|Home | About | Journals | Submit | Contact Us | Français|
Acidosis is a common feature of the human brain during ischemic stroke and is known to cause neuronal injury. However, the mechanism underlying acidosis-mediated injury of the human brain remains elusive. We show that a decrease in the extracellular pH evoked inward currents characteristic of acid-sensing ion channels (ASICs) and increased intracellular Ca2+ in cultured human cortical neurons. Acid-sensing ion channels in human cortical neurons show electrophysiological and pharmacological properties distinct from those in neurons of the rodent brain. Reverse transcriptase-PCR and western blot detected a high level of the ASIC1a subunit with little or no expression of other ASIC subunits. Treatment of human cortical neurons with acidic solution induced substantial cell injury, which was attenuated by the ASIC1a blockade. Thus, functional homomeric ASIC1a channels are predominantly expressed in neurons from the human brain. Activation of these channels has an important role in acidosis-mediated injury of human brain neurons.
Acid-sensing ion channels (ASICs) are ligand-gated channels activated by protons (Waldmann et al, 1997). They belong to the amiloride-sensitive Na+ channel/Degenerin superfamily (Bianchi and Driscoll, 2002). In rodents, they are enriched in neurons of both peripheral and central nervous systems (Krishtal, 2003; Wemmie et al, 2006). So far, seven subunits encoded by four genes have been identified (Wemmie et al, 2006). All subunits except ASIC2b and ASIC4 can form functional homomultimeric channels, with distinct electrophysiological and pharmacological properties. Recent animal studies have shown that activation of Ca2+-permeable ASIC1a channels in central nervous system neurons has an important role not only in synaptic plasticity, learning, and memory (Wemmie et al, 2002) but also in acidosis-mediated intracellular Ca2+ accumulation and neuronal injury (Yermolaieva et al, 2004; Xiong et al, 2004; Gao et al, 2005; Herrera et al, 2008).
Stroke is a leading cause of mortality and morbidity and the most common reason for long-term disability. Unfortunately, there remains no effective treatment for stroke patients other than the use of thrombolytics, which has limited success because of a short therapeutic time window and the potential of intracranial hemorrhage (Caplan, 2004). Long-recognized intracellular Ca2+ accumulation, particularly through glutamate receptor activation, has an important role in ischemic brain injury (Choi, 1988). However, recent clinical trials using the glutamate antagonists have failed to show protection in ischemic brain injury (Ikonomidou and Turski, 2002). Although multiple factors may account for the failure of the trials, one could speculate that glutamate-independent Ca2+ loading pathway(s) also contribute to the Ca2+ toxicity in ischemia. Indeed, recent studies have clearly shown that activation of Ca2+-permeable ASIC1a channels has an important role in glutamate receptor-independent, acidosis-mediated, neuronal injury in ischemia (Yermolaieva et al, 2004; Xiong et al, 2004; Gao et al, 2005). It is noteworthy that the ASIC1a blockade has a >5h therapeutic time window in rodent models of focal ischemia (Pignataro et al, 2007).
Although the presence of ASICs, their electrophysiological/pharmacological properties, and the potential physiologic/pathologic roles have been studied in animal cells, no studies of ASICs have been conducted on native human brain neurons. The distribution, subunit combination, electrophysiological/ pharmacological properties, and thus the function of an ion channel may vary dramatically from species to species. For example, the ortholog of rodent ASIC2b mRNA has not been identified in humans, whereas ASIC3 shows a widespread distribution in human tissues, in contrast to its restricted localization in rodents (Babinski et al, 1999, 2000). In addition, animal cells or experimental models do not adequately mimic human pathophysiology (Hackam, 2007; Savitz, 2007). For these reasons, the findings in animal cells cannot be simply translated into humans. Studies dedicated to human brain neurons are thus a critical step for establishing ASICs as novel therapeutic targets for human stroke.
Using the whole-cell patch-clamp technique, biochemical/molecular biologic analysis, fluorescent Ca2+-imaging, and cell injury assays, we studied the electrophysiological/pharmacological properties of ASICs in neurons isolated from human cortical tissues. The role of ASICs in acidosis-mediated injury of human brain neurons was examined.
Human brain tissues from a total of 12 patients, aged 23 to 71 years, were used for the studies. The protocol for obtaining and using human brain tissues was approved by the Institutional Review Board Committee of the Legacy Clinical Research Center (Protocol number: FWA00001280). Cortical tissue samples were obtained, with consent, from patients undergoing craniotomies for the removal of brain tumor. Small tissues were removed for access to and removal of brain tumors. These cortical tissues were normally considered surgical waste that otherwise would have been discarded (Frenkel et al, 1998).
