Introduction

Tissue ischaemia leads to rapidly declining oxygen levels and diminished capacity for oxidative phosphorylation. The consequential switch to anaerobic metabolism, to maintain cellular ATP production, results in increased generation of metabolic acid. This, in combination with the lack of perfusion, culminates in a rapidly developing tissue acidosis. For example, during myocardial ischaemia, pHo falls to ≈6.8 within the first 8 min and ≈6.2 after 20 min [20, 21, 59, 109]. Tissue acidoses together with elevation of extracellular K+, ATP and bradykinin are all thought to play a role in the excitation of sensory neurones within ischaemic tissue and the generation of ischaemic pain [5, 22, 24, 70, 76, 80, 85].

Acid is thought to excite sensory neurones by activating pH-sensitive cation channels [12, 68]. Two main types of proton-gated cation channel have been found in sensory neurones: acid-sensing ion channels (ASICs) and the transient receptor potential vanilloid 1 receptor (TRPV1). ASICs are predominantly sodium selective and can be blocked by amiloride [46–48, 99]. The ASIC family contains four distinct genes, of which ASICs 1 and 3 are both expressed in sensory neurones [46–48, 100]. ASICs and, in particular, ASIC3 have been implicated in cardiac sensory neurone activation by ischaemia [10, 78]. Although thought to be primarily gated by protons, activation of ASIC3 may be further enhanced by the presence of lactate [41, 59]. TRPV1 is expressed predominantly in small sensory neurones from the dorsal root ganglia (DRG) [29, 33, 58, 60, 104, 111] and is a polymodal sensor. TRPV1 may be activated by capsaicin, heat, protons and proalgesic substances. This ability of TRPV1 to integrate both physical and chemical stimuli suggests that it plays a major role in pain transduction [16, 44, 82]. Unlike most ASICs, TRPV1 is highly permeable to both Na+ and Ca2+ (P Ca/P Na between 1 and 10) [8, 32, 96, 97].

In addition to acting as pain sensors, sensory neurones also release neuropeptides during ischaemia (e.g. calcitonin gene-related peptide (CGRP) or substance P) which may act locally to modulate tissue function and/or blood flow within the ischaemic zone [37, 50]. Whilst electrical excitation of sensory nerves by acidosis/ischaemia most probably results directly from membrane depolarisation (due to activation of the above proton-gated cation channels), secretory responses are likely to be mediated via a rise in intracellular Ca2+. There are numerous factors that could contribute to changes in [Ca2+] i during ischaemia including voltage-gated Ca2+ entry secondary to electrical excitation, direct Ca2+ influx through proton-gated channels, Ca2+ release from internal stores and modulation of Ca2+ uptake, buffering or extrusion. In the present study, we have therefore investigated the effects of acidosis on intracellular Ca2+ regulation in small, capsaicin-sensitive, sensory neurones (15–30 μm) from cervicothoracic DRG. These neurones were exposed to four different acid stimuli with pHo values 6.8, 6.2 (with and without lactate), and 5.0 simulating the initial phase of an ischaemic event, more prolonged ischaemia and an extreme acidosis of the type typically used to study acid-sensitive cation channel function in vitro. Using a pharmacological approach, we have characterised the Ca2+-entry pathways and stores that contribute to elevation of [Ca2+] i during acidosis. In addition, since anoxia and aglycaemia can also have profound effects on Ca2+ metabolism [34], we combined these stimuli with acidosis to more closely simulate ischaemic conditions and to investigate their collective effect on [Ca2+] i .

Materials and methods

Neurone dissociation

Adult Wistar rats of either sex aged between 6 and 8 weeks (130–170 g) were sacrificed by an overdose of halothane (4%). Cervicothoracic DRG (C4-Th6) were removed under sterile conditions and were immediately transferred into cooled (on ice) Ca2+- and Mg2+-free phosphate-buffered saline (PBS), pH 7.4. After cleaning the ganglia from surrounding tissue, the ganglia were incubated in an enzymic dispersion media comprising 10 mg/ml collagenase type I (208 U/mg, Worthington, CLS-1, MON4393), 1 mg/ml trypsin (9.3 U/mg, Sigma, T-4665), in PBS and with 60 μM CaCl2 and 33 μM MgCl2. The ganglia were incubated at 37°C for 35 min. Following enzyme treatment, the ganglia were washed once in PBS (Ca2+- and Mg2+-free) and once in Dulbecco’s modified Eagle’s medium (DMEM; containing 10% fetal bovine serum, 1.2 mM l glutamine), before mechanical trituration in 1.5 ml of DMEM. The dissociated cells were then washed twice by centrifugation (at 1,000×g for 5 min) followed by resuspension in fresh DMEM. Following the final wash, the cell pellet was resuspended in 500 μl basal TNB-100 containing protein–lipid complex (Biochrom, Berlin, Germany), penicillin (100 IU/ml), streptomycin (100 μg/ml) and 10 μM/ml nerve growth factor (TNB). Following a second trituration, the neurones were seeded onto poly-l-lysine and laminin-coated coverslips (6 mm in diameter) and incubated in sterile culture dishes in a humidified chamber at 37°C and 5% CO2/95% air for 2 h. After this incubation period, a further 3-ml TNB was added to each culture dish. The neurones were then kept in the incubator for at least 30 min before being used for experiments. These neurones were typically used within 2 days of isolation.

Fluorescence measurements

Measurements of Fura-2 fluorescence were performed using a microspectrofluorometer based on an epifluorescence microscope (Nikon Diaphot 200, Japan) fitted with photomultiplier tubes (PMT; Thorn, EMI, UK) to detect emitted fluorescence and a motor driven monochromator (Cairn Instruments, Kent) with xenon lamp to provide the excitation light source. Fura-2 was excited alternately at 340 and 380 nm (±8 nm) for 250 ms at each wavelength with the cycle repeated at 1 Hz. Emitted fluorescence from Fura-2 was passed through a bandpass filter centre wavelength 510 nm (±20 nm). Bandpass filtered fluorescence was detected using a PMT air cooled to −20°C (Thorn, EMI, UK). The output from the PMT was integrated over each illumination period and recorded on a microcomputer using a micro CED1401 and Spike 4 software (Cambridge Electronic Design). For Fura-2, the ratio of fluorescence at 340 nm relative to that at 380 nm (R) was also calculated and recorded using Spike 4 software.

Selection and superfusion of neurones

Neurones were placed in a recording chamber with a volume of approximately 100 μl mounted on the stage of the epifluorescence microscope (see below). This chamber was perfused with solutions at a flow rate of approximately 2 ml min. Solutions were delivered from reservoirs kept in a water bath to the recording chamber via medical grade stainless steel tubing articulated by short sections of Pharmed tubing (Norton performance plastics, UK). A mechanically driven two-way tap which allowed a rapid change between two different solutions was placed within a few inches of the recording chamber. A heating coil was placed around a short section of tubing between the tap and the chamber to ensure solutions remained at 37°C. This arrangement allows rapid solution exchange and tight control over the gas content and temperature of solutions.

Neurones were observed and fluorescence recorded through a X40 fluorescence grade objective (n.a. 1.30). Sensory neurones were initially selected by soma size (15–30 μm) and then tested for a response (a robust increase in [Ca2+] i ) to capsaicin. This capsaicin test was usually conducted at the end of the experiment and proved positive in >80% of the neurones selected by the above size criteria. Only capsaicin-positive neurones are included in this study.

Loading of neurones with Fura-2-AM

To introduce Fura-2 into neurones, they were incubated in either a 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) buffered saline (for in vivo calibrations) or a bicarbonate buffered saline (for experiments) containing 5 μM Fura-2-AM (Molecular Probes, Leiden, The Netherlands) at room temperature for 25 min in a dark chamber. The HEPES buffered saline used comprised (in millimolars): HEPES 20, glucose 11, KCl 4.5, MgCl2 1, CaCl2 2.5, and NaCl 117, pH 7.4 at room temperature.

