Microglial cells are the host macrophages in the central nervous system and respond to brain injury and various neurological diseases. In this process, microglial cells undergo multiple morphological and functional changes from the resting cell toward a fully activated, phagocyting tissue macrophage. In culture, bacterial lipopolysaccharide (LPS) is a frequently used tool to induce this activation. By using calcium-imaging and patch-clamp techniques, we investigated the effect of hydrogen peroxide (H2O2), which is released by macrophagic cells themselves, on the intracellular calcium concentration and ion currents in cultured rat microglia. Application of 0.1–5 mM H2O2 for several minutes induced small responses in untreated cells but a large calcium influx and cation current in LPS-treated cells. In both untreated and LPS-treated microglia, internal perfusion of ADP-ribose (ADPR) via the patch pipette elicited large cation currents. Both stimuli, H2O2 and ADPR, have been reported to activate the recently cloned nonselective cation channel TRPM2. RT-PCR analysis from cultured rat glial and neuronal cells confirmed a strong expression of TRPM2 in rat microglia but not in astrocytes and cerebellar granule cells. In situ hybridizations from mouse brain showed a distribution of TRPM2, which is compatible with the expression in microglial cells. In conclusion, we describe here a novel calcium influx pathway in microglia coupled to hydrogen peroxide and ADPR and provide evidence that this pathway involves TRPM2. The increased sensitivity to H2O2 in LPS-stimulated cells suggests a role for TRPM2 in the calcium signaling of activated microglia.
- nonselective cation channel
- transient receptor potential channel
- activated microglia
microglial cells are activated in response to brain injury and various neurological diseases. Several stimuli, including bacterial cell wall components such as lipopolysaccharide (LPS), induce this activation process. Like other activated macrophages, microglia remove bacteria and cellular debris and produce a diverse range of mediators of the inflammatory cascade, including proteases, arachidonic acid derivatives, cytokines, nitric oxide (NO), and hydrogen peroxide (H2O2). Thus microglial cells are a key factor in the immune defense and tissue repair in the central nervous system (CNS).
A variety of receptors and receptor-ligand interactions leading to Ca2+ signals have been described in microglial cells (21). Because multiple cellular functions depend on the intracellular Ca2+ concentration ([Ca2+]i), it may also play an important role in the activation process of microglia. Recently, the chronic elevation of basal [Ca2+]i has been shown in cultured mouse microglia activated by LPS (9). The increased [Ca2+]i was associated with a reduced signaling upon stimulation with UTP and complement factor 5a (9). However, intensified or specific Ca2+ signals in activated microglia have not been reported to our knowledge. A potential extracellular mediator inducing elevation of [Ca2+]i in these cells may be the mild oxidant H2O2. Recently, it has been discovered that H2O2, released by macrophagic cells themselves, not only induces stress responses and cell death but is also a second messenger for signal transduction and signal amplification in lymphocytes (1, 17). Furthermore, it has been suggested that the production of H2O2 in macrophages is involved in the activation of these and neighboring cells (28). The underlying mechanism of this H2O2-induced signaling seems to be a modulated protein phosphorylation through cysteine oxidation (29). Another mechanism, coupling H2O2 to an influx of Ca2+, was discovered by the molecular and functional characterization of the nonselective cation channel TRPM2. This channel, a member of the superfamily of transient receptor potential (TRP) proteins, is activated by H2O2 (6, 32) and intracellular ADP-ribose (ADPR) (26, 30). TRPM2 has been described in neutrophil granulocytes (7) and the monocytic cell line U937 (26, 30).
In this study, we investigated the effect of H2O2 on activated and unstimulated microglia. Micromolar concentrations of H2O2 induced a robust increase in Ca2+ influx and cation currents in activated cells. ADPR elicited cation currents in both activated and unstimulated cells. Characteristics of these currents and results of RT-PCR and in situ hybridizations strongly suggest the expression of TRPM2 and its involvement in microglial calcium signaling and cation influx.
MATERIALS AND METHODS
Primary cell cultures. Glial cell cultures were prepared from brains of newborn Wistar rats (Tierzucht Schönwalde, Germany) as described previously (4). In brief, brain tissue was carefully freed of blood vessels and meninges. Tissue was trypsinized with 1% trypsin and 0.05% DNase in Hanks' buffered saline solution (HBSS) for 5–8 min, dissociated with a fire-polished pipette, and washed twice. Mixed glial cells were cultured for 7–9 days in Dulbecco's modified Eagles medium (DMEM; Invitrogen, Groningen, The Netherlands) supplemented with 10% fetal calf serum (FCS) (Invitrogen), 100 U/ml penicillin, and 100 μg/ml streptomycin under 5% CO2 atmosphere at 37°C. Microglial cells were then separated from the underlying astrocytic monolayer by gentle shaking of the bottles. Supernatants were collected and centrifuged, and the microglial cell pellet was immediately frozen in liquid nitrogen and stored at –80°C until RNA preparation. One aliquot of the same preparation was resuspended and plated on glass coverslips at a density of 105 cells/cm2 for further experiments, which were performed 1–2 days after seeding. For some experiments, microglial cells were activated with 100 ng/ml LPS (E. coli K-235 or Salmonella minnesota L-9764; Sigma, Deisenhofen, Germany) for 20–26 h.