Cortical tissues were enzyme treated and isolated according to the method described previously (Brewer et al, 2001). Briefly, the tissue obtained in the operating room was rapidly transferred to a laminar flow sterile hood in the culture laboratory in a sterile 50mL polystyrene tube containing 25mL ice-cold sterile medium (Hibernate A with 2% B27 medium supplement). The meninges and white matter were removed from the tissue with a mini scalpel and forceps in a 35mm dish in a 2mL transporting medium. The tissue was cut in 0.5mm slices, digested with papain (3.5mg/mL), and dissociated into a single cell suspension as described previously (Brewer et al, 2001). The cell suspension was enriched for neurons by centrifugation. Cells were counted and plated in poly--ornithine-coated culture dishes at a density of 1 × 106 cells per 35mm diameter dish. Cells were cultured in Neurobasal A medium (Life Technologies, Carlsbad, CA, USA) with 2% B27, 5ng/mL fibroblast growth factor-2 (Life Technologies), and 0.5mmol/L glutamine in a humidified atmosphere with 5% CO2. In general, neurons were used for experiments 3 to 4 days after culture. Both transporting and culture media contained 1% penicillin/streptomycin (Invitrogen, Carlsbad, CA, USA).
Acid-sensing ion channel currents were recorded using the conventional whole-cell patch-clamp and fast-perfusion technique as described previously (Xiong et al, 2004). Data were acquired using an AXOPATCH 200B amplifier with pCLAMP 8.1 software (Axon Instruments, Foster City, CA, USA). Unless otherwise specified, cells were voltage clamped at −60mV. Patch pipettes were pulled from borosilicate glass (1.5mm diameter; WPI, Sarasota, FL, USA) on a two-stage puller (PP83, Narishige, Tokyo, Japan). Pipettes had a resistance of 2 to 4MΩ when filled with the intracellular solution (see below). Membrane capacitance was recorded for each neuron as a measure of cell size. For rapid changes of extracellular solutions, a multibarrel perfusion system (SF-77, Warner Instruments, Hamden, CT, USA) was used.
Standard extracellular solution (ECF) contained (in mmol/L): 140 NaCl, 5.4 KCl, 2 CaCl2, 1 MgCl2, 20 HEPES, 10 glucose. The Na+-, K+-, and Mg2+-free high Ca2+ solutions contained (in mmol/L): 140 choline chloride, 10 CaCl2, 25 HEPES, 10 glucose. pH was adjusted with NaOH/HCl, and osmolarity was adjusted to 320 to 335mOsm. For solutions with pH≤6.0, HEPES was replaced by 2-(4-morpholine)-ethane sulfonic acid (MES) for more reliable pH buffering (Bassler et al, 2001; Chu et al, 2002). Intracellular solution contained (in mmol/L): 140 cesium fluoride (CsF), 2 TEACl, 5 EGTA, 10 HEPES, 1 CaCl2, 4 MgCl2; pH 7.3 adjusted with CsOH, 290 to 300mOsm.
Solutions of progressively decreasing pH were applied at intervals of 2mins to allow for complete recovery of ASIC channels from desensitization. During each experiment, a voltage step of −10mV from the holding potential was applied periodically to monitor cell capacitance and access resistance. To ensure high quality of voltage clamp, only recordings with an access resistance of <10MΩ and a leak current <100pA at −60mV were included for data analysis. ZnCl2 and amiloride were purchased from Sigma (St Louis, MO, USA). Spider venom containing Psalmotoxin 1 (PcTX1) was purchased from Spider Pharm (Yarnell, AZ, USA). Synthesized PcTX1 was purchased from Peptide International (Louisville, KT, USA).
