the baseline [Ca2+]i (93 ± 11 nM, n = 13) and the Ca2+ buffering capacity of the pipette solution permitted channel activation but were unable to sustain the response. This interpretation is in line with a previous study conducted on hippocampal neurons, in which intracellular application of ADPR under normal conditions did not elicit a current response (Olah et al. 2009). However, when neurons were depolarized to activate voltage-gated Ca2+ channels and raise [Ca2+]i, application of ADPR was able to evoke a TRPM2 channel-mediated current.
The ADPR-induced changes in membrane current and [Ca2+]i were strongly inhibited with two nonselective TRPM2 channel blockers, flufenamic acid and clotrimazole. Flufenamic acid is a nonsteroid anti-inflammatory agent that also modulates other TRP channels, as well as voltage-gated Na+ and K+ channels, Cl− channels, and GABAA receptors (Hill et al. 2004b). Clotrimazole is an antifungal agent that also blocks sarco(endo)plasmic reticulum Ca2+-ATPase, voltage-gated Ca2+ channels, and Ca2+-activated K+ channels (Hill et al. 2004a; Naziroglu et al. 2007). In the absence of more selective blockers, these drugs have been routinely used to identify TRPM2 channel responses in heterologous expression systems, primary cell cultures, and brain slices (Fonfria et al. 2006; Hill et al. 2006; Olah et al. 2009; Togashi et al. 2006; Wilkinson et al. 2008). The responses evoked in our study by intracellular application of ADPR, and the inhibition of these responses by TRPM2 blockers, are consistent with functional expression of TRPM2 channels in dopaminergic SNc neurons. However, although unlikely, we cannot exclude the possibility that ADPR acted through a yet-unknown mechanism, independent of TRPM2 channel activation, which was also affected by flufenamic acid and clotrimazole.
Responses of SNc neurons to H2O2.
Exogenous application of H2O2 was used to determine whether TRPM2 channels are involved in the response of SNc neurons to oxidative stress. H2O2 was chosen since this oxidant has been suggested to activate TRPM2 channels by direct interaction (Naziroglu and Luckhoff 2008; Wehage et al. 2002). Although H2O2 is a cellular toxin at high concentrations, there is substantial evidence that at low concentrations it also functions as a signaling molecule (Avshalumov et al. 2007; Veal et al. 2007). In the nigrostriatal system, H2O2 regulates dopamine release from SNc neurons via a KATP channel-dependent mechanism (Chen et al. 2001, 2002). Although the effects of H2O2 on KATP channel activation in SNc neurons are well documented (Avshalumov et al. 2005; Geracitano et al. 2005), it is not known whether this oxidant also activates TRPM2 channels in these neurons.
The effects of H2O2 have been tested in brain slices typically at high micromolar or low millimolar concentrations (Avshalumov et al. 2005; Geracitano et al. 2005; Katsuki et al. 1997; Nistico et al. 2008). Similar concentrations were also used to activate TRPM2 channels in expression systems (Hill et al. 2006; Kaneko et al. 2006; Naziroglu and Luckhoff 2008; Wilkinson et al. 2008). At 0.2–5 mM, H2O2 only moderately affected mitochondrial function in hippocampal slices (Gerich et al. 2009), indicating minimal toxicity. Although in the present study H2O2 was bath applied at 1–5 mM, the actual concentration in the tissue is likely to be significantly lower because of degradation by glutathione peroxidase and catalase (Dringen et al. 2005). For example, in Jurkat T cells, cytoplasmic H2O2 levels were estimated to be one-seventh of the extracellular concentration (Antunes and Cadenas 2000).
H2O2 induced an outward current that was only moderately inhibited in the presence of KATP channel blockers. We hypothesized that either a residual KATP current or another unidentified outward current was masking the expected TRPM2 channel-mediated inward current. Consistent with this hypothesis, blocking all K+ conductances with Cs+ and TEA revealed a H2O2-induced inward current that was sensitive to TRPM2 channel blockers (see below).
H2O2 evoked an increase in [Ca2+]i, consistent with opening of Ca2+-permeable TRPM2 channels. As expected, this response was insensitive to KATP channel blockers, but, surprisingly, the [Ca2+]i increase was enhanced when K+ conductances were blocked. This may be a consequence of improved space-clamp conditions associated with increased cell input resistance. Alternatively, it can be hypothesized that, in this experimental model, cesium inhibits Na+-K+-ATPase. Degradation of the transmembrane Na+ gradient would lead to reduced efficiency of the Na+/Ca2+ exchanger and a slower rate of Ca2+ efflux.
H2O2 failed to evoke a [Ca2+]i rise in the absence of extracellular calcium, demonstrating an extracellular source of this ion. This is consistent with activation of TRPM2 channels in the plasma membrane. In both cultured striatal and cortical neurons, the H2O2-evoked [Ca2+]i rise was also dependent on extracellular calcium (Kaneko et al. 2006; Smith et al. 2003). In contrast, Gerich et al. (2009) reported that in cultured hippocampal neurons the H2O2-evoked [Ca2+]i rise was due to Ca2+ release from intracellular stores. The discrepancy between these and our results suggests that the mode by which H2O2 induces [Ca2+]i rise is cell type specific. Alternatively, it could be due to differences in experimental conditions. Gerich et al. studied calcium responses in cultured neurons that, unlike cells in brain slices, are deprived of their normal environment and hence may differ in calcium handling. In addition, in their experiments the cell membrane potential was not clamped, and therefore other voltage-sensitive mechanisms may have influenced [Ca2+]i.