Astrocyte cultures were separated by removal of remaining microglia and O2A-precursor cells by vigorous shaking of the bottles (18). Supernatants were discarded, and monolayers were washed twice with Dulbecco's phosphate-buffered saline (Invitrogen) to remove nonadherent cells and cell debris. The monolayer was then trypsinized with 0.05% trypsin and 0.02% EDTA for 10 min. Cells were collected in DMEM and centrifuged, and the astrocytic cell pellet was immediately frozen in liquid nitrogen and stored at –80°C until RNA preparation.
Cerebellar granule cells (CGC) were prepared from 5-day-old Wistar rats as described previously (3). Briefly, cerebelli were aseptically removed from skulls and freed from the meninges. The tissue was chopped with a scalpel and disintegrated by 1% Trypsin and 0.05% DNase for 7 min at 37°C. Trypsin was blocked by adding an excess of FCS-containing medium. The cells were carefully triturated with a fire-polished pipette. After centrifugation, the cells were filtered through 40 μm of nylon mesh and seeded on poly-l-lysine-coated (PLL) culture dishes at a density of 3 × 105 cell/cm2. Neurons were cultured in Eagle's basal medium (BME) supplemented with 10% heat-inactivated FCS, 25 mM KCl, 2 mM l-glutamine, and penicillin-streptomycin. Twenty-four hours after plating, the medium was exchanged by defined serum-free medium (DMEM supplemented with 32 mM glucose, 25 mM KCl, 10 μg/ml insulin, 300 μM Na-selenite, 62 ng/ml progesterone, 16 μg/ml putrescine, 100 μg/ml transferrin, 400 ng/ml triiodothyronine, 4 mM thyroxin, 2 mg/ml gentamycin, 10 μg/ml BSA, and 2 mM l-glutamine) containing 10 μM cytosine arabinofuranoside to prevent proliferation of nonneuronal cells. Neurons were cultivated for 7 days in vitro at 37°C under 5% CO2. The cell pellet was then harvested, immediately frozen in liquid nitrogen, and stored at –80°C until RNA preparation.
Culture and transfection of HEK-293 cells. HEK-293 cells were cultured in Earle's minimal essential medium (Invitrogen), supplemented with 10% FBS, 100 U/ml penicillin, and 100 μg/ml streptomycin under 5% CO2 atmosphere at 37°C. Cells were plated in 85-mm dishes onto glass coverslips and transiently transfected with 1.5–2 μg of DNA and 5 μl of FUGENE 6 transfection reagent (Roche, Indianapolis, IN) in 95 μl of OptiMEM medium (Invitrogen) 2 days later. Two overlapping cDNA fragments were amplified and verified according to the sequence of human TRPM2 (26). The fragments were subcloned in pcDNA3.1 vector to generate a cDNA allowing cytomegalovirus promotor-driven expression of a TRPM2-YFP protein. Fluorescent measurements and electrophysiological studies were carried out 2–3 days after transfection.
Calcium measurements. [Ca2+]i was monitored using the fluorescent Ca2+ indicator fura 2-AM. Cells were loaded in a standard bath solution composed of 140 mM NaCl, 5 mM KCl, 1 mM MgCl2, 2 mM CaCl2, 10 mM glucose, and 10 mM HEPES (pH 7.4 with NaOH), supplemented with 4 μM fura 2-AM (Molecular Probes, Eugene, OR) and 0.005% Pluronic F-127 (Molecular Probes) for 60 min at room temperature. Cells were then rinsed with extracellular solution and allowed to de-esterify for at least 30 min at room temperature. Measurements of [Ca2+]i were performed by using an inverted microscope (Axiovert 100, Zeiss) equipped with a Fluar objective (×40/1.30, oil immersion, Zeiss) and an imaging system (T. I. L. L. Photonics, Gräfelfing, Germany). The illumination was generated by a xenon arc lamp. Monochromator settings and data acquisition were controlled by Fucal 5.12 C software (T. I. L. L. Photonics). The fluorescence emission of rectangular areas within single cells was long-pass filtered at 520 nm and recorded with a charge-coupled device camera. After correction for background fluorescence, the fluorescence ratio F340/F380 was calculated. For calibration of fluorescence signals in terms of Ca2+ concentration, we performed intracellular measurements as described previously (14).