Fura-2 fluorescent Ca2+ imaging was performed as described previously (Chu et al, 2004; Xiong et al, 2004). Human neurons grown on poly--ornithine-coated 25 × 25mm2 glass coverslips were washed three times with ECF and incubated with 5μmol/L fura-2-acetoxymethyl ester for ~30mins at room temperature. Neurons were then washed three times and incubated in normal ECF for 30mins. Coverslips with fura-2-loaded neurons were transferred to a perfusion chamber on an inverted microscope (Nikon TE300, Nikon, Melville, NY, USA). Cells were illuminated using a xenon lamp (75W) and observed with a × 40 ultraviolet fluor oil-immersion objective lens. Video images were obtained using a cooled CCD camera (Sensys KAF 1401, Photometrics, Tucson, AZ, USA). Digitized images were acquired, stored, and analyzed in a personal computer controlled by Axon Imaging Workbench software (AIW2.1, Axon Instruments). The shutter and filter wheel (Lambda 10-2, Sutter Instrument, Novato, CA, USA) were also controlled by AIW to allow timed illumination of cells at either 340 or 380nm excitation wavelengths. Fura-2 fluorescence was detected at an emission wavelength of 510nm. Ratio images of 340/380 were analyzed by averaging pixel ratio values in circumscribed regions of cells within the field of view. The values were exported from AIW to SigmaPlot (Systat Software, Inc., San Jose, CA, USA) for further analysis and plotting.
Cultured human cortical neurons grown on etched coverslips (Bellco Biotechnology, Vineland, NJ, USA; Cat no. 1916-90025) were randomly divided into different treatment groups. Desired cells were captured as phase-contrast images, and their alphanumeric coordinates on the coverslip were noted. After a 2-h acid treatment, cells were washed three times with Neurobasal medium and incubated at 37°C. Staining of the alive/dead cells was performed at 24h after the 2-h acid treatment as described previously (Xiong et al, 2004; Wang et al, 2006). Briefly, cells were incubated in ECF containing Hoechst (which stains all cell nuclei) and propidium iodide (PI, which stains dead cell nuclei) for 10mins. Neuronal injury was then measured using the Image J software (Wayne Rasband at NIH, Bethesda, MD, USA) and presented as a ratio of dead to live. To exclude the potential contribution made by other Ca2+-permeable channels to acid toxicity, blockers for glutamate receptors (10μmol/L MK-801 ((5S,10R)-(−)-5-Methyl-10,11-dihydro-5H-dibenzo[a,d]cyclohepten-5,10-imine hydrogen maleate), 20μmol/L CNQX (6-cyano-7-nitroquinoxaline-2,3-dione) and voltage-gated Ca2+ channels (e.g., 5μmol/L nimodipine) were added in all groups.
Total RNAs were isolated from a human brain sample with Trizol reagent (Invitrogen) according to the manufacturer's protocol. Equal amounts of total RNA were reverse transcribed and PCR amplified with Superscript II (Invitrogen) using specific primers for GAPDH (glyceraldehyde 3-phosphate dehydrogenase) and ASIC1a, 2a, 2b, 3, and 4. The primer sequences were as follows; GAPDH forward (5′-ATGCTGGTGCTGAGTATGTCGTG-3′), GAPDH reverse (5′-TTACTCCTTGGAGGCCATGTAGG-3′), ASIC1a forward (5′-CTGGAGATACTGCAGGACAAAG-3′), ASIC1a, 2a and 2b reverse (5′-GGGGCATCCCCTGGCATGTG-3′), ASIC2a forward (5′-GGCCCTGCTGGATGTCAACCTG-3′), ASIC2b forward (5′-AGTGGTTCCGCAAGCTGGCGG-3′), ASIC3 forward (5′-GAGGTGGGGATCCGAGTGCAG-3′), ASIC3 reverse (5′-CTGCTGGGGGCTGCACACTGG-3′), ASIC4 forward (5′-CAGATGCCGATCGAGATTGTG-3′), ASIC4 reverse (5′-TTGCAGAGGGTGACAGCCGG-3′). The reverse transcriptase (RT)-PCR products were electrophoresed on 1.5% gel and visualized. Similar results were obtained from the other three independent samples.
Human brain samples were homogenized, and then lysed in lysis buffer (50mmol/L Tris-HCl, pH 7.5, 100mmol/L NaCl, 1% Triton X-100, and protease inhibitor). After centrifugation at 12,000g at 4°C for 30mins, the supernatants were collected. Protein concentration was estimated using the Bio-Rad protein assay (Bio-Rad, Hercules, CA, USA). Thereafter, 60mg of proteins was mixed with Laemmli sample buffer and boiled at 95°C for 10mins. The samples were resolved by 10% SDS-PAGE, followed by electrotransfer to polyvinylidene difluoride membranes. For visualization, blots were probed with antibodies to ASIC1 (rabbit anti-mouse/human, 1:1,000; Sigma), ASIC2a (1:1,000; Alpha Diagnostics, San Antonio, TX, USA), or actin (1:2,000; Abcam, Cambridge, MA, USA), and detected using horseradish peroxidase-conjugated secondary antibody (1:1,000; Cell Signaling, Danvers, MA, USA) and an enhanced luminescence kit (Amersham Pharmacia Biotech, Piscataway, NJ, USA). For immunofluorescent staining of the ASIC1a subunit, ASIC1a antibody at 1:100 dilution and goat anti-rabbit secondary antibody conjugated with Cy3 (Jackson Lab, Bar Harbor, MA, USA; Cat no. 111-165-144) at 1:750 dilution were used.