In vivo calibration of Fura-2

Fura-2-loaded neurones (see above) were incubated in a high-K+ zero-Ca2+ HEPES buffered solution (consisting of 150 mM KCl, 5 mM NaCl, 1 mM ethylenediaminetetraacetic acid and 1 mM ethylene glycol tetraacetic acid (EGTA)) containing 10 μM ionomycin (Sigma, Dorset, UK) for 10–20 min. After this incubation, the neurones were placed in a flow through incubation chamber mounted on the microspectrofluorometer and perfused with the same high-K+ zero-Ca2+ HEPES solution but containing 1 μM ionomycin and at 37°C. After a further 5-min incubation in this solution, Fura-2 fluorescence was recorded from five identified sensory neurones. The ratio of fluorescence obtained under these conditions was deemed equivalent to the calibration constant R min [30]. The perfusate was then changed to a high-K+ high-Ca2+ HEPES saline containing 150 mM KCl, 5 mM NaCl, 10 mM CaCl2 and 1 μM ionomycin. The change in fluorescence ratio was followed in one of the five identified neurones until it reached a new stable value, and then the fluorescence ratio in it and the other four identified neurones were recorded and deemed to be equivalent to the calibration constant R max. The ratio of fluorescence at 380 nm in Ca2+-free buffer divided by that obtained in high Ca2+ buffer (S f2/S b2) was also calculated for each neurone. The mean values obtained for R min, R max and (S f2/S b2) were then used to calibrate measurements of the fluorescence ratio in subsequent experiments using the equation [Ca2+] = (R − R min)/(R max − R) × S f2/S b2 × Kd [30].

Solutions

The standard bicarbonate buffered Tyrode solutions contained (in millimolars) NaCl 117, KCl 4.5, CaCl2 2.5, MgCl2 1, HCO −3 23 and glucose 11. Glucose-free Tyrode solution was prepared by replacing glucose with 11 mM sucrose. For Ca2+-free solution, the CaCl2 was omitted and 1 mM EGTA was added. For Tyrode solution with elevated KCl concentration (50 mM), the NaCl concentration was reduced to 71.5 mM; all other constituents remained the same. Equilibration of these solutions with 5% CO2 and 95% air achieved normocapnic conditions with pH 7.4 at 37°C. Moderate hypercapnic acidosis (pHo 6.8) was achieved by increasing the CO2 content of the equilibrating gas from 5% to 20%. To achieve a combined respiratory and metabolic acidosis, Tyrode solution with reduced NaHCO3 (10 mM, 130 mM NaCl) was equilibrated with 20% CO2, which resulted in a pHo 6.2. pH 5.0 solutions were attained by lowering NaHCO3 to 2 mM (and increasing NaCl to 138 mM) whilst equilibrating with 20% CO2. Solutions simulating a lactic acidosis contained (in millimolars) Na-lactate 15, NaCl 126, KCl 4.5, CaCl2 2.5, MgCl2 1 and glucose 11. The pH of this solution was adjusted to 6.2 at 37°C by equilibration with 20% CO2 and addition of NaOH.

Anoxic solutions were generated by replacing the air in the gas used to equilibrate the above solutions with nitrogen (i.e. gas mixtures were x% CO2 + 100 − x% N2) and with the further addition of 0.5 mM Na2S2O4 [72] following 15–30 min prior equilibration with the appropriate oxygen-free gas mixture. The addition of Na2S2O4 did not cause any detectable change in solution pH. All solutions were equilibrated with appropriate gas mixes at 37°C in a water bath for at least 30 min before use.

Drugs

(E)-3-(4-t-butylphenyl)-N-(2,3-dihydrobenzo[b][1,4] dioxin-6-yl)acrylamide (AMG 9810, 1 μM), N-(4-tertiarybutylphenyl)-4-(3-cholorphyridin-2-yl) tetrahydropryazine-1(2H)-carbox-amide (BCTC, 1 μM), capsazepine (CPZ; 10 μM), cyclopiazonic acid (CPA; 10 μM), carbonyl cyanide-p-trifluoromethoxyphenylhydrazone (FCCP; 1 μM) and amiloride (100 μM) containing solutions were prepared from stock solutions in dimethyl sulfoxide. Capsaicin was added from stock solutions in ethanol. NiCl2 (2.5 mM), CdCl2 (100 μM), MnCl2 (1 mM) and GdCl3 (0.5–1 mM) containing solutions were prepared from 0.5 to 1 M stock solutions in H2O.

Statistics

Values were expressed as mean ± standard error of mean (SEM). Statistical significance was tested using the paired Student’s t test, or Wilcoxon signed-rank test for experiments with non-Gaussian distribution. Statistical testing of in vitro calibration data was performed using one-way analysis of variance and post hoc analyses were carried out using Bonferroni’s multiple comparison, calculated by SPSS 12.0 software for windows. Level of significance was set at p < 0.05.

Results

Acidosis evoked rise in [Ca2+] i in sensory neurones

Tissue acidosis is believed to be a major activator of sensory neurones transmitting ischaemic pain. Here, we investigated changes in [Ca2+] i in response to acidosis. Four different types of acidosis were tested: 20% CO2 in normal, 23 mM, HCO −3 (pHo 6.8) simulating a simple hypercapnic acidosis; 20% CO2 in 10 mM HCO −3 and 20% CO2 in 15 mM Na-lactate (both pHo 6.2) simulating a mixed hypercapnic and metabolic acidosis (with and without lactate ions present); and 20% CO2 in 2 mM HCO −3 (pHo 5.0) simulating a very severe hypercapnic/metabolic acidosis.

In the majority of sensory neurones (65%) exposure to acidosis (either pHo 6.8, 6.2 or 5.0) evoked a rise in [Ca2+] i . This rise in [Ca2+] i typically had a biphasic kinetic with an initial Ca2+ transient leading into a sustained elevation of [Ca2+] i which remained throughout the exposure to acid (Fig. 1a). In a few neurones, however, only an initial Ca2+ transient was obvious with [Ca2+] i returning back to baseline whilst still under acidic conditions (e.g. Fig. 1c). Another small subgroup of neurones lacked an obvious Ca2+ transient and instead responded with a rapid rise in [Ca2+] i up to a plateau that was then sustained throughout exposure to acidosis (Fig. 1b).

Fig. 1
figure 1

Extracellular acidosis evokes rise in [Ca2+] i in sensory neurones. a–c Effects of a hypercapnic (20% CO2, pHo 6.8) and a mixed acidosis (20% CO2, pHo 6.2) on [Ca2+] i in capsaicin-sensitive neurones in the presence and absence of extracellular Ca2+. In the presence of extracellular Ca2+, acidosis evoked an initial transient followed by a sustained rise in [Ca2+] i in the majority of neurones (e.g. a and c). The amplitude of this Ca2+-transient amplitude was defined as the maximal increase in [Ca2+] i within the first minute of exposure (see c). The amplitude of the sustained rise was defined as the mean increase in [Ca2+] i during the final 30 s of the exposure period (see a). The Δ[Ca2+] i of the transients were thus defined as the difference between the sustained rise and the transient peak (Δtrans., see a). d Comparison of the amplitude of initial Ca2+ transients evoked by acidosis pHo 6.8 or 6.2 in the presence and absence of extracellular Ca2+ (**p < 0.01). e, f Comparison of Ca2+ responses evoked by mixed acidosis (20% CO2, pHo 6.2) in the presence and absence of lactate (15 mM). f Comparison of the amplitude of the Ca2+ transient in response to lactic acidosis normalised to the response to acidosis in the absence of lactate. Time scale bars in a–c and e 200 s. Exposure periods are indicated by horizontal bars. Bar charts present means + SEM; numbers of experiments are given in parenthesis

Full size image

The peak amplitudes of acid-induced Ca2+ transients were quantified as the maximal increase in [Ca2+] i attained within the first minute of exposure to acidosis relative to baseline (e.g. Fig. 1c). The sustained elevation was defined as the averaged rise in [Ca2+] i from the final 30 s of acidosis exposure relative to baseline (Fig. 1a). The Δ[Ca2+] i of the transients (Δ trans) were thus defined as the difference between the sustained elevation and the transient peak (e.g. Fig. 1a). A Ca2+ transient was only considered to have occurred where this Δ[Ca2+] i measurement was greater than the mean + 2 times the standard deviation of the sustained elevation in [Ca2+] i . The duration of the Ca2+ transient was characterised by measuring the time taken for the [Ca2+] i to fall to half of the peak value (t 1/2).

