Slowing of the Time Course of Acidification Decreases the Acid-Sensing Ion Channel 1a Current Amplitude and Modulates Action Potential Firing in Neurons

Abstract

Acid-sensing ion channels (ASICs) are H⁺-activated neuronal Na⁺ channels. They are involved in fear behavior, learning, neurodegeneration after ischemic stroke and in pain sensation. ASIC activation has so far been studied only with fast pH changes, although the pH changes associated with many roles of ASICs are slow. It is currently not known whether slow pH changes can open ASICs at all. Here, we investigated to which extent slow pH changes can activate ASIC1a channels and induce action potential signaling. To this end, ASIC1a current amplitudes and charge transport in transfected Chinese hamster ovary cells, and ASIC-mediated action potential signaling in cultured cortical neurons were measured in response to defined pH ramps of 1-40 s duration from pH 7.4 to pH 6.6 or 6.0. A kinetic model of the ASIC1a current was developed and integrated into the Hodgkin-Huxley action potential model. Interestingly, whereas the ASIC1a current amplitude decreased with slower pH ramps, action potential firing was higher upon intermediate than fast acidification in cortical neurons. Indeed, fast pH changes (<4 s) induced short action potential bursts, while pH changes of intermediate speed (4-10 s) induced longer bursts. Slower pH changes (>10 s) did in many experiments not generate action potentials. Computer simulations corroborated these observations. We provide here the first description of ASIC function in response to defined slow pH changes. Our study shows that ASIC1a currents, and neuronal activity induced by ASIC1a currents, strongly depend on the speed of pH changes. Importantly, with pH changes that take >10 s to complete, ASIC1a activation is inefficient. Therefore, it is likely that currently unknown modulatory mechanisms allow ASIC activity in situations such as ischemia and inflammation.

Keywords: ASIC; Hodgkin-Huxley model; acidification; kinetic model; neuronal signaling; patch-clamp.

Figure 1

Experiments analyzing the kinetics of ASIC1a. This figure reports the experiments done for creating a kinetic model of ASIC1a function. (A) Kinetic models describing ASIC function; left, model containing three functional states (C, closed; O, open; D, desensitized). Right, Hodgkin-Huxley (HH)-type model, in which activation transitions are horizontal, and (de)sensitization transitions are vertical, containing an open, a closed and two desensitized states, D and DC. (B–E) Currents were measured by outside-out (B,C) or whole-cell (D,E) voltage-clamp to −60 mV from CHO cells stably expressing ASIC1a. (B) ASICs were activated every 40 s by switching from pH 7.4 to acidic test stimuli for 3 s. The current appearance and decay were each fitted to a single exponential. A schematic protocol and representative current traces are shown. The two traces are from different patches. (C) To determine the kinetics of deactivation, the ASICs were activated every 40 s by switching from pH 7.4 to the stimulation pH 6.0 for 120 ms, and then exposed to pH 7.4. (D) The kinetics of entry into steady-state desensitization (SSD) were determined with the basic protocol shown on top. Channels were first exposed to pH 6.0 for 5 s (control response), followed by a recovery at pH 7.4 for 40 s. Next, the channels were exposed to a given conditioning pH for different time periods (one duration per sweep), followed by a second switch to pH 6.0 to assess the fraction of the channels having desensitized. Middle panel, pH protocol and representative current traces of an SSD experiment with conditioning pH 7.1. Bottom panel, kinetics of current decay as a function of exposure duration, shown for different conditioning test pH conditions (n = 2–11). (E) Recovery from desensitization; protocol shown in top panel. Channels were first exposed to pH 6.0 for 10 s, leading to opening and subsequent desensitization. Next, they were exposed to a given conditioning pH for increasing time periods, followed each time by a second exposure to pH 6.0. Middle panel, pH protocol and current traces of a representative experiment with conditioning pH 7.3. Bottom panel, kinetics of current recovery from desensitization, as a function of exposure duration, shown for different recovery test pH (n = 3–13). (F) Rate constants derived from the different protocols used in B-E, shown as a function of pH. n = 3–36 (opening), 11 (deactivation), 2–19 (current decay), 4–8 (SSD kinetics), and 3–11 (recovery from desensitization).