The pH50 values for ASICs were obtained by fitting with Hill equations as previously described (Wang et al, 2006): I=a/(1+(pH50/pH)n), where a is the current amplitude, pH50 the pH at which a half-maximal current is activated, and n the Hill coefficient. All data are presented as mean±s.e.m. Student's t-test was used to determine the statistical difference. The criterion of significance was set at P<0.05.
Acid-activated currents were studied in human cortical neurons 2 to 4 days after culture using whole-cell patch-clamp and fast-perfusion techniques as described previously (Xiong et al, 2004). Neurons were easily identified by their pyramidal or round cell body with some processes and the presence of tetrodotoxin (TTX)-sensitive voltage-gated Na+ current (Figures 1A and 1B). In addition, human cortical neurons express glutamatergic and GABAergic neurotransmitter receptors as evidenced by positive responses to bath perfusion of NMDA (N-methyl--aspartate) and GABA (γ-aminobutyric acid) (Figures 1C and 1D).
In the majority of human cortical neurons, a rapid change of pH in the extracellular solution from 7.4 to below 7.0 activated a transient inward current with a small steady-state component (Figure 1E). The amplitude of acid-activated inward currents increases with decreasing pH values and a maximal amplitude was achieved at a pH of ~5.0 (Figures 1E and 1F). Detailed dose–response analysis yielded a pH50 of 6.60±0.02 and a Hill coefficient of 1.79±0.04 (n=10). At a pH of 6.0, an average current amplitude of −1,144.08±104.11pA or a current density of −30.57±3.01pA/pF was recorded (n=56).
The current–voltage (I–V) relationship of the acid-activated currents was constructed by plotting the peak amplitude of the currents activated at pH 6.0 against the membrane potentials. As shown in Figures 1G and 1H, the acid-activated currents in human cortical neurons has a near linear I–V relationship with a reversal potential of ~+60mV (n=6). This reversal potential is close to the anticipated Na+ equilibrium potential. Taken together, these electrophysiological studies provide strong evidence supporting the expression of functional ASICs in human cortical neurons.
The effect of amiloride, a commonly used nonspecific inhibitor of ASICs in animal cells (Waldmann et al, 1997; Xiong et al, 2004), was tested on acid-activated currents in human cortical neurons. After recording stable inward currents activated at pH 6.0, 100μmol/L amiloride was added to both pH 7.4 and pH 6.0 solutions. As illustrated in Figure 2A, the acid-activated currents in human cortical neurons were potently and reversibly inhibited by amiloride, providing further evidence supporting the expression of functional ASICs in these neurons. In the presence of 100μmol/L amiloride, the amplitude of ASIC currents was decreased by ~90% (from −1,209.46±136.47pA to −59.23±11.25pA, n=9, P<0.01) (Figure 2B).
Next, we tested the effect of PcTX1, a specific inhibitor for Ca2+-permeable homomeric ASIC1a channels (Escoubas et al, 2000; Xiong et al, 2004). As shown in Figure 2C, the addition of 10nmol/L synthetic PcTX1 (Peptide International, Louisville, KY, USA), a near saturating concentration for inhibiting the homomeric ASIC1a channels expressed in CHO cells (not shown), dramatically reduced the amplitude of the acid-activated currents in human cortical neurons to 28.68%±4.36% of control (n=10, P<0.01) (Figure 2D). The reduction of the ASIC current by PcTX1 in human cortical neurons is more pronounced than the current in rodent brain neurons where only ~50% of the current was inhibited (Baron et al, 2002; Xiong et al, 2004). This finding may suggest that the Ca2+-permeable homomeric ASIC1a channels are the predominant configuration of ASICs responsible for acid-activated currents in human cortical neurons.