In neurones responding with an initial Ca2+ transient, the mean Δtrans evoked by hypercapnic acidosis alone (pHo 6.8) was 112 ± 24.5 nM (n = 10) and that evoked by a combined hypercapnic/metabolic acidosis (pHo 6.2) was 871 ± 352 nM (n = 7). These acid-evoked Ca2+ transients were substantially attenuated in Ca 2+o -free media to 7 ± 0 nM (n = 10, **p < 0.01) for a hypercapnic acidosis and 120 ± 22 nM (n = 7, *p < 0.05) for a combined hypercapnic/metabolic acidosis (see Fig. 1a, c, d). Thus, for both types of acid stimuli, the transient elevation in [Ca2+] i in response to acidosis was strongly dependent upon the presence of extracellular Ca2+. In contrast, the sustained rise in [Ca2+] i was unaffected by Ca2+-free solutions (see below). Acid-evoked Ca2+ transients were often relatively brief with t 1/2 = 27.8 ± 1.1 s at pHo 6.8 (n = 27) and 28.1 ± 1.3 s at pHo 6.2 (n = 27).

Acidosis due to tissue ischaemia is often accompanied by an increase in lactate from anaerobic respiration. Lactate has been reported to augment acid-evoked inward currents in sensory neurones [41]. We therefore compared a simple mixed acidosis (20% CO2, pH 6.2) with an equivalent acidosis in the presence of 15 mM lactate (also 20% CO2, pH 6.2). Eighty percent of neurones showed similar Ca2+ transients to both stimuli, and in these neurones, there was no significant difference between the amplitudes of these Ca2+ transients (Na-lactate 106 ± 11%, n = 26, p = 0.442, data normalised to pHo 6.2 with 20% CO2; Fig. 1f). A few neurones responded only to one or other of the two stimuli.

Correlation between capsaicin- and acidosis-evoked Ca2+ transients

In this study, we have focussed upon sensory neurones selected by morphology and sensitivity to capsaicin. The capsaicin receptor (TRPV1) is also acid sensitive and would be expected to contribute to acid-evoked Ca2+ influx described above. We therefore sought to compare the amplitude of [Ca2+] i responses to capsaicin (100 nM) with those evoked by acidosis. A significant positive correlation was observed between the amplitude of the Ca2+ transient evoked by capsaicin (100 nM) and that evoked by pHo 6.8 (r = 0.49, p < 0.05, n = 22) and pHo 5.0 (r = 0.69, p < 0.05, n = 11; Fig. 2a, c), but not by pHo 6.2 (r = 0.4, p = 0.081, n = 20; Fig. 2b). It is, however, evident that this correlation was not particularly strong. We also noted a striking paradox in that a significant number of neurones (approximately 15–20%) responding to capsaicin did not respond to acidosis with a transient Ca2+ influx. This failure to respond to acidosis could not be attributed to low levels of functional expression of the capsaicin receptor since we found no significant difference between the amplitude of the capsaicin-evoked Ca2+ transients in acid responders (308 ± 78 nM, n = 39) vs. acid nonresponders (295 ± 86 nM, n = 12, p = 0.886; Fig. 2d). The importance of TRPV1 and other Ca2+ influx pathways in mediating acid-evoked Ca2+ influx was therefore investigated further.

Fig. 2
figure 2

Correlation between response to capsaicin and acidosis. a–c Correlation between maximum amplitudes of Ca2+ transients evoked by acidosis vs. capsaicin (100 nM). There was a weak positive correlation between the Ca2+ response to capsaicin and that to pHo 6.8 and 5.0 (r = 0.49, p < 0.05, n = 22 for pHo 6.8; r = 0.69, p < 0.05, n = 11 for pHo 5.0) but not pHo 6.2 (r = 0.4, p = 0.081, n = 20). d The maximum amplitudes of [Ca2+] i response to capsaicin (100 nM) in neurones responding (responder) or not responding (nonresponder) to external acidosis with a biphasic rise in [Ca2+] i . There was no difference in the [Ca2+] i response to capsaicin in these two subgroups (p = 0.886). e The amplitudes of Ca2+ transients evoked by extracellular acidosis significantly increase with falling pHo values (1 compared to pHo 6.8; 2 compared to pHo 6.2; **p < 0.01; ***p < 0.001)

Full size image

Influence of divalent and trivalent cations on Ca2+-entry pathways

The former experiments indicated that external acidosis has a dual action on [Ca2+] i . An initial Ca2+ transient was evoked by activation of a Ca2+-entry pathway, whereas the sustained elevation of [Ca2+] i was triggered independently from external Ca2+.

In order to determine the Ca2+-entry pathways involved in mediating the transient response to acidosis, we needed to find ways of selectively inhibiting the different types of Ca2+-permeable channels that may be involved. Divalent and trivalent cations are often useful in this respect. Cd2+ is widely used to block voltage-gated Ca2+ channels especially in electrophysiological experiments. In these sensory neurones, Cd2+ (100 μM) similarly inhibited voltage-gated Ca2+ entry elicited by 50 mM KCl (by 89% ± 7%, n = 5; Fig. 3a). The application of capsaicin in the presence of Cd2+ (100 μM), however, led to a rise in fluorescence ratio that exceeded that observed in the absence of Cd2+ and either did not recover or recovered only partially within the observation period (Fig. 3b). Removal of external Ca2+ failed to inhibit the effect of capsaicin in the presence of Cd2+ although it was inhibited by capsazepine (Fig. 3c). These data suggest that Cd2+ is able to permeate TRPV1 channels. Once inside the cell, Cd2+ will inevitably bind to Fura-2, as this has a much higher affinity for Cd2+ than for Ca2+, which will cause similar changes in Fura-2 fluorescence to that observed with Ca2+ [36]. Thus, Cd2+ is inappropriate for use as a blocker of voltage-gated Ca2+ entry under conditions in which TRPV1 channels might be active (e.g. such as in responses to acidosis).

Fig. 3
figure 3

Inhibition of Ca2+-entry pathways by divalent and trivalent cations. a Voltage-gated Ca2+ entry triggered by KCl exposure (50 mM for 5 s) is almost fully inhibited by 100 μM Cd2+ (to 11% of control). b Application of Cd2+ (100 μM) in Ca2+-containing solution amplified the [Ca2+] i response to capsaicin (100 nM). Note also that the [Ca2+] i response to capsaicin in the presence of Cd2+ often failed to fully recover. c In Ca2+-free solution, capsazepine (10 μM) prevented a rise in Fura-2 fluorescence in response to capsaicin and Cd2+ application (note that following second exposure to capsaicin and Cd2+ in Ca2+-free medium lacking capsazepine Fura-2 fluorescence increased dramatically and did not recover). d Inhibition of voltage-gated Ca2+ influx (50 mM KCl) by 2.5 mM Ni2+. e Ni2+ (2.5 mM) enhanced Ca2+ transients evoked by capsaicin. f Gd3+ (0.5 mM) inhibited almost completely capsaicin evoked rise in [Ca2+] i . The response to capsaicin recovered partially during a final capsaicin application omitting Gd3+. Exposure periods are indicated by horizontal bars. Time scale bars, 100 s

Full size image

Ni2+ at relatively high (millimolars) concentrations can also be used to block voltage-gated Ca2+ entry [62]; 2.5 mM Ni2+ thus blocked high-K+ (voltage-gated) Ca2+ influx (Fig. 3d) in these neurones but it did not inhibit the response to capsaicin; instead in the presence of Ni2+, we measured a slightly enhanced rise in [Ca2+] i (130 ± 10%, n = 5, compared to control; Fig. 3e). This enhanced response to capsaicin in the presence of Ni2+ recovered back to baseline after wash out of capsaicin (unlike that seen in the presence of Cd2+). Mn2+ ions can also pass through many Ca2+ permeable channels and, upon gaining access to the cytosol, will quench Fura-2 fluorescence. Fura-2 fluorescence quenching by Mn2+ is therefore a useful technique with which to monitor the activation of various Ca2+-entry pathways. As shown in Fig. 4a, 2.5 mM Ni2+ prevents high K+-induced Fura-2 quenching by Mn2+ but does not prevent capsaicin induced Fura-2 quenching by Mn2+ (Fig. 4b). Ni2+ is therefore a good blocker of voltage-gated Ca2+ entry but has little effect upon Ca2+ or Mn2+ entry through TRPV1.