Activation of TRPM2 channels highly depends on resting [Ca2+]i, a property believed to be mediated by calmodulin (Starkus et al. 2007; Tong et al. 2006). In fact, sufficiently high [Ca2+]i can activate TRPM2 channels by itself (Du et al. 2009). In the present study, the effect of baseline [Ca2+]i on the H2O2-evoked [Ca2+]i rise was tested by increasing [Ca2+]pipette, and consequently in patched neurons. In this condition ([Ca2+]i 115 ± 18 nM, n = 7), application of H2O2 evoked a larger [Ca2+]i increase, consistent with this property of TRPM2 channels.
The H2O2-evoked [Ca2+]i rise was significantly reduced by ACA and clotrimazole. ACA is another nonselective TRPM2 channel blocker that is structurally similar to flufenamic acid (Harteneck et al. 2007). ACA was used in these experiments since flufenamic acid caused instability during prolonged patch-clamp recordings, possibly due to a mild toxicity effect. The H2O2-induced inward current, unmasked with Cs+ and TEA, was reduced after preincubation with clotrimazole but not ACA. The inconsistent effect on whole cell current suggests that a significant component of the response was mediated by other unidentified channels. However, the effectiveness of both blockers in inhibiting the [Ca2+]i rise supports TRPM2 channel activation in SNc neurons.
It is still debated whether H2O2-mediated activation of TRPM2 channels requires ADPR as a diffusible messenger. In support of a direct gating mechanism, others have reported TRPM2 activation by H2O2 in not only whole cell but also membrane patch recordings (see, e.g., Hill et al. 2006; Naziroglu and Luckhoff 2008). Our results are consistent with this mechanism since H2O2-evoked current and Ca2+ responses were observed despite cytoplasmic dialysis associated with whole cell patch-clamp recording. However, there is evidence indicating that oxidative stress-mediated activation of TRPM2 channels requires PARP-1 (Buelow et al. 2008; Fonfria et al. 2004; Perraud et al. 2005), an enzyme involved in the synthesis of ADPR and commonly associated with repair of damaged DNA (Kuhn et al. 2005). In our experiments, the PARP inhibitor DPQ did not inhibit the H2O2-evoked inward current and [Ca2+]i rise. While this suggests that in our system PARP-1 activity is not required for oxidative stress-mediated activation of TRPM2 channels, ADPR may still act as a diffusible messenger that originates from another enzymatic pathway or is released form an intracellular store (Partida-Sanchez et al. 2007).
Comparison between responses evoked by ADPR and H2O2.
Our results indicate activation of TRPM2 channels by exogenous ADPR and H2O2 in SNc neurons. However, two major differences were observed between the responses evoked by these two stimuli. First, ADPR evoked a significantly larger inward current. Second, nonselective TRPM2 blockers were more effective at blocking ADPR-evoked responses. These results are consistent with the concept that ADPR is a potent, selective activator of TRPM2 channels. In contrast, H2O2 potentially not only activates TRPM2 channels but also modulates other redox-sensitive processes.
Our results provide novel information on TRPM2 mRNA and protein expression in SNc neurons and characterize changes in membrane currents and [Ca2+]i evoked by TRPM2 channel activators. Intracellular application of ADPR evoked a rise in [Ca2+]i and an inward current, while H2O2 evoked a complex response comprising a [Ca2+]i rise and an inward current that was masked by activation of K+ conductance. Although our study focused on the oxidative stress-sensing function of TRPM2 channels, there are also considerable data on the role of these channels in mediating cell death (see introduction). In the context of PD, a disorder in which a number of factors contribute to degeneration of SNc neurons (see, e.g., Sulzer 2007), TRPM2 channel may act as a central element that integrates factors such as oxidative stress, calcium dysregulation, and mitochondrial dysfunction (Freestone et al. 2009). These channels may also play a role in other neurodegenerative and acute brain disorders associated with an increased level of oxidative stress. These considerations, as well as our observations regarding TRPM2 channel function in SNc neurons, warrant further studies to determine the pathophysiological role of these channels in neurological disorders.
This work was supported by the Marsden Fund of the Royal Society of New Zealand. K. K. H. Chung was supported by the Tertiary Education Commission of New Zealand and P. S. Freestone by the Neurological Foundation of New Zealand.
No conflicts of interest, financial or otherwise, are declared by the author(s).
Author contributions: K.K.H.C. and J.L. conception and design of research; K.K.H.C. and P.S.F. performed experiments; K.K.H.C. and P.S.F. analyzed data; K.K.H.C., P.S.F., and J.L. interpreted results of experiments; K.K.H.C. prepared figures; K.K.H.C. drafted the manuscript; K.K.H.C., P.S.F., and J.L. edited and revised the manuscript; K.K.H.C., P.S.F., and J.L. approved the final version of the manuscript.
- Copyright © 2011 the American Physiological Society