To gain more information on the subunit composition of ASICs in human cortical neurons, we tested the effect of Zn2+ on the ASIC currents. Previous studies have shown that, depending on the subunit combination, ASIC currents can show distinct responses to the extracellular Zn2+. ASIC1a-containing channels, e.g., show high-affinity Zn2+ inhibition, with an IC50 of ~15nmol/L (Chu et al, 2004). Thus, chelating the ambient Zn2+ (e.g., by TPEN (N,N,N',N'-tetrakis(2-pyridylmethyl)ethylenediamine) often produces a dramatic potentiation of the activity of ASIC1a-containing channels (Chu et al, 2004). In contrast, ASIC2a-containing channels show low-affinity Zn2+ potentiation; the currents are significantly increased in the presence of high concentrations of Zn2+ (e.g., 100 to 300μmol/L) (Baron et al, 2001). As shown in Figure 2E, addition of the high-affinity zinc chelator TPEN significantly potentiated the ASIC current in human cortical neurons, indicating the presence of ASIC1a-containing channels. In the presence of 10μmol/L TPEN, the peak amplitude of the ASIC current was increased from −894.69±290.09pA to 1,036.65±340.90pA (n=9, P<0.01) (Figures 2E and 2F). Different from the rodent brain where the ASIC currents in a large percentage of neurons are potentiated by the addition of high concentrations of Zn2+ (Baron et al, 2002), application of 150μmol/L Zn2+ did not increase the amplitude of the ASIC currents in human cortical neurons (n=17, Figure 3). The lack of potentiation of the ASIC current in human cortical neurons by high concentration of Zn2+ suggests the absence of functional ASIC2a-containing channels (Baron et al, 2001).
Similar to other ligand-gated channels, e.g., NMDA and GABA receptor-gated channels, the responses of ASICs are transient in nature, suggesting that ASICs desensitize or inactivate in the continuous presence of acidosis. This property is important as it determines how long ASIC-induced responses (membrane depolarization, Ca2+ entry, etc.) may last in the continuous presence of acidosis. For this reason, we performed a detailed analysis of the desensitization properties of the ASIC currents in human cortical neurons. We first analyzed the decay time constant of the currents activated at different pH values. The ASIC currents were activated by fast perfusion of neurons with acidic solutions (pH 7.0, 6.5, 6.0, 5.5, 5.0) for 6secs. The decay time constant (τ) was derived by fitting with single exponential function: y=Ae−t/τ, where y is the relative amplitude, t the time, and τ the time constant (method). As shown in Figures 4A and 4B, the time constant for desensitization of ASICs is strongly dependent on the test pH values; the rate of desensitization increased with decreasing pH values. The time constants of desensitization were: 1,164.9±39.9, 1,057.3±31.6, 775.1±33.9, 464.2±29.6, and 292.5±35.0msecs at pH 6.5, 6.0, 5.5, 5.0, and 4.5, respectively.
We next analyzed the rate of recovery of ASICs from desensitization in human cortical neurons using pairs of acid pulses as described previously (Wang et al, 2006). Neurons were exposed to pH 6.0 for 15secs to achieve a near complete desensitization of the ASIC currents, followed by perfusion with pH 7.4 solution for various durations (1, 3, 5, 15, 25, 50, and 100secs) before the second acid exposure. The peak amplitude of the second set of currents was then normalized to the first set and plotted against the time interval between the end of the first and the beginning of the second acid exposures. A time constant for the recovery of ASICs from desensitization in human cortical neurons (τ) of 0.90±0.23secs (n=6) was recorded (Figures 4C and 4D). This time constant for recovery from desensitization is shorter than for the ASICs in mouse cortical neurons (Wang et al, 2006). Thus, ASICs in human cortical neurons can recover faster than the ASICs in mouse cortical neurons.
In addition to desensitization, the steady-state inactivation of ASICs in human cortical neurons was studied. Neurons were first exposed to extracellular solutions at different conditioning pH values ranging from 7.6 to 6.7 for ~4mins before the currents were activated by decreasing the pH to 6.0. The amplitude of ASIC currents recorded with different conditioning pH values was normalized to the one recorded with a conditioning pH of 7.6 in which no apparent inactivation occurs. The normalized amplitude was then plotted against the conditioning pH values to generate the steady-state inactivation curve (Wang et al, 2006). We obtained a half-maximal pH value (pH50) for steady-state inactivation of ASICs in human cortical neurons at 7.03±0.02 (n=7, Hill coefficient 1.92±0.04, Figures 4E and 4F). This value is lower than the pH50 for ASIC inactivation in mouse cortical neurons where pH 7.2 was reported (Wang et al, 2006). Therefore, ASICs in human cortical neurons are more resistant to acid inactivation.