Patch-clamp experiments. Membrane currents were measured in whole cell, cell-attached, and outside-out configurations of the patch-clamp technique. Currents were recorded using an EPC-7 amplifier (HEKA, Lambrecht, Germany), subsequently low-pass filtered at 1 kHz, digitized with a sampling rate of 5 or 10 kHz, and analyzed using pCLAMP software (version 7.0; Axon Instruments). The pipette resistance varied between 2 and 5 MΩ. Whole cell currents were elicited by voltage ramps from –100 to +100 mV (400-ms duration) applied every 5 s from a holding potential of 0 mV.
Pipettes for whole cell recordings were filled with a solution composed of 130 mM CsCH3O3S, 10 mM CsCl, 2 mM MgCl2, and 10 mM HEPES (pH 7.2 with CsOH). In some cases, this solution contained 1 mM ADPR or 1 mM NAD+. The standard bath solution contained 140 mM NaCl, 2 mM CaCl2, 1 mM MgCl2, and 10 mM HEPES (pH 7.4 with NaOH). For Na+- and divalent cation-free conditions, the bath solution contained 140 mM N-methyl-d-glucamine (NMDG+) and 10 mM HEPES (pH 7.4 with HCl). ADPR, NAD+, H2O2, and all other chemicals were obtained from Sigma. Relative cation permeabilities were calculated as described previously (5).
Single-channel currents were continuously measured at different potentials. Pipettes were either filled with 130 mM CsCH3O3S, 10 mM CsCl, 2 mM MgCl2, and 10 mM HEPES (pH 7.4) for outside-out recordings or with 130 mM NaCl, 2 mM CaCl2, 1 mM MgCl2, and 10 mM HEPES (pH 7.4) for cell-attached recordings. Single-channel amplitudes at different potentials were calculated from current traces of 2–4 s using amplitude histograms fitted to Gaussian functions.
Extraction of mRNA and RT-PCR. The frozen cell pellets were quickly thawed, and total RNA was isolated using RNeasy (Qiagen, Hilden, Germany). Approximately 10–20 μg of total mRNA were isolated from (5–10) × 106 microglial cells, 5–10 μg of total mRNA from (4–6) × 106 granule cells, and 20–30 μg of total RNA from (5–10) × 106 astrocytes. Poly(A)+ RNA was purified from total RNA using Oligotex (Qiagen) and was reverse-transcribed to generate first-strand cDNAs. PCR analysis was performed with these cDNAs by using the oligonucleotides (sense: 5′-AGCGCCCGGCATCTCCTCTATC-3′, antisense 5′-ATTGTCTGTGTTCCTTGGGTCATCC-3′) for 30 cycles at 94°C for 30 s, 50°C for 30 s, and 72°C for 2 min. As a control, glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was amplified using gene-specific primers.
Northern blot and in situ hybridizations. The cDNA coding for mouse TRPM2 was amplified from a Marathon mouse brain cDNA library (BD Biosciences, Clontech, Heidelberg, Germany) by using the oligonucleotides (sense: 5′-TCTCTTGGGTGTTTTTATTTT-3′, antisense 5′-GCCCAGAAGCCACAGTCAGA-3′) and was subcloned in TOPOpcDNA3.1 cloning kit (Invitrogen). The clones carrying the insert in sense and antisense orientation were isolated; the sequences were determined and used as probes for Northern blot and in situ hybridization.
Total RNA from various mouse tissues was extracted using Trizol reagent (Invitrogen). A standard agarose gel checked integrity of RNA. For Northern blot hybridization, 10 μg of total RNA per lane were loaded on a denaturing formamide/formaldehyde agarose gel. After electrophoresis, the size-fractionated RNA was transferred by capillary transfer on a charged nylon membrane (Tropilon Plus; Applied Biosystems, Weiterstadt, Germany). RNA was fixed on the membrane surface by ultraviolet light and by heating in an incubator at 80°C for 2 h. DNA fragments (0.5, 0.7, and 0.9 kB fragments after digestion of mTRPM2 with PstI and XbaI) were radiolabeled according to a standard random-priming procedure using [α-32P]dATP in parallel reactions. Unincorporated nucleotides were removed, and a mixture of all reactions was used as probe for hybridization. Hybridization and washing steps were performed with Ultrahyb hybridization mix (Ambion, Huntingdon, UK) according to the standard protocol. Posthybridization wash steps were performed at 55°C beginning with 2× SSC and 0.1% (wt/vol) SDS and ending up with 0.2× SSC + 0.1% (wt/vol) SDS. The blots were exposed for 48 h with intensifying screen on KODAK XAR films.