Fig. 4
figure 4

Mn2+ quenching of Fura-2 fluorescence. This figure shows Fura-2 fluorescence at each individual excitation wavelength (340 and 380 nm) and the excitation ratio (340:380). In a, voltage-gated Ca2+ influx is triggered by brief depolarisation using KCl (50 mM for 5 s) causing a transient rise in Fura-2 fluorescence ratio. In the presence of Ni2+ (2.5 mM) to block voltage-gated Ca2+ channels, depolarisation by KCl in the presence of 1 mM Mn2+ (1 mM) does not cause quenching of Fura-2 fluorescence. Thus, Ni2+ blocks Mn2+ entry through voltage-gated Ca2+ channels. In b, application of capsaicin also evokes a transient rise of Fura-2 ratio. In Ca2+-free solutions and in the presence of Ni2+ (2.5 mM) to block voltage-gated Ca2+ channels, the simultaneous application of capsaicin and Mn2+ (1 mM) resulted in a rapid quench of Fura-2 fluorescence at both wavelengths. Thus, activation of capsaicin receptors causes Mn2+ influx independently of voltage-gated Ca2+-channel activity. c Exposure to acidosis (pHo 6.2) triggered a transient rise of Fura-2 ratio. In the presence of Ni2+ (2.5 mM) to block voltage-gated Ca2+ channels, the simultaneous exposure to acidosis (pHo 6.2) and Mn2+ (1 mM) caused a immediate quench of Fura-2 fluorescence of both individual wavelengths. Exposure periods are indicated by horizontal bars. Time scale bars, 100 s

Full size image

We next tested the effects of the trivalent cation Gd3+ upon the Ca2+ response evoked by capsaicin. Gd3+ (0.5 mM) substantially inhibited the [Ca2+] i response to capsaicin (reduced to 6 ± 4%, n = 6, p < 0.001 compared to control). This inhibition appeared to be only partially reversible (the third exposure to capsaicin following wash of gadolinium was 23 ± 7%, n = 6, p < 0.01, compared to control; Fig. 3f). Gd3+ is therefore an effective inhibitor of TRPV1-mediated Ca2+ entry although it lacks specificity, i.e. it also blocks voltage-gated Ca2+ channels [62]. Reversibility may also be slowFootnote 1.

In summary, Ni2+ blocks voltage-gated Ca2+ and Mn2+ entry, but not Ca2+ or Mn2+ influx through TRPV1. Gadolinium (0.5 mM) blocks both voltage-gated Ca2+ channels and TRPV1, and Cd2+ blocks voltage-gated Ca2+ entry but actually permeates TRPV1 sufficiently well to induce a major increase in Fura-2 fluorescence. These features can therefore be used to discriminate between a Ca2+-influx pathway utilising voltage-gated channels and one using TRPV1.

Acidosis-activated Ca2+-entry pathways in DRG neurones

In order to determine the route of Ca2+ entry in response to acidosis, we have studied the effects of a variety of pharmacological agents upon the amplitude of the Ca2+ transient generated at pHo 6.8, 6.2 and 5.0. In these experiments, control acid stimuli were applied before and after exposure to the pharmacological agent and the test acid stimulus; data were then normalised to the amplitude of the first control response (set as 100%).

Ni2+ failed to inhibit the Ca2+ transients evoked in response to pHo 6.8, 6.2 or 5.0 (Fig. 5). Indeed, there was some evidence that at pHo 6.2, Ni2+ enhanced the response to acidosis (276 ± 42%, n = 5, p < 0.01) when compared to the first control. We also noted a tendency for the second, post-Ni2+, control response (recovery) to be amplified relative to the first control (Fig. 5d) at pHo 6.8 (127 ± 10%, n = 6, p < 0.05), 6.2 (230 ± 87%, p < 0.05, n = 5) and 5.0 (248 ± 62%, n = 5, p < 0.05). Acidosis (pH 6.2) also activated Mn2+ influx and Fura-2 fluorescence quenching in the presence of 2.5 mM Ni2+ (Fig. 4c).

Fig. 5
figure 5

Effects of Ni2+ on acid-evoked Ca2+ transients. Three different levels of external acidosis were used (pHo 6.8, a; 6.2, b; 5.0, c). A control exposure to acidosis was followed by an exposure to acidosis in the presence of Ni2+ (2.5 mM) and finally by a second exposure to acidosis omitting Ni2+. a Ca2+ transients evoked by pHo 6.8 were unaffected by Ni2+. b Ca2+ transients evoked by pHo 6.2 were enhanced in the presence of Ni2+ and the post-Ni2+ (control) response to pHo 6.2 was also augmented when compared to the first control response. c Ca2+ transients in response to pHo 5.0 were unaffected by Ni2+ although the post-Ni2+ control response was again enhanced compared to the first control. d Summary of the effects of Ni2+ on Ca2+ transient response to acidosis. The data are presented as mean + SEM. Number for each experiment (n = 6). Data on the effects of Ca2+-free solutions upon acid-evoked Ca2+ transients are included for comparison. Statistical significance was assessed using Student’s paired t test (*p < 0.05; **p < 0.01; ***p < 0.001). Exposure periods in a–c are indicated by horizontal bars. Time scale bars in all recordings, 200 s

Full size image

Gadolinium (Gd3+; 0.5 mM) completely blocked Ca2+ transients evoked by pHo 6.8 in all neurones investigated (n = 5) and almost completely blocked the response to pHo 6.2 (only one neurone in five showed any response; Fig. 6). Gadolinium also suppressed the second control (recovery) response to pHo 6.8 but not that to pHo 6.2. Application of capsaicin (100 nM) at the end of each experiment and after a rest period of several minutes evoked robust Ca2+ transients confirming that any effects of Gd3+ upon TRPV1 channels were ultimately reversible (Fig. 6).

Fig. 6
figure 6

Effects of Gd3+ and Cd2+ on acidosis-evoked Ca2+ transients. a Gd3+ (0.5 mM) completely inhibited Ca2+ transients as a response to pHo 6.8. Ca2+ transients did not recover during a final exposure to pHo 6.8; however, the final application of capsaicin (100 nM), to characterise the neurones, still evoked a rise in [Ca2+] i . b Ca2+ transients evoked by pHo 6.2 were significantly reduced by Gd3+ (0.5 mM) below 10% compared to control transients. Following Gd3+ removal, the Ca2+ response to pHo 6.2 recovered fully. c In the presence of 100 μM Cd2+, acidosis evoked a strong increase of Fura-2 fluorescence which only partially recovered. d Summary of the effects of Gd3+ on Ca2+ transients evoked by acidosis (normalised to the first control transient, arbitrarily set as 100%). Values are presented as mean + SEM (n = 6 for each condition). Statistical significance was assessed using Student’s paired t test (**p < 0.01; ***p < 0.001). Exposure periods are indicated by horizontal bars. Time scale bars in all recordings, 200 s

Full size image

Application of acid stimuli in the presence of Cd2+ caused a massive rise in Fura-2 fluorescence that almost reached saturation (i.e. the maximal value for the fluorescence ratio encountered during Fura-2 calibration). Following removal of acidosis, the fluorescence ratio did not fully recover during the observation period (Fig. 6c). This observation is similar to findings seen during capsaicin application in the presence of Cd2+.

Amiloride (100 μM), an inhibitor of ASICs, did not significantly attenuate the Ca2+ transients evoked by any of the three acid stimuli tested, nor did it affect the postamiloride control response to acidosis (Fig. 7).