Intracellular Ca2+ overload is essential for neuronal injury associated with ischemia and other neurologic disorders (Choi, 1994, 1995). To know whether ASICs represent a new Ca2+ loading pathway for human brain neurons, we determined whether ASICs in these neurons are Ca2+ permeable. Using an ion-substitution protocol with Na+- and K+-free extracellular solutions containing only 10mmol/L Ca2+ as the charge carrier (osmolarity adjusted with nonpermeable cation NMDG) (Xiong et al, 2004), detectable inward currents were recorded in the majority of human cortical neurons in response to the pH decrease from 7.4 to 6.0 (holding potential was −60mV, n=5, Figure 5A), suggesting that ASICs in these neurons are Ca2+ permeable.
To provide further evidence that activation of ASICs triggers an increase of [Ca2+]i in human cortical neurons, fluorescent Ca2+ imaging was performed to determine whether a decrease in pH can induce an increase of [Ca2+]i. As shown in Figure 5B, reduction of pHo from 7.4 to 6.0, in the presence of blockers of other major Ca2+ entry pathways (e.g., 10μmol/L MK-801 and 20μmol/L CNQX for glutamate receptors; 5μmol/L nimodipine for voltage-gated Ca2+ channels), triggered a clear increase of [Ca2+]i, as indicated by an increased intensity of 340/380nm ratio image (Figure 5B, n=6). This acid-induced increase of [Ca2+]i was inhibited by amiloride (Figure 5B). Taken together, these data strongly suggest that ASICs in human cortical neurons are Ca2+ permeable and that activation of these channels triggers intracellular Ca2+ loading.
To determine whether ASIC activation is involved in acidosis-mediated injury of human cortical neurons, and whether ASIC blockers can serve as neuroprotective agents for human stroke, we studied the effect of ASIC1a blockade on acidosis-induced injury of human cortical neurons. Cell injury was studied using neurons grown on photo-etched coverslips (25mm in diameter, Bellco Biotechnology, Cat no. 1916-90025) wherein individual neurons can be tracked at different time points according to their alphanumeric coordinates. After 3 days in culture, cells were washed three times with normal extracellular solution and randomly divided into the following treatment groups: (1) control: treated with pH 7.4 extracellular solution; (2) acidosis: treated with pH 6.0 solution; (3) acidosis with amiloride: treated with pH 6.0 solution in the presence of amiloride; and (4) acidosis with PcTX1: treated with pH 6.0 solution in the presence of PcTX1. Cells were treated with different solutions for 2h, followed by reverting to normal cell culture medium and placed in the incubator. To eliminate possible effects of acidosis on the functions of glutamate receptors and voltage-gated Ca2+ channels, which might complicate data interpretation, MK-801 (10μmol/L), CNQX (20μmol/L), and nimodipine (5μmol/L) were added in the extracellular solutions of all groups. Live and dead cells were counted 24h after a 2-h treatment with different solutions by changes in cell morphology and by PI labeling of nuclei of dead cells (Xiong et al, 2004). As shown in Figure 6A, treatment of neurons with acidic solution (pH 6.0) triggered clear neuronal injury as evidenced by cell body swelling, loss of cell process, and positive PI staining. In contrast, neurons treated with 7.4 solutions showed near normal morphology and negative PI staining. Figure 6B is a summary of the acid injury experiment. Neurons treated with acidic solution (pH 6.0) triggered 73.42%±15.54% of cell death (n=31 cells from five separate cultures). In contrast, neuron treatment with normal solution (pH 7.4) had only 10.96%±4.9% of cell death (n=65 cells from five separate cultures). The acid-induced neuronal cell death in amiloride and PcTX1-treated cultures was 30.11%±17.85% and 17.78%±17.78% (n=79 to 85 cells from five separate cultures, P<0.05), respectively. Taken together, these data suggest that acidosis can injure human cortical neurons through activation of ASIC1a channels and that ASIC1a blockade can attenuate such injury.