Paraffin sections (5 μm) from mouse tissues were mounted on silane-coated glass slides. DNA prepared from two different E. coli clones carrying the entire reading frame of mTRPM2 in sense and antisense orientations was used for generation of probes. Plasmid DNAs were linearized by XhoI recognizing the corresponding sites of the vector backbone (pcDNA3.1). In vitro transcription was performed using T7 MAXIscript RNA polymerase (AMBION) in the presence of [α-35S]UTP. Probe length was adjusted to 150 bases by alkaline treatment. Sections were partially digested by 0.5% pronase in 0.2 M HCl and acetylated with 0.25% acetic anhydride in 0.1 M triethanolamine (pH 8.0) after deparaffinization and rehydration. Hybridization was carried out at 50°C for 16–18 h with antisense and sense RNA at 350,000 cpm in hybridization buffer containing 50% formamide, 0.3 M NaCl, 10 mM NaPO4, pH 6.8, 0.02% polyvinylpyrrolidone, 0.02% Ficoll 400, 0.02% BSA, 1 mg ml–1 yeast tRNA, 5 mM EDTA, 10 mM dithiothreitol (DTT), and 10% dextran sulfate. Sections were washed in posthybridization solution containing 50% formamide, 10 mM DTT, 0.5 M NaCl, 0.1 M Tris·HCl, 0.05 M EDTA, and 0.1 M NaH2PO4/Na2HPO4, pH 6.8, at 52°C for 30 min and 4 h and were then incubated in TES buffer (10 mM Tris·HCl, pH 7.5, 1 mM EDTA, and 0.5 M NaCl) at 37°C for 15 min. After treatment with RNase A (20 μg ml–1 in TES buffer) for 30 min at 37°C and after incubation in TES buffer at 37°C for 15 min, slides were washed in 2× SSC and 0.1× SSC for 30 min at 21°C. After dehydration, the slides were exposed to Kodak BIOMAX film for 6 days and to Kodak NTB-2 film solution for 4 mo. After development, slides were counterstained with hematoxylin/eosin.
Statistical analysis. All pooled data from fluorescent measurements are given as means ± SE from n experiments containing 10–14 cells, unless otherwise noted. All pooled data from patch-clamp experiments are expressed as means ± SE from n cells. To test for significant differences, a nonpaired t-test was used.
H2O2-induced calcium influx in microglia. To examine the response of H2O2 on cultured microglial cells, we performed side-by-side experiments on both untreated and LPS-treated cells, loaded with the fluorescent Ca2+ indicator fura 2-AM. LPS has been validated to induce activation of microglial cells and is commonly used in studies on inducible macrophage/microglia functions and bacterial CNS infections. Responses to a high concentration of H2O2 (5 mM) in both activated and control cells are shown in Fig. 1A. In most activated cells (>90%), application for 2 min induced a biphasic change in [Ca2+]i: a rapid, transient increase followed by a second increase reaching a plateau. The initial [Ca2+]i transient was also present in the absence of extracellular Ca2+ (data not shown), suggesting a release of Ca2+ from intracellular stores. In control cells, the transient response was almost absent and the second [Ca2+]i increase reached lower levels. In both preparations, however, the second part of the Ca2+ signal persisted after washout of H2O2 and disappeared in the absence of extracellular Ca2+, suggesting an irreversible activation of Ca2+ influx via the plasma membrane. To test whether a more physiological concentration of H2O2 also induces responses in microglia, we applied 100 μM H2O2 for >2 min to activated and control cells. We found that only a prolonged application of 5–10 min sufficed to induce an increase in [Ca2+]i in both preparations (Fig. 1B). These Ca2+ signals were irreversible, disappeared in Ca2+-free solution, and showed larger amplitudes in activated cells, similar to the phase of the response observed with 5 mM H2O2. We further tested the effect of La3+ (100 μM), a blocker of several Ca2+ conducting pathways, on the H2O -induced Ca2+ influx. A consistent inhibition was not observed; La3+ induced only a very small and transient decrease in [Ca2+]i in both activated and control cells (Fig. 1B). The increases in [Ca2+]i evoked by application of H2O2 (5 mM for 2 min and 100 μM for 10 min) are summarized in Fig. 1C. Changes in [Ca2+]i were calculated from values before application and after washout of H2O2. The [Ca2+]i increases evoked by 5 mM H2O2 were 94 ± 3 nM (n = 7) and 268 ± 28 nM (n = 9) for control and activated cells, respectively. With 100 μM H2O2, control and activated cells showed an elevation by 119 ± 17 nM (n = 6) and 357 ± 48 nM (n = 6), respectively. The basal [Ca2+]i in control and activated cells estimated before application of H2O2 was 190 ± 7 nM (n = 13) and 221 ± 6 nM (n = 15), respectively. In conclusion, micromolar H2O2 induced a long-lasting Ca2+ influx in microglia, which is clearly increased in activated cells.