Fig. 7
figure 7

Sensitivity of acidosis-evoked Ca2+-influx to amiloride. a–c The effects of amiloride (100 μM) on acid-evoked Ca2+ transients was tested at three different pHo values (5.0, 6.2, 6.8). Note that the intrinsic fluorescence of amiloride causes only a very small baseline shift indicating negligible interference with Fura-2 fluorescence measurements. d Summary of the effects of amiloride on acid-evoked Ca2+ transients. Data are normalised to the initial control responses (arbitrarily set as 100%). Note that although there is a tendency for amiloride to reduce the effect of acidosis by 10–20%, this failed to reach statistical significance. The effects of Ca2+-free solutions on the acid-evoked Ca2+ transients are included for comparison. Statistical significance was assessed using Student’s paired t test (**p < 0.01; ***p < 0.001). Exposure periods are indicated by horizontal bars. Each experimental group includes six independent recordings. Time scale bars in all recordings, 200 s

Full size image

CPZ (10 μM), an antagonist of capsaicin-mediated activation of TRPV1 [81], strongly inhibited the Ca2+ response to capsaicin (0.1–1 μM) as expected (p < 0.01, n = 6; see Supplementary material 1). In contrast, the Ca2+ transients evoked by three different acid stimuli were not attenuated by CPZ. Instead, there was a tendency for the Ca2+ transients in response to acidosis to be amplified slightly by the application of CPZ; this was significant at pHo 6.2 (205 ± 60%, n = 6, p < 0.05 when compared to first control). Ca2+ transients evoked by the final, control, exposure to pHo 6.2 or 5.0 were also significantly larger than the first control Ca2+ transients (Fig. 8). Acid-evoked (pHo 6.2) Ca2+ transients were greatly reduced by BCTC (1 μM) to 22 ± 5% (n = 12, p < 0.01). AMG 9810 also substantially reduced initial Ca2+ transients to 15 ± 3% (n = 18, p < 0.001; Fig. 9b, d) when compared to first control. Ca2+ transients evoked in response to lactic acidosis (15 mM Na-lactate, pHo 6.2) were similarly reduced by both BCTC (to 15 ± 4%, n = 17, p < 0.001) and by AMG 9810 (to 11 ± 3%, n = 8, p < 0.01; Fig. 9c, d). For both BCTC and AMG 9810, the Ca2+-transient response to acidosis remained reduced compared to initial controls after washout of the drug (Fig. 9d). A similar lack of rapid/full reversibility was also noted in the ability of both BCTC and AMG 9810 to antagonise responses to capsaicin (see Supplementary material 1).

Fig. 8
figure 8

a–c Effects of CPZ on acidosis-evoked Ca2+ transients. Experiments were conducted using three different levels of acidosis pH 6.8 (a), pH 6.2 (b) and pH 5.0 (c). Note that in contrast to its effects on capsaicin-evoked Ca2+ transients, capsazepine failed to inhibit acid-evoked Ca2+ transients. d Mean data from experiments in a–c. Ca2+ transients were normalised to control Ca2+ transients (set to 100%). Note that instead of inhibiting the response to acidosis, CPZ amplified the response to pHo 6.2. The effects of removal of extracellular Ca2+ on the acid-evoked Ca2+ response are included for comparison. Exposure periods are indicated by horizontal bars. Number of each experimental condition: n = 6. Statistical significance was assessed using Student’s paired t test (*p < 0.05; **p < 0.01; ***p < 0.001). Time scale bars in all other recordings, 200 s

Full size image
Fig. 9
figure 9

Effects of BCTC and AMG 9810 (AMG) on acid (pHo 6.2) evoked Ca2+ transients. a, d BCTC (1 μM) reduced Ca2+ transients evoked by a mixed acidosis (pHo 6.2, 20% CO2) to 22%. The effects of BCTC were only partially reversible. b, d AMG 9810 (AMG, 1 μM) reduced mixed acidosis-evoked Ca2+ transients to 15% when compared to first control acidic stimulus. The effects of AMG were also only partially reversible. c, d BCTC and AMG 9810 significantly inhibited Ca2+ transients evoked by a lactic acidosis (pHo 6.2, 20% CO2 15 mM Na-lactate). Breaks in recordings (a–c) before the second control acidic stimulus were approximately 10 min (to try to facilitate reversal of effects of BCTC and AMG 9810). Time scale bars in all recordings, 200 s. Statistical significance was assessed using Student’s paired t test (**p < 0.01; ***p < 0.001, number of experiments are in parenthesis). rec. recovery, i.e. second postdrug control acid stimulus; lactate = 15 mM Na-lactate

Full size image

Acidosis triggers a sustained Ca2+-release from internal stores

The sustained rise in [Ca2+] i evoked by hypercapnic acidosis pHo 6.8 was 52 ± 6 nM (n = 17) and that evoked by a mixed acidosis pHo 6.2 was 73 ± 13 nM (n = 13). There was no statistically significant difference between the sustained rise in [Ca2+] i in response to these two stimuli (p = 0.263). The sustained rise in [Ca2+] i was not reduced by Ca2+-free solutions (pHo 6.8, 42 ± 5 nM (n = 10, p = 0.223) and pHo 6.2, 38 ± 7 nM (n = 7, p = 0.097)).

In further experiments, we investigated whether this sustained Ca 2+o -independent rise in [Ca2+] i was due to acid-induced Ca2+ liberation from intracellular Ca2+ stores. Since we observed no differences between the sustained rises in [Ca2+] i evoked either by pHo 6.8 or pHo 6.2, the following series of experiments were conducted using a single acid stimulus (hypercapnic acidosis pHo 6.8). Depletion of endoplasmic reticulum (ER) stores using CPA (10 μM, a sarco-/endoplasmic reticulum Ca2+ ATPase (SERCA) inhibitor) significantly reduced the sustained rise in [Ca2+] i to 24 ± 12 nM (n = 4, p < 0.05; Fig. 10a, d). Application of caffeine (30 mM) to deplete ER stores also reduced the sustained rise in [Ca2+] i to 11 ± 2.4 nM (n = 5, p < 0.01; Fig. 10b, d). Depletion of mitochondrial Ca2+ stores by FCCP (1 μM) did not influence the sustained rise in [Ca2+] i evoked by pHo 6.8 (55 ± 11 nM, n = 6, p = 0.235; Fig. 10c, d).

Fig. 10
figure 10

ER Ca2+ stores contribute to the acidosis evoked sustained rise in [Ca2+] i . a–c In Ca2+-free media acidosis (pHo 6.8) evoked a sustained rise in [Ca2+] i . a Effects of ER store depletion with CPA (10 μM) on the acid evoked rise in [Ca2+] i . b Effects of ER store depletion with caffeine (30 mM). c Effects of mitochondrial Ca2+-store depletion with the uncoupler FCCP (1 μM). d Summary of effects of CPA, caffeine and FCCP on the acidosis evoked rise in [Ca2+] i in Ca2+-free media. Statistical significance was assessed using Student’s paired t test (**p < 0.01; *p < 0.05 compared to acidic exposure in Ca2+-free solution). Exposure periods are indicated by horizontal bars. Time scale bars, 200 s

Full size image

Effects of anoxia on the [Ca2+] i response to acidosis

In vivo, the intense acidosis that accompanies ischaemia is primarily a consequence of the anaerobic generation of lactic acid due to a lack of oxygen delivery. Ischaemic acidosis is consequently invariably accompanied by anoxia. The following experiments were therefore conducted to investigate the effects of oxygen deprivation on the neuronal response to acidosis.