Figure 2

Development of the model of the ASIC1a current. Experimental data are shown in blue, the initial functions a_∞ and s_∞ fit to the experimental data are shown in green and the corresponding simulated data in red. Currents were measured from CHO cells stably expressing ASIC1a, in the whole-cell mode (A,C; B, right panel, desensitization from the closed state and recovery from desensitization) or from excised outside-out patches [B, left panel (opening and deactivation) and right panel, desensitization from the open state]. (A) Steady-state activation (left) and steady-state desensitization (SSD, right). For measuring steady-state activation curves, cells were exposed to the conditioning pH 7.4, and the solution was changed every 50 s for 5 s to an acidic test pH to activate ASICs, n = 8–10. To determine the SSD pH dependence, channels were activated by a first exposure to pH 6 for 8 s (I₁), which was followed by a 40-s exposure to pH 7.4, a 60-s exposure to the test conditioning pH and a 8-s exposure to pH 6 (I₂). After this, channels were exposed during 42 s to pH 7.4, and subsequently a new cycle, with a different test conditioning pH was started. The I₂/I₁ ratio was then plotted as a function of the conditioning pH, n = 4–9. (B) Rates of opening and deactivation (left) and desensitization and recovery from desensitization (right). Experimental data are as in Figure 1F; protocols as described in the legend of Figure 1. (C) Same as in (A), but using the modified functions a_∞ and s_∞ of the final model. In all simulations, the model was subjected to the same pH protocols and its output to the same analyses as in the experiments.

Figure 3

Model of the ASIC1a current: simulations vs. experiments. Experimental data are shown in blue and model results in red. Currents were measured from CHO cells stably expressing ASIC1a, in the whole-cell mode (A,B; D except for desensitization from the open state) or from excised outside-out patches (C; D, desensitization from the open state). (A) Steady-state activation. (B) Steady-state desensitization (SSD). Experimental data in A and B are the same as in Figures 2A,C; the corresponding protocols are described in the legend of Figure 2. (C) Rate of opening and deactivation. (D) Rate of desensitization and recovery from desensitization. Experimental data in (C,D) are the same as in Figure 1F. In all simulations, the model was subjected to the same pH protocols and its output to the same analyses in the experiments.

Figure 4

Slowing of the pH change reduces the current amplitude and the transported charge. (A) Pseudocolor image by the pH-sensitive fluorescent dye 5(6)-FAM of the same cell at pH 7.4 and at pH 6.0. (B) Fluorescence intensities (green traces, upper panel) as a function of time in response to pH ramps of different speed from pH 7.4 to pH 6.0. The 10–90% fall-time (FT) of the fluorescence change corresponds to the interval between the two red squares (Materials and Methods). Each fluorescence trace (top panel) was measured simultaneously with the corresponding ASIC1a current trace (middle panel). Bottom panel, traces created by the kinetic model in response to the above pH ramps. The experimental data in (B–I) are from whole-cell patch-clamp at −60 mV. (B–H) Current and transported charge values are normalized to the corresponding response measured in the same cell to a rapid pH change (ramp FT < 4 s). (C,D) ASIC1a peak current amplitudes in response to pH 7.4–6.0 ramps (C) or pH 7.4–6.6 ramps (D) of different duration, shown as function of pH ramp FT (n = 14–20). The model prediction is indicated with black symbols and the connecting line. (E) Normalized peak current amplitudes for ramp durations classified into intermediate (ramp FT 4–10 s) and slow (ramp FT > 10 s; n = 4–16). ****p < 0.0001, different from rapid change (ANOVA, Tukey post-hoc test). (F,G) Transported charge (Q), the current integral over time, is indicated for the same protocols as shown in (C,D) (n = 12–17). The model prediction is indicated with black symbols and the connecting line. (H) Normalized transported charge for ramp durations classified as in E (n = 4–16). *p < 0.05; ****p < 0.0001, different from rapid change (ANOVA, Tukey post-hoc test). (I) Time constant of ASIC current appearance in pH 7.4–6.0 (red symbols) and pH 7.4–6.6 ramps (orange), plotted as a function of the pH ramp FT. The r² values were 0.906 (p < 0.001, pH 7.4–6.0 ramps) and 0.846 (p < 0.001, pH 7.4–6.0 ramps), n = 21–29. (J,K) Heat map of the simulated peak ASIC1a current (J) or charge transport (K) as a function of the pH ramp lowest value (pH) and duration (FT) in seconds. Shown are the amplitude of the maximal current (I_peak, μA cm⁻²) and the transported charge (Q, mC cm⁻²) according to the scaling bar.