To provide biochemical/molecular biologic evidence that ASICs are expressed in human brain neurons, RT-PCR, western blot, and immunofluorescent staining were used. Reverse transcriptase-PCR detected rich expression of ASIC1 mRNA with weak expression of ASIC2 mRNA (Figure 7A). No clear expression of ASIC2b, ASIC3, or ASIC4 mRNA was detected. Consistent with RT-PCR data, western blot showed a strong band for ASIC1 with a much weaker band for ASIC2, although variation was noticed among different human samples (Figure 7B). Immunofluorescent staining also showed positive staining for the ASIC1a subunit in human cortical neurons (Figure 7C). Thus, biochemical and molecular biologic data support the electrophysiological and pharmacological findings that ASIC1a is a predominant ASIC subunit expressed in human brain neurons. Different from the findings by Babinski et al, our RT-PCR analysis did not show a clear expression of ASIC3 in human brain tissues. However, this finding is consistent with the study by Ishibashi and Marumo (1998)who reported an exclusive expression of hASIC3 in testis.
Stroke is a leading cause of death and long-term disabilities. Unfortunately, there is no effective treatment for stroke patients other than the use of thrombolysis (e.g., with tissue plasminogen activator), which has a limited time window and a potential side effect of intracranial hemorrhage (Caplan, 2004; Wang et al, 2004). The recent failure of clinical trials using glutamate antagonists and a free radical scavenger as neuroprotective agents further emphasizes the need for new therapeutic targets and neuroprotective strategies (Ikonomidou and Turski, 2002; Savitz and Fisher, 2007). The findings in recent animal studies showing that activation of Ca2+-permeable ASICs has an important role in acidosis-mediated glutamate receptor-independent neuronal injury disclosed a potential novel target for stroke intervention (Yermolaieva et al, 2004; Xiong et al, 2004; Benveniste and Dingledine, 2005). In rodent models of focal ischemia, ASIC1a blockade provided ~60% reduction of the infarct volume (Xiong et al, 2004). Moreover, ASIC1a blockade showed a therapeutic time window of >5h (Pignataro et al, 2007), far beyond that of the ~1h time window for glutamate receptor antagonists (Hoyte et al, 2004). Thus, targeting ASIC1a channels may prove to be an effective strategy for the intervention of human stroke.
Although the electrophysiological/pharmacological properties of ASICs and their role in acidosis-mediated neuronal injury have been studied in animal preparations, no study has been performed on native neurons from the human brain. Natively expressed channels are species specific because of the potential differences in their subunit repertoire and corresponding variants, in addition to spatiotemporal differences in the regulation of developmental and anatomic patterns of expression. For these reasons, the findings in animal cells cannot be easily translated into the human species. Studies by Babinski et al (2000), e.g., have shown that the ortholog of rodent ASIC2b mRNA is not present in human tissues, whereas ASIC3 shows a widespread distribution in human tissues in contrast to a restricted localization to sensory ganglia in rodents (Babinski et al, 1999). In addition to the potential differences in subunit combinations, animal cells or experimental models may not adequately mimic human pathophysiology (Hackam, 2007; Savitz and Fisher, 2007). Test animals are often young, rarely have comorbidities, and are not exposed to the range of competing and interacting interventions that humans often receive (Hackam, 2007). Therefore, additional preclinical studies using human brain neurons are essential in moving forward with ASIC antagonists as novel and potentially highly translatable therapeutics for stroke. Moving directly from rodent studies to clinical trials may be a substantial factor in misinterpreting efficacy, dosing, and time widow considerations critical in human clinical trial design. These translational issues have been brought into clear focus by the failure of the Stroke-Acute Ischemic NXY Treatment (SAINT) II trial, a trial based on preclinical data fulfilling the Stroke Therapy Academic Industry Roundtable (STAIR) criteria (Savitz and Fisher, 2007).
We showed that, in the majority of human cortical neurons, decreasing pHo from 7.4 to below 7.0 evoked transient, amiloride-sensitive, inward current characteristics of ASICs. The pH50 for ASIC currents in human cortical neurons is ~6.6, which is substantially higher than that in mouse cortical neurons (~6.0) (Chu et al, 2004; Xiong et al, 2004; Wang et al, 2006). This finding suggests that ASICs in human cortical neurons have higher sensitivity to acid stimulation than do ASICs in rodent brain. The increased sensitivity to acid by ASICs in human cortical neurons suggests that ASICs in the human brain are more likely to be activated in physiologic and/or pathologic conditions in which only moderate pH decreases occur.
This difference in pH sensitivity may be explained by the finding that homomeric ASIC1a channels, which have higher sensitivity to H+, are the predominant ASIC configuration in human brain neurons. In contrast, in the rodent brain neurons, a combination of homomeric ASIC1a, heteromeric ASIC1a/ASIC2a, and homomeric ASIC2a channels coexist (Askwith et al, 2004; Chu et al, 2004; Xiong et al, 2004). The reduced sensitivity of ASIC2a-containing channels to H+ contributes to the lower pH50 value for ASICs in rodent brain neurons.