H2O2-induced currents in activated microglia. Single cells were analyzed using the whole cell and cell-attached patch-clamp technique. To minimize the contribution of K+ channels to the whole cell currents, the intracellular solution contained Cs+ instead of K+ and the extracellular solution was K+ free. We tested the response of H2O2 on both control (untreated) and activated (LPS-treated) microglial cells. Control cells showed no significant responses on application of 0.5 or 5 mM H2O2 for 2–7 min (n = 5). The same treatment, however, gradually induced currents in activated cells (n = 6 of 10). When extracellular cations were replaced by the large cation NMDG+, H2O2-induced currents were strongly reduced in the inward direction, but not in the outward direction (Fig. 2, A and B). After substitution of NMDG+ by Na+ and divalent cations, inward currents were restored, shifting the reversal potential close to 0 mV (Fig. 2B), consistent with permeation through nonselective cation channels.
We further used cell-attached measurements for investigation of single-channel properties underlying the H2O2-induced whole cell current. Before the experiments, LPS-treated microglial cells were bathed in a solution containing 5 mM H2O2 for about 5 min because activated cells were not stable enough in the cell-attached configuration over a long period of time. In 3 of 6 cells, we obtained channel openings as shown in Fig. 2C. The channel activity was characterized by a long open time over several hundred milliseconds and absolute values for unitary currents of 3–4 pA at test potentials of –50 and +50 mV. This suggests the activation of channels showing a linear current-voltage (I-V) relationship and a single-channel conductance of 60–80 pS. In untreated (n = 5) and LPS-treated (n = 7) cells, which were not incubated with H2O2, similar channel activity was absent.
A long open time and a conductance of 58–76 pS for single channels (26, 30) and the gradual activation of currents by H2O2 in perforated-patch (6) and conventional whole cell recordings (32) have been reported for heterologously expressed TRPM2. We therefore suppose the involvement of this channel in H2O2-induced currents found in microglial cells.
Characteristics of ADPR-induced currents in microglia. To obtain further evidence for the contribution of TRPM2 to microglial function, we tested the effect of intracellular ADPR on whole cell currents in control and activated cells. The nucleotide ADPR has been reported to open TRPM2, very likely by a mechanism involving a COOH-terminal region of TRPM2 with specific ADPR pyrophosphatase activity (yielding AMP and ribose 5-phosphate) (26). Half-maximal activation of TRPM2 was observed at 70–90 μM ADPR and maximal activation at 300–500 μM ADPR (26, 30). Infusion of 1 mM ADPR via the patch pipette by establishing the whole cell configuration rapidly evoked currents in untreated (Fig. 3A) and LPS-treated microglia (Fig. 3B). These currents were reversibly suppressed by substitution of Na+ and divalent cations by NMDG+ (Fig. 3, A and B). The I-V relationship of ADPR-induced currents was linear and showed a reversal potential near 0 mV (Fig. 3C), indicative of nonselective cation channels. The expression of ADPR-gated channels in untreated and LPS-treated cells of the same preparation was estimated from current amplitudes at –80 mV. The mean ADPR-induced currents, sensitive to NMDG+, were –1.18 ± 0.36 nA (n = 5) and –1.14 ± 0.10 nA (n = 3) in activated and unstimulated cells, respectively. To determine whether tested cells showed properties of unstimulated or activated microglia, we analyzed currents evoked immediately after establishing the whole cell configuration. Unstimulated microglial cells lack outward currents activated at potentials positive to –60 mV (12). However, upon exposure to inflammatory agents (such as LPS), cultured microglia express an additional voltage-gated K+ current activated positive to –50 mV (24, 25). Figure 3D shows the initial currents in both cell types before development of the ADPR-induced conductance. In most LPS-treated cells, depolarization positive to –30 mV elicited a voltage-dependent outward current, whereas all untreated cells showed a linear I-V relationship. Our bathing solution containing zero K+ masked all inward K+ conductances. The inhibition of outward K+ currents by internal Cs+, however, was delayed depending on the perfusion rate of the pipette solution.