Figure 5

Neuronal model behavior (classical Hodgkin-Huxley model). Shown are two simulations of the neuronal response to a pH drop to pH 6.0 of 1 (A) resp. 6 s duration (B). The center column shows a zoom of 120 ms duration selected toward the middle of the firing. From top to bottom, the figures present the time course of the pH, the membrane potential (V_m), the open probabilities of the gates m, h, and n (see Materials and Methods, section Incorporation of I_ASIC1a into the Hodgkin-Huxley Model of the Nerve Action Potential), the ASIC behavior in terms of activation of the activation (a) and the sensitization gate (s), and of the product a·s, and the current related to the voltage-gated sodium (Na) and potassium (K) channels, the leak current (L), and the ASIC-mediated sodium current (ASIC). For both ramp types, phase-plane plots of the first AP, an AP in the middle and the last AP of the burst are shown in the right column.

Figure 6

Neuronal modeling and experiments in cortical neurons to assess the response to slow pH changes. (A) Simulated neuronal activity in response to pH 7.4–6.0 ramps using the Hodgkin-Huxley (HH) neuronal model in which the ASIC1a model was integrated. Membrane potential changes in response to the pH change (bottom panel) are shown in the middle panel; selected parts of the traces are shown in the top panel on an expanded time scale. (B,C) representative experiments with WT (pH 7.4–6.0 ramps) and ASIC2^−/− cortical neurons (pH 7.4–6.6 ramps) in current-clamp. Baseline current was injected to obtain a resting membrane potential close to −60 mV. For the slow pH ramp with ASIC2^−/− cortical neurons, one trace with practically no (bottom), and one with high activity (top) is shown. (D,E) Number of action potentials (APs) induced by rapid (ramp FT < 4 s), intermediate (ramp FT = 4–10 s) and slow (ramp FT > 10 s) pH 7.4–6.0 ramps (D) or pH 7.4–6.6 ramps (E) of WT, ASIC1a^−/− and ASIC2^−/− cortical neurons (n = 4–15). (F) pH dependence of ASIC currents in WT, ASIC1a^−/− and ASIC2^−/− cortical neurons. Currents were measured in whole-cell patch-clamp at −60 mV. Normalized current amplitudes are plotted, n = 2–10. The pH₅₀ values were 6.29 ± 0.23 (WT, n = 10) and 6.67 ± 0.03 (ASIC2^−/−, n = 4). For ASIC1a^−/− neurons, the current amplitudes still increased between pH 4.0 and 3.2, and the pH₅₀ is therefore ≤ 4.10 ± 0.04 (n = 2). (G,H) Number of APs of WT and ASIC2^−/− cortical neurons, plotted as a function of the pH ramp FT, in pH 7.4–6.0 ramps (G, n = 23–28) and pH 7.4–6.6 ramps (H, n = 23–36). (I,J) Proportion of pH 7.4–6.0 (I) or pH 7.4–6.6 ramps (J) that induced at least 1 AP, presented as probability (n = 5–16).