The desensitization property of ASICs is another important parameter for these channels as it determines how long ASIC-induced responses (e.g., membrane depolarization and Ca2+ entry) can last in conditions in which prolonged acidosis occurs (e.g., during brain ischemia or inflammation) and how fast they can recover after desensitization. Similar to ASICs in rodent brain neurons, ASICs in human cortical neurons show a pH-dependent increase in desensitization. However, compared with the ASICs in mouse cortical neurons, ASICs in human cortical neurons show a faster recovery from desensitization. A time constant for recovery of ASICs from desensitization in human cortical neurons is ~0.9secs, in contrast to ~1.5secs in mouse cortical neurons (Wang et al, 2006). This finding suggests that ASICs in human brain neurons could be more rapidly reactivated by repeated pH decreases or fluctuations expected to occur during high-frequency synaptic activities (e.g., long-term potentiation and epileptic seizures). This property might make the ASICs in human brain neurons more physiologically and pathologically relevant than the ASICs in rodent brain neurons. As homomeric ASIC1a channels in heterologous expression systems show slower recovery, the reason for the fast recovery of the ASIC current in human brain neurons warrants future investigation.
The steady-state inactivation of ASICs in human cortical neurons also differs from that in mice. The pH50 for steady-state inactivation of ASICs in human cortical neurons is ~7.0 (Figures 4E and 4F), which is more acidic than the currents in mouse cortical neurons, in which a pH50 of ~7.2 is reported (Wang et al, 2006). The more acidic pH50 for steady-state inactivation of ASICs indicates a reduced inactivation of ASICs by preexisting moderate pH drops, thus making more channels available for activation by subsequent severe acidosis.
Pharmacologically, ASIC currents in human cortical neurons show increased sensitivity to PcTX1, a specific antagonist for homomeric ASIC1a channels, further suggesting that homomeric ASIC1a channels likely have a predominant role in mediating acid-activated currents in these neurons. In contrast, ASIC currents in human cortical neurons are not affected by high concentrations of extracellular Zn2+, which is known to potentiate ASIC2a-containing channels (Baron et al, 2001). Thus, human cortical neurons likely lack functional ASIC2a-containing channels. Our RT-PCR and western blot analysis showing low levels of ASIC2a mRNA and protein also support the lack of ASIC2a-containing channels in these neurons. This is in contrast to the rodent brain neurons where high levels of both ASIC1a and ASIC2a are expressed and the acid-activated currents are mediated by a combination of homomeric ASIC1a, ASIC2a, and heteromeric ASIC1a/ASIC2a channels (Baron et al, 2002; Askwith et al, 2004; Chu et al, 2004; Xiong et al, 2004).
Using an ion-substitution protocol and Fura-2 fluorescent imaging techniques, we show that acidosis can induce [Ca2+]i accumulation in human brain neurons through ASIC1a activation in the presence of blockers for NMDA, AMPA, and voltage-gated Ca2+ channels. This acid-induced increase of [Ca2+]i is inhibited by pharmacological ASIC blockade, suggesting that activation of ASICs can induce glutamate receptor and voltage-gated Ca2+ channel-independent accumulation of [Ca2+]i in human brain neurons.
Using an in vitro cell toxicity model, we further showed that acidosis induces glutamate-independent injury of human cortical neurons. This acid-induced injury of human cortical neurons is inhibited by the blockade of ASIC1a channels. Taken together, these findings suggest that acidosis can injure human brain neurons through ASIC1a activation, and that targeting ASIC1a channels may be an effective neuroprotective strategy for human stroke, wherein acidosis is a common feature (Siesjo, 1988).
It is worth mentioning that the brain tissues used in these studies were obtained from patients with brain tumor. It is not clear whether the physiologic/pharmacological properties of ASICs in these neurons are different from neurons in nontumor patients. Ongoing studies in neurons isolated from cortical tissues of trauma patients showed similar electrophysiological/pharmacological properties of ASICs (not shown).
Owing to the limitation of current available pharmacological agents, e.g., the lack of specificity for amiloride and a large molecule for PcTX1, future studies will also consider using molecular biologic approaches, e.g., ASIC gene knockdown, to test the role of ASICs in acidosis-mediated injury of human brain neurons.
The authors declare no conflict of interest.