From the stable activation of large whole cell currents by ADPR, we suggested a considerable high density of active channels in membrane patches from microglial cells. We therefore produced outside-out patches from successful whole cell configurations after infusion of ADPR. Figure 3E shows a representative recording of single-channel activity in the presence of 1 mM ADPR at the internal face of the patch. Sometimes single-channel events were only resolvable after rundown of channel activity, indicating a considerable number of active channels. The channels were characterized by a long open time of several hundred milliseconds and a slope conductance of 65 ± 3 pS (n = 3) at negative membrane potentials. At voltages between –70 and –20 mV, a linear I-V relationship was observed. Recordings at positive membrane potentials were not evaluated because the stability of the patch decreased dramatically.
We next verified a possible inhibition of ADPR-induced currents by La3+, a blocker of most Ca2+-permeable channels and some members of the TRP family. However, application of increasing concentrations of La3+ (0.1–1 mM) did not significantly block the ADPR-evoked current in microglia (n = 5; Fig. 3F). Addition of 100 μM La3+ to the standard bath solution sometimes induced a small and transient decrease, followed by a slight potentiation of currents (Fig. 3F). This effect and the poor efficacy of La3+ on ADPR-induced current at all are in accordance with its weak effect on Ca2+ responses (see Fig. 1B). To verify the action of La3+ on cloned TRPM2-channels, we transfected HEK-293 cells with a TRPM2-YFP construct and performed similar whole cell patch-clamp experiments. External La3+ (0.1–1 mM) was also not sufficient to induce a significant block of heterologously expressed TRPM2 (data not shown).
Besides ADPR, the nucleotide NAD+ (the oxidized form of NADH) has been reported to activate TRPM2 (6, 30). We compared the efficacy of ADPR and NAD+ on microglial currents by performing alternating whole cell experiments with either 1 mM ADPR or 1 mM NAD+ in the pipette solution. Single-cell experiments of 3–7 min in duration were performed from three independent microglial cultures (data not shown). The mean NMDG+-sensitive currents at –80 mV evoked by ADPR and NAD+ were –1.91 ± 0.34 nA (n = 9) and –0.04 ± 0.02 nA (n = 9), respectively. To verify the missing effect of NAD+ in microglia, we recorded whole cell currents in TRPM2-transfected HEK-293 cells. Perfusion with 1 mM NAD+ for at least 4 min was also not sufficient to induce currents through TRPM2 (data not shown).
We further studied permeation properties of the ADPR-gated whole cell currents with respect to the main divalent cations Ca2+ and Mg2+. Standard bath solution containing Na+ and divalent cations was first exchanged by NMDG+ solution and subsequently by 100 mM CaCl2. Application of NMDG+ completely suppressed inward currents, whereas 100 mM CaCl2 partially restored the currents (Fig. 4A). To estimate permeability for Mg2+, we consecutively applied NMDG+ solution, standard solution, and 100 mM MgCl2. Inward currents were clearly increased in the presence of Mg2+ compared with NMDG+ solution (Fig. 4B). Respective I-V relationships were constructed from the measurements examining permeability for either Ca2+ (Fig. 4C) or Mg2+ (Fig. 4D). These I-V relationships yielded reversal potential (Erev) values of +1.4 ± 0.2 mV (n = 8), –5.4 ± 0.2 mV (n = 5), and –13 ± 0.6 mV (n = 3) for standard solution, 100 mM CaCl2, and 100 mM MgCl2, respectively. From the shifts of Erev during single experiments, we calculated permeability ratios as described previously (5). The ion concentrations were corrected for the respective activity coefficients (0.76 for 140 mM NaCl, 0.517 for 100 mM CaCl2, and 0.535 for 100 mM MgCl2). Thus permeability ratios of ADPR-gated channels in microglia were 0.71 ± 0.01 for PCa/PNa (n = 5) and 0.47 ± 0.01 for PMg/PNa (n = 3).
Expression of TRPM2 in cultured glial and neuronal cells from rat brain. To identify rat TRPM2 in our cultured microglial cells, we performed RT-PCR experiments on total RNA using a specific primer set for mouse TRPM2. We were furthermore interested in whether other cells of glial/neuronal origin express TRPM2. Cerebellar granule cells and astrocytes derived from newborn rats were cultured for acquisition of RNA. In parallel experiments, an amplified product of the expected size (588 bp) was only detected in microglia (n = 7) but not in granule cells (n = 3) and most astrocytic preparations (n = 4 of 7) (Fig. 5). Because some RT-PCR experiments from astrocytic cultures (n = 3 of 7) revealed a weak TRPM2-specific signal, we measured whole cell currents in cultured astrocytes using 1 mM ADPR. All tested cells did not respond to infusion of ADPR (n = 8). Hence, a possible expression of TRPM2 in astrocytic cultures most likely results from contamination with microglial cells.