Figure 7

Analysis of acid-induced action potential firing in cultured cortical neurons. The experimental conditions were as described in the legend to Figure 6. (A,D) Firing time, calculated as the duration between the beginning of the first and the end of the last AP in AP bursts, plotted for pH 7.4–6.0 ramps (A) and pH 7.4–6.6 ramps (D) with ramp FT < 4 s (rapid) and >4 s (Int./slow), n = 8–20. Only experiments with at least one AP were considered. (B,E) Firing time, plotted as a function of the pH ramp FT, for pH 7.4–6.0 ramps (B) and pH 7.4–6.6 ramps (E), n = 14–22. (C,F) AP firing frequency, calculated as AP number/firing time, for pH 7.4–6.0 (C) and pH 7.4–6.6 ramps (F), n = 7–12. *p < 0.05; **p < 0.01; ***p < 0.001, different between the indicated conditions. Statistical tests were Kruskal-Wallis and Dunn's post-hoc test for A and F, ANOVA and Tukey post-hoc test for (C,D).

Figure 8

Phase-plane analysis of neuronal action potentials. The data are from current-clamp experiments with WT cultured cortical neurons; baseline current was injected to obtain a resting membrane potential close to −60 mV. The pH ramp protocols produced in control experiments a fall time of 2.0 ± 0.1 s (rapid), 4.4 ± 1.8 s (intermediate), and 17.8 ± 6.7 s (slow; n = 3, mean ± SEM), respectively. Of each induced AP burst, the properties of the first, the last, and of an AP in the center of the burst was submitted to the phase-plane analysis. Red, purple, and blue data points were obtained from rapid, intermediate and slow pH ramps, respectively. (A) Representative APs and corresponding phase-plane plots of an AP burst induced by a pH 7.4–6.6 ramp of intermediate speed. (B,D) Threshold of AP generation in response to pH 7.4–6.0 (B) or pH 7.4–6.6 ramps (D), n = 4–8. (C,E) Maximal AP upstroke velocity measured in experiments with AP induction by pH 7.4–6.0 (C) or pH 7.4–6.6 ramps (E), n = 4–8. These parameters were determined from phase-plane plots as indicated in Materials and Methods. A two-way ANOVA test identified significant contributions of the AP position to the AP threshold (p < 0.001 in pH 7.4–6.0 ramps, p < 0.05 in pH 7.4–6.6 ramps) and the maximal AP upstroke velocity (p < 0.0001 in pH 7.4–6.0 ramps, p < 0.01 in pH 7.4–6.6 ramps) and of the ramp speed to the AP threshold (p < 0.05 for pH 7.4–6.0 and pH 7.4–6.6 ramps).

Figure 9

Heat maps of the simulated neuronal activity as a function of the pH ramp lowest value and duration (RT) in seconds, for ramps starting at pH 7.4. Shown are the number of APs (A) and the firing time (s) (B) for the neuronal Hodgkin-Huxley model integrating ASIC1a.

Figure 10

Responses of the neuronal models to the variability of the conductance parameters. Responses in terms of firing time and number of APs of the Hodgkin-Huxley neuronal model integrating the ASIC1a model with pH 7.4–6.0 ramps. The differently colored lines highlight the effect of varying the maximal conductance of ASIC1a (g_ASIC), sodium (g_Na), and potassium (g_K) channels, as well as of the leak current (g_L) by ±10% of the optimized corresponding parameters. The firing time (A) and the number of APs (B) are shown in purple for optimized parameters, and as blue and red lines for optimized corresponding parameters −10% (blue) and +10% (red). Note that the speed of pH ramps is presented here as total ramp time, and not ramp FT, as in other figures.

Figure 11

Responses of the neuronal models to variations of Na_v parameters. The neuronal modeling results are summarized for pH ramps of three durations [short, FT (=fall time) 1 s; intermediate (FT 5 s); slow (FT 18 s)], and in which the following Na_v parameters were varied: Kinetics of the m-gate, f_m (A–D), shift of the V_1/2 of the m-gate, s_m (E–H), kinetics of the h-gate, f_h (I–L), and shift of the V_1/2 of the h-gate, s_h (M–P). The responses are shown in terms of number of APs (1st column from left), firing time (2nd column from left), and two parameters from phase-plane plot analysis of the first, middle and last AP in the burst, the spike threshold (3rd column from left) and the maximal AP upstroke velocity (1st column from right). The colors of the symbols refer to the speed of the pH ramp, and the symbol types in the two right columns refer to the position of the studied AP within the burst, as indicated in the figure.