Exposure to anoxia plus hypercapnic acidosis (pHo 6.8) evoked Ca2+ transients with a peak Δ[Ca2+] i of 353 ± 214 nM (n = 8; Fig. 11b). These Ca2+ transients were significantly larger (p < 0.05, n = 8) than those evoked by pHo 6.8 alone (112 ± 60 nM). Anoxia plus a mixed acidosis (pHo 6.2) also evoked Ca2+ transients (498 ± 25 nM, n = 7) that were significantly (p < 0.05, n = 7) greater than those evoked by pHo 6.2 alone (310 ± 31 nM; Fig. 11a). Since the amplitude of the transient [Ca2+] i response to acidosis tended to vary widely between cells, these data were normalised to the Ca2+ transient evoked under normoxic conditions (set as 100%) and are presented in Fig. 11e. In addition to increasing Ca2+-transient amplitude, anoxia also appeared to slow the recovery (rate of decline) of the [Ca2+] i transient. Under control conditions, recovery rates (measured over the range 200–250 nM) were 3.7 ± 0.5 nM/s at pHo 6.8 and 4.3 ± 0.8 nM/s at pHo 6.2. In anoxia, these rates were slowed to 2.2 ± 0.5 nM/s (p < 0.05, n = 8) at pHo 6.8 and 1.7 ± 0.6 nM/s (n = 8, p < 0.01) at pHo 6.2. Ca2+ transients generated in response to anoxic–acidosis were either completely prevented or were significantly reduced in Ca2+-free medium (pHo 6.8, ∆[Ca2+] i  = 11 ± 4 nM, n = 6, p < 0.05; pHo 6.2, ∆[Ca2+] i  = 79 ± 14 nM, n = 5, p < 0.05; Fig. 11d, f).

Fig. 11
figure 11

Effects of anoxia on acid and capsaicin-evoked Ca2+ signalling. a, b Effects of anoxia on acid (pHo 6.8 and pHo 6.2) evoked Ca2+ responses. c Effects of anoxia on Ca2+ response to capsaicin (100 nM for 10 s). d Effects of removal of extracellular Ca2+ on Ca2+ response to acidic anoxia. e Summary of the effects of anoxia on the initial Ca2+ transient generated in response to acidosis and capsaicin. Data are normalised to the control response (in absence of anoxia) which is arbitrarily set to 100%. Values are means + SEM. f Summary of the effects of Ca2+-free media on the initial [Ca2+] i transient generated in response to acidic anoxia. Data are means + SEM. Statistical significance (in e and f) was assessed using Student’s paired t test (**p < 0.01; *p < 0.05). Numbers of experiments are given in parenthesis. Exposure periods (in a–d) are indicated by horizontal bars. Time scale bars, 200 s

Full size image

For comparison, we also investigated the effects of anoxia upon the Ca2+ transients evoked by capsaicin (100 nM for 10 s). Capsaicin-evoked Ca2+ transients were also amplified under anoxic conditions to 140 ± 6% of control (p < 0.01, n = 6; Fig. 11c, e). Moreover, the recovery rates of these Ca2+ transients were significantly slowed from 9 ± 1.6 nM/s under normoxia to 3.4 ± 0.8 nM/s during anoxic conditions (n = 6, p < 0.01).

The sustained rise in [Ca2+] i induced by anoxia and acidosis was also significantly greater than that evoked by acidosis alone. At pHo 6.8 in anoxia, the sustained rise in [Ca2+] i was 84 ± 9 nM (n = 13) vs. 52 ± 6 nM (n = 17) in pHo 6.8 alone (p < 0.01). Similarly at pHo 6.2 in anoxia, the sustained rise in [Ca2+] i was 104 ± 25 nM (n = 12) vs. 73 ± 13 nM (n = 13) in pHo 6.2 alone (p < 0.05).

Ca2+ responses to a combination of anoxia, acidosis and aglycaemia

In addition to anoxia, under ischaemic conditions, there may also be consumption of exogenous glucose. So, we also investigated the effects of combining acidosis and anoxia with aglycaemia. In the following experiments combining extracellular acidosis (pHo 6.8 or 6.2) with anoxia and aglycaemia (aaa 6.8 and aaa 6.2) resulted in a rapid transient increase in [Ca2+] i followed by a sustained rise in [Ca2+] i in the majority (65%) of neurones (see, e.g. Fig. 12a, e). In a small subpopulation of neurones (30%), however, the initial Ca2+ transient was absent or very small and only a sustained rise in [Ca2+] i was seen (e.g. Fig. 12b).

Fig. 12
figure 12

Effects of combining anoxia, aglycaemia and acidosis on [Ca2+] i . a–f Acidic stimuli (pHo 6.8 or pHo 6.2) together with anoxia and acidosis (aaa) were applied simultaneously and compared with aglycaemia (a, b), an equivalent acidosis (c, d) or anoxia (e, f) alone. In this series of experiments, the order of application of stimuli was randomised to exclude any systematic bias due to sensitisation or desensitisation. Time scale bars are 200 s. g, h Summary data showing comparison of effects of combined anoxia, aglycaemia and acidosis (g pHo 6.8, h pHo 6.2) with each individual stimulus in isolation on the sustained elevation of [Ca2+] i . Data are means + SEM. Statistical significance was assessed by Student’s paired t test (*p < 0.05; **p < 0.01); the number of experiments are given in parenthesis

Full size image

By comparison, the responses to anoxia alone or aglycaemia alone consisted only of a monophasic sustained increase in [Ca2+] i (see Fig. 12a, b, e, f). The sustained rise in [Ca2+] i evoked by the combination of anoxia, acidosis and aglycaemia was found to be larger than that evoked by any of these stimuli in isolation (see Fig. 12g, h; although the difference between pHo 6.8 and pHo 6.8 + anoxia and aglycaemia just failed to reach statistical significance). This suggests that the effects of these stimuli on sustained [Ca2+] i are additive (see “Discussion” section).

As noted above, Ca2+ transients were only observed under acid conditions and not with anoxia or aglycaemia alone. The amplitude of these acid-evoked [Ca2+] i transients were, however, augmented by both anoxia and the combination of anoxia and aglycaemia with the effects of anoxia and aglycaemia being larger than the effects of anoxia alone (see Fig. 13). We also noted that the recovery rates of Ca2+ transients evoked by aaa pHo 6.8 (1.9 ± 0.5 nM/s, n = 26) and aaa pHo 6.2 (2.5 ± 0.4 nM/s, n = 16) were again significantly reduced when compared to pHo 6.8 and pHo 6.2 alone (p < 0.05), but were not different to the recovery rates of Ca2+ transients evoked by anoxia pHo 6.2 or anoxia pHo 6.8.

Fig. 13
figure 13

Comparison of Ca2+-transient amplitude in response to acidosis, anoxic acidosis and anoxic aglycaemic acidosis. Studies were conducted at two levels of acidosis pHo 6.8 and pHo 6.2. Data are mean + SEM with n in parenthesis. Statistical comparisons, indicated by bars, were conducted using Mann–Whitney U test (*p < 0.05; **p < 0.01)

Full size image

Identification of Ca2+-influx pathways activated by aaa pHo 6.8

As previously observed for acidosis alone, the initial Ca2+ transient in response to aaa was greatly inhibited in Ca2+-free media whereas the sustained elevation in [Ca2+] i was independent of external Ca2+ (see Fig. 14). Having determined that the initial Ca2+ transients in response to the combination of acidosis, anoxia and aglycaemia is dependent upon Ca2+ influx, we next sought to identify the ion channels responsible for this Ca2+ influx. Neurones were exposed to aaa pHo 6.8 alone (control) and then to aaa 6.8 in the presence of amiloride, Ni2+, capsazepine, Gd3+ and then to aaa 6.8 again (second control). The amplitudes of the resulting Ca2+ transients were normalised to the first control response (arbitrarily set as 100%). Amiloride (100 μM) reduced Ca2+ transients to 50 ± 8% (n = 6, p < 0.01) compared to control (Fig. 15a, e). Ni2+ (2.5 mM) had no significant effect on Ca2+ transients (81 ± 21%, n = 5, p = 0.413; Fig. 15b, e). Capsazepine (10 μM) also had no significant effect upon the Ca2+ transient (86 ± 7% compared to control, n = 5, p = 0.1; Fig. 15c, e). In the presence of Gd3+ (1 mM), Ca2+ transients evoked by aaa pHo 6.8 were reduced to 45 ± 9% (n = 6, p < 0.05; Fig. 15d, e). BCTC (1 μM) and AMG 9810 also inhibited Ca2+ transients evoked by aaa pHo 6.2 (Fig. 16a, b, d). Ca2+ transients were reduced to 9 ± 6% (n = 6, p < 0.001) by BCTC and to 11 ± 8% (n = 6, p < 0.001) by AMG 9810, respectively. Ca2+ transients evoked by lactic acidosis in the presence of anoxia and aglycaemia (20% CO2 in 15 mM Na-lactate) were similarly reduced by AMG 9810 to 11 ± 5% (n = 6, p < 0.001; Fig. 16c, d). The effects of BCTC and AMG 9810 were again largely irreversible within the time frame of these experiments (Fig. 16a–c). Thus, the Ca2+ transient evoked by aaa was strongly inhibited by BCTC and AMG 9810 (although not completely abolished) was roughly halved by amiloride and Gd3+ but was unaffected by Ni2+ or capsazepine.