Expression pattern of TRPM2 in mouse tissues and brain. We finally verified the expression of TRPM2 in native tissues, particularly in the brain. For Northern blot hybridization, fragments of the cloned TRPM2 cDNA were isolated, radiolabeled, and used as a probe to hybridize membranes carrying total RNA isolated from different mouse tissues. A distinct band of about 5.5 kb was visible only in the autoradiogram in the lane of mouse brain (Fig. 6A). No signals were detected in liver, kidney, lung, small intestine, or skeletal muscle, suggesting a preferential expression of TRPM2 in the brain. In situ hybridizations were generated from adult mouse brain in sense or antisense direction to control the specificity of the signals. Autoradiograms from brain slices showed an expression of TRPM2 in the cerebral cortex, thalamus, hippocampus, midbrain, and cerebellum (Fig. 6B). In the olfactory bulb, septum, and caudate putamen, no or very low levels of TRPM2 mRNA were visible.
In summary, TRPM2 is expressed in cultured microglial cells, but not in astrocytes and cerebellar granule cells, and can be detected in native brain tissue. The characteristics of ADPR- and H2O2-induced responses provide further evidence for a functional role of this channel in microglia.
This study provides evidence for the expression and a functional role of TRPM2 in microglial cells, the host macrophages of the brain. We could show a close similarity between properties of ADPR-gated channels and H2O2-induced currents in microglia and in a cell line with heterologously expressed TRPM2. Micromolar concentrations of H2O2 induced a long-lasting Ca2+ influx in microglia that was strongly enhanced in activated cells.
Expression of TRPM2, originally named TRPC7, was initially found in human brain, particularly in cerebral cortex, occipital pole, frontal lobe, amygdala, caudate nucleus, and putamen (22). A widespread distribution of TRPM2, i.e., in lung, spleen, eye, and brain, was described later (6). Moreover, within the immune system, TRPM2 has been detected in abundance in cells of the monocytic lineage, including various cultured macrophage cell lines, peripheral blood monocytes, and neutrophils (27). Our in situ hybridizations from mouse brain showed a wide expression of TRPM2 with increased signal intensity in the hippocampus, cerebral cortex, thalamus, and midbrain. Microglia comprise up to 20% of the total glial cell population in the brain. Densely populated areas include the hippocampus, olfactory telencephalon, basal ganglia, and substantia nigra; average microglia densities were found in the cerebral cortex, thalamus, and hypothalamus (16). Thus the distribution of TRPM2 is compatible with an expression in microglia. Further evidence for a preferential localization in this cell type is provided by our RT-PCRs from cultured rat glial and neuronal cells showing TRPM2-specific signals only in microglia but not in astrocytes and cerebellar granule cells.
Intracellular ADPR is a specific stimulus for TRPM2 (26, 30). By using whole cell and single-channel patch-clamp techniques, we determined properties of ADPR-gated channels in microglia that indeed resemble those of TRPM2. The single-channel conductance was 65 pS in microglia vs. 58–76 pS for heterologously expressed TRPM2 (26, 30). The relative permeability for Ca2+ over monovalent cations (PCa/PCs) has been reported to be 0.67 (30). We found values for PCa/PNa of 0.71 and a so far unreported permeability for Mg2+ (PMg/PNa) of 0.47. A similar value for PMg/PNa was calculated from measurements in TRPM2-transfected HEK-293 cells (Kraft R, unpublished observation). Likewise, sensitivity to La3+ has not been reported so far. We show that this often used Ca2+ channel inhibitor exerts only negligible influence on ADPR-gated currents in microglia and TRPM2-transfected HEK-293 cells at concentrations of 0.1 to 1 mM. Besides ADPR, the nucleotide NAD+ has been reported to activate TRPM2 (6, 30). However, we and others (26, 32) could not confirm the finding of a direct NAD+-mediated opening of TRPM2 channels (6, 30). The reason for this contradiction is presently unclear; however, NAD+ can be converted to ADPR by enzymatic activity of NAD+ glycohydrolases, such as the membrane-bound enzyme CD38 (10, 20).