Fig. 14
figure 14

a, b Effects of removal of extracellular Ca2+ upon [Ca2+] i responses to combined acidosis (a pHo 6.8, b pHo 6.2) anoxia and aglycaemia. The initial Ca2+ transient in b was out of scale. c, d Summary of the effects of Ca2+ removal on both the [Ca2+] i transient and the sustained rise in [Ca2+] i in response to combined acidosis anoxia and aglycaemia. Statistical significance was assessed by Student’s paired t test (*p < 0.05; **p < 0.01). Numbers of experiments are given in parentheses. Exposure periods are indicated by horizontal bars. Time scale bars, 100 s

Full size image
Fig. 15
figure 15

Identification of calcium entry pathways activated in acidic (pHo 6.8) anoxic aglycaemia. a Effects of the ASIC inhibitor amiloride (100 μM). b Effects of the voltage-gated Ca2+-channel inhibitor Ni2+ (2.5 mM). c Effects of the capsaicin antagonist capsazepine (CPZ; 10 μM). d Effects of the trivalent cation Gd3+. e Summary of effects of amiloride, Ni2+ capsazepine and Gd3+ on Ca2+-transient amplitude in response to acidic anoxic aglycaemia. Data are normalised to the control transient recorded in the absence of drug (set to 100%). Statistical significance was assessed using Student’s paired t test (*p < 0.05; **p < 0.01; ***p < 0.001). Exposure periods are indicated by horizontal bars. Time scale bars are 200 s

Full size image
Fig. 16
figure 16

Effects of the TRPV1 inhibitors BCTC (1 μM) or AMG 9810 (1 μM, AMG) on Ca2+ transients evoked by acidosis, anoxia and aglycaemia (aaa). a, b Mixed acidosis (20% CO2 in pH 6.2). c Mixed acidosis with lactate (20% CO2, 15 mM Na-lactate, pH 6.2). d Comparison of aaa evoked Ca2+-transient amplitudes in the presence of BCTC and AMG 9810. Ca2+-transient amplitude is expressed as a percentage of the first control response to aaa. Note that both inhibitors reduced initial Ca2+ transients significantly (for both types of acidosis) but did not completely abolish them. Interruptions in the recordings between the application of the inhibitor and the second control were approximately 10 min (to allow for washout of AMG 9810 or BCTC). Despite this wash period Ca2+ transients did not fully recover following removal of either BCTC or AMG 9810. Data are means + SEM; statistical significance was assessed using Student’s paired t test (***p < 0.001). Exposure periods are indicated by horizontal bars. Time scale bars are 200 s

Full size image

Discussion

In this study, we have investigated the effects of acidosis, anoxia and aglycaemia on [Ca2+] i in sensory neurones. Acidosis is considered to be one of the primary stimuli responsible for sensory neurone excitation in ischaemia [68]. It also plays a pivotal role in triggering neuronal release of CGRP [25, 73, 113]. The secretion CGRP in response to a number of stimuli, including acidosis, is critically dependent upon Ca2+ signalling [23, 28, 56].

In the first part of this study, we investigated the effects of acidosis alone on intracellular [Ca2+] i in small capsaicin-sensitive neurones arising from dorsal root ganglia as these are thought to be responsible for sensing ischaemic conditions [17, 68, 88]. In the second part of the study, we have compounded the effects of acidosis with anoxia and aglycaemia since these conditions not only accompany ischaemia but can have profound effects upon Ca2+ regulation in sensory neurones [34, 55, 69, 75].

Effects of acidosis alone on [Ca2+] i in sensory neurones

We have employed four different levels and types of acidosis in this study: a relatively mild hypercapnic acidosis (pH 6.8; 20% CO2 and normal bicarbonate) representative of the early stages of ischaemia in the heart, a more severe mixed acidosis (pH 6.2; 20% CO2 and reduced bicarbonate) and an equivalent mixed acidosis including 15 mM Na-lactate representative of the latter stages of ischaemia and a very severe acidosis (pHo 5.0) representative of the stimuli frequently used to activate acid sensitive cation channels in vitro [10, 52, 110]. All four acid stimuli evoked a biphasic elevation of [Ca2+] i in the majority of neurones investigated. The initial phase was characterised by a Ca2+ transient occurring within the first minute of exposure to acidosis. This initial transient was chiefly dependent on external Ca2+. The Ca2+ transient was then followed by a sustained elevation of [Ca2+] i . This second sustained phase of the [Ca2+] i response was not depended upon Ca2+ influx. We will first address possible mechanisms for these Ca2+ signalling events.

Pharmacological identification of Ca2+ influx pathways

The most likely mechanisms for acid-evoked Ca2+ influx in neurones are through (a) ASIC1a channels [101, 102, 110], (b) TRPV1 (or capsaicin) receptors [17, 89] and (c) voltage-gated Ca2+ channels secondary to acid-evoked membrane depolarisation [15, 40, 87]. In order to discriminate between these entry pathways, we needed to find suitable pharmacological tools.

ASIC channels are widely reported to be inhibited by amiloride [106] although this is not entirely selective as amiloride may also inhibit T-type voltage-gated Ca2+ channels [86, 95]. Recent evidence also indicates that under moderate acidosis (pH 7.0), amiloride may activate ASIC3 leading to sustained inward currents [108].

Although there are several different types of voltage-gated Ca2+ channels in sensory neurones [3, 4], there are a number of divalent cations that can be used to inhibit all high voltage-activated channels (cadmium) and both high and low voltage-activated channels (nickel) [9, 13, 62, 64, 71, 74, 95]. Using high potassium solutions to depolarise, we confirmed that 100 μM Cd2+ substantially blocked and 2.5 mM Ni2+ fully blocked voltage-gated Ca2+ entry. We next tested the effects of these cations on capsaicin evoked Ca2+ influx to see if they could be used to discriminate between TRPV1 and voltage-gated Ca2+ entry. Capsaicin-induced Ca2+ transients were not inhibited by either Ni2+ or Cd2+; indeed, the effects of capsaicin appeared to be potentiated. In particular, exposure to capsaicin in the presence of Cd2+ led to a substantial and mostly irreversible increase in the Fura-2 fluorescence ratio. We suggest that this is probably due to Cd2+ entry through TRPV1 and subsequent binding of Cd2+ to Fura-2. Cd2+ binding to Fura-2 has similar effects on its fluorescence to those of Ca2+ binding but Fura-2 has a much greater affinity for Cd2+ than for Ca2+ [36]. Ni2+ ions can also bind to Fura-2 but when they do so, they quench fluorescence. No fluorescence quenching was observed during exposure to capsaicin in the presence of Ni2+ suggesting that Ni2+ does not pass through TRPV1 to any significant extent. Ni2+ was also ineffective in inhibiting capsaicin-evoked Mn2+ entry (TRPV1 channels are permeable to Mn2+ [66]), whereas it fully blocked voltage-gated Mn2+ entry. Mn2+ quenching of Fura-2 fluorescence in the presence of Ni2+ may therefore also serve as an indicator of TRPV1 activation. One possible problem with the use of Ni2+ is that it appeared to slightly potentiate the effects of capsaicin. Sensitisation of TRPV1 channels by other external cations, Na+ and Mg2+, has been previously described [1]. Despite this sensitisation, the effects of Ni2+ and Cd2+ should permit discrimination between voltage-gated and TRPV1-mediated Ca2+ influx.