Activation of TRPM2 by H2O2 has been reported from two approaches. One group has shown H2O2-induced opening of TRPM2 channels in the whole cell configuration (32); another group detected channel activation only in perforated-patch experiments (6). The latter finding suggests the involvement of a diffusable second messenger released by H2O2, whereas the other observation does not exclude a direct oxidant-induced TRPM2 gating. In both microglia and TRPM2-transfected HEK-293 cells, we observed current responses in the whole cell configuration, which also raises the possibility of a direct effect of H2O2 on TRPM2. Experiments on COOH-terminal truncated TRPM2 channels, lacking the ADPR pyrophosphatase domain, revealed a loss of ADPR-mediated gating but a retention of H2O2-induced activation (32). This observation strongly supports the hypothesis of independent gating mechanisms by H2O2 and ADPR.
To our knowledge, this is the first report describing H2O2-induced calcium signals and ion currents in microglia. It is known that H2O2 interacts with a number of ion transport mechanisms, including ion channels in the plasma membrane and the endoplasmatic reticulum (13). Some of these effects were reported on Ca2+-permeable channels, such as the suppression of voltage-gated (L-type) Ca2+ channels and the activation of ryanodine-sensitive Ca2+-release channels (13). In microglial cells, we found a H2 O2-induced Ca2+-influx via the plasma membrane and, under certain conditions (in activated cells and at high concentrations of H2O2), an additional Ca2+ release from internal stores. Whereas the mechanism of the latter Ca2+ transport was not subjected to further investigation, we could show the involvement of a La3+-insensitive cation channel in the Ca2+ entry pathway in microglia. Application of 100 μM H2O2 induced this Ca2+ influx within 10 min. A major source of H2O2 in brain tissue are microglial cells themselves. In a study using 3 × 105 microglial cells per milliliter, nanomolar concentrations of H2O2 were measured after 30 min of stimulation with PMA (2). In vivo, however, microglial cells are not arranged in a monolayer with a large cell-free extracellular space, but in a three-dimensional tissue with densely packed cells. We can therefore expect an increase in H2O2 concentration up to micromolar levels and suggest a physiological role for H2O2 in the calcium signaling and cation transport in microglial cells. Millimolar H2O2 concentrations (5 mM) dramatically shortened the latency of Ca2+ responses (≤2 min) in our experiments. A similar dose-dependent kinetics has been reported for heterologously expressed TRPM2 (6). In some other cell types, i.e., the insulin-secreting cell line CRI-G1 (8), pancreatic β-cells (15), and striatal neurons (31), H2O2 (at millimolar concentrations) has been shown to induce Ca2+ influx or activation of nonselective cation channels. Moreover, for CRI-G1 cells (11) and another insulinoma cell line RIN-5F (6), the expression of TRPM2 has been verified, suggesting identity of H2O2-activated channels with TRPM2.
Interestingly, the H2O2-induced Ca2+ responses in microglia differed significantly from unstimulated to LPS-treated cells, whereas the ADPR-induced conductances were quite similar. From the increased H2O2-mediated signaling, we therefore suggest an increased sensitivity of TRPM2 for H2O2 in activated microglia. One possible explanation for this effect can be a modification of TRPM2 or TRPM2-interacting proteins during activation of cells due to a change in the phosphorylation state. Recently, it has been reported that treatment of microglia with LPS leads to a prolonged decrease of conventional protein kinase C activity (23).
Furthermore, TRPM2-mediated currents might be amplified in activated microglial cells by two mechanisms. First, only activated microglial cells express outward rectifying currents (24, 25) that counteract depolarization by a nonselective conductance. The more negative membrane potential will enhance the current through TRPM2 with its reversal potential at about 0 mV. Second, activated microglia have a higher resting Ca2+ level. This will enhance TRPM2-mediated currents because it is a Ca2+-dependent channel showing a half-maximal activation at 340 nM (19).
The chronic elevation of basal [Ca2+]i in activated microglia as described earlier (9) could be confirmed in our study. The elevated basal [Ca2+]i was furthermore linked to the reduced ability of activated microglia to respond to external stimuli, such as ATP, UTP, and C5a (9). Thus H2O2-induced Ca2+ signaling is conversely regulated with respect to the receptor-mediated effects. It has been speculated that the downregulation of receptor-mediated Ca2+ signaling represents an intrinsic mechanism of activated cells to shut off sensitivity toward external influences (9). Consistent with this idea, the upregulation of H2O2-induced Ca2+ signaling could stabilize the activated state of microglia because H2O2 is continuously produced by these cells themselves. An enhanced activity of TRPM2 due to increased H2O2-sensitivity might elevate basal [Ca2+]i, which is an essential prerequisite for important functions of activated microglia, namely, the release of cytokines and NO (9).
We thank Inge Reinsch for technical assistance.
This work was supported by the Deutsche Forschungsgemeinschaft and Fonds der Chemischen Industrie.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2004 the American Physiological Society