The TRPV1 antagonist capsazepine was, as expected, highly efficient in blocking the effects of capsaicin [31, 51, 53]. Whether it can also block acid-evoked activation of TRPV1 is, however, uncertain (see below). We therefore sought other blockers of Ca2+ entry through TRPV1. Gadolinium (Gd3+) is a small lanthanide that blocks various types of calcium channels at submillimolar concentrations [7, 14, 49, 62]. Gd3+ has been reported to activate and sensitise TRPV1 channels at low concentrations, but at higher concentrations (>300 μM), it inhibits capsaicin-evoked membrane currents [1, 92]. In the present study, Gd3+ inhibited both voltage-activated Ca2+ entry and capsaicin-evoked Ca2+ influx, but we could only inhibit capsaicin-evoked Ca2+ transients with Gd3+ concentrations of 500 μM and above. Two newer TRPV1 antagonists BCTC and AMG 9810 with reported ability to block proton activation of TRPV1 were also tested [27, 93]. These produced a profound block of capsaicin-evoked Ca2+ transients that was largely irreversible.

Acid-evoked Ca2+ influx pathways

Protons activate strong inward sodium currents in DRG neurones which have both transient and sustained components [10]. These currents are mediated mainly by ASIC3, a proton-sensitive cation channel [29, 57, 67, 77]. ASICs are thought to play a key role in transmitting ischaemic pain. They are directly activated by extracellular acidosis which relieves channel blockade by Ca2+ ions through the protonation of the Ca2+-binding sides [41, 42, 78]. The ion selectivity of these channels is not restricted to Na+; ASIC1a is also permeable to Ca2+ [19, 99, 105, 110] and contributes to the Ca2+ overload in neurones of ischaemic brain [107]. In the present study, amiloride was administered to evaluate the contribution of ASICs to the acid-evoked Ca2+ transients [10, 18, 99]. Surprisingly although there was a general tendency for amiloride to reduce the acid-evoked Ca2+ transient at all pH values tested, this effect did not reach statistical significance and robust Ca2+ transients invariably remained. Thus, although we cannot completely exclude a minor role for ASICs in mediating some of the acid-evoked Ca2+ influx, they are clearly not the main cause.

We also tested the effects of Ni2+ and Cd2+ on the Ca2+ response to acidosis. Both of these agents failed to inhibit the acid-evoked Ca2+ transient. Ni2+ also failed to inhibit acid-evoked Mn2+ influx (see Fig. 4c). These data indicate that the Ca2+ transient is not primarily a consequence of membrane depolarisation followed by voltage-gated Ca2+ entry and point instead to another acid-activated cation channel. The fact that acid evoked a sustained rise in Fura-2 fluorescence ratio in the presence of Cd2+ suggests that this channel is also permeable to Cd2+. Acid-evoked Ca2+ transients were, however, completely blocked by 500 μM Gd3+ and blocked by 85–90% by BCTC and AMG 9810. These characteristics, resistance/sensitisation by Ni2+, permeability to Mn2+ and Cd2+, block by high levels of Gd3+ and block by BCTC and AMG 9810 are identical to those of TRPV1 as described above. We also noted a positive correlation between the amplitudes of acidosis- and capsaicin-evoked Ca2+ transients (see Fig. 2).

Although the above data point to TRPV1 as being the main pathway for Ca2+ entry, it was notable that capsazepine failed to inhibit the acid-evoked Ca2+ transients. There are conflicting reports in the literature regarding the capsazepine sensitivity of both proton-activated inward currents in rat sensory neurones and the cloned rat TRPV1 channel. Tominaga and Tominaga originally reported that 10 μM capsazepine inhibited proton activated currents through cloned rat TRPV1 channels by approximately 80% [89], but others have subsequently found no effect at concentrations up to 30 μM [61, 65]. Similarly in rat sensory neurones, there are some reports that the sustained proton-activated currents are inhibited by capsazepine [52] and others which show no effect of capsazepine on proton-evoked currents or ion fluxes [11, 98]. Absence of effects of capsazepine upon proton activation of TRPV1 is not surprising since the capsaicin/capsazepine binding site and proton binding site are at different locations (capsaicin binds an intracellular domain [45] whereas protons bind to an extracellular domain [43]). The lack of consistency with respect to the effects of capsazepine on proton-activated TRPV1 currents in rat is, however, puzzling. Nevertheless, as there is clear precedent for lack of effect of capsazepine on proton activation of rat TRPV1 and given the strong inhibitory effects of two other TRPV1 antagonists, the failure of capsazepine to antagonise the proton-activated Ca2+ influx cannot be considered to exclude a role for TRPV1.

In summary, the primary route for acid-evoked Ca2+ influx is via TRPV1. We cannot entirely exclude some contributory role for parallel voltage-gated Ca2+ entry (secondary to membrane depolarisation) since possible sensitisation of TRPV1 by Ni2+ may obscure any minor contribution from voltage-gated Ca2+ channels. The lack of any apparent major role for ASIC channels may seem surprising, but it should be noted that many of these channels inactivate very rapidly [10, 42, 78] so any Ca2+ influx will be short-lived and could easily be obscured by the subsequent Ca2+ influx through TRPV1.

Ca2+ liberation from internal stores during external acidosis

During exposure to acidosis, there is also a rise in [Ca2+] i that is independent from Ca 2+o . The sustained nature of this [Ca2+] i rise suggests a change in the equilibrium between Ca2+ fluxes into and out of the cytosol (note that a discrete Ca2+-release event alone, for example, displacement of Ca2+ from internal buffers by H+, would be expected to only result in a transient increase in [Ca2+] i as the Ca2+ released is subsequently extruded). We have not investigated the cause of this sustained rise in [Ca2+] i in any detail, but we have observed that ER Ca2+-store depletion using either caffeine or CPA reduces this acid induced rise in [Ca2+] i . One possible explanation therefore is that intracellular acidosis causes a sustained increase in Ca2+ leak from the ER.

Whilst a number of studies indicate that a direct effect of intracellular acidosis is to inhibit the ryanodine receptor [6, 84], there is another indirect pathway whereby Ca2+ release from internal stores may be triggered by the activation of proton-sensing G protein-coupled receptors [39, 54, 90]. These receptors have recently been described in approximately 75% of small nociceptive DRG neurones in which they are colocalised with ASICs [38]. In addition to the possibility of enhanced Ca2+ release, acidosis may also reduce Ca2+ reuptake into the ER by inhibiting SERCA. Moreover, our observation that a small (acid evoked) sustained increase in [Ca2+] i also occurs in the presence of CPA (in Ca2+-free media) suggests that some inhibition of plasma membrane Ca2+ extrusion is also likely. Determining which of these pathways are indeed responsible for the sustained Ca2+ release from internal stores seen acidic conditions clearly requires further investigation.

Effects of combining acidic stimuli with anoxia and aglycaemia

When acidosis was combined with anoxia, or with both anoxia and aglycaemia, there was a systematic increase in the amplitude of the Ca2+ transient (Fig. 13). The effects of acidosis are therefore potentiated by loss of oxygen and loss of glucose both of which are conditions present in ischaemia. There are likely to be a number of factors contributing to this effect. In analysing the cause of the Ca2+ transient in anoxia, aglycaemia and acidosis, we found that this transient was exclusively dependent on extracellular Ca2+ and was strongly inhibited by BCTC and AMG 9810, but was not inhibited by voltage-gated Ca2+-channel blockers (Ni2+). The major source of Ca2+ influx would therefore again seem to be mainly via TRPV1. We did, however, note that amiloride also reduced the Ca2+ transient under these conditions. A role for ASICs would therefore seem to be more prominent in response to anoxic–aglycaemic–acidosis than to acidosis alone. One possible explanation for this is that oxygen and glucose deprivation has recently been reported to both increase the amplitude and slow the inactivation of ASIC1a currents in mouse hippocampal neurones [26]. Thus, the amount of Ca2+ influx through ASIC1a may be enhanced. Further factors which we anticipate will also contribute to the enhancement of the Ca2+ transient by anoxia and aglycaemia include rather more general changes to Ca2+ metabolism including inhibition of plasma membrane Ca2+ extrusion and inhibition of SR Ca2+ uptake [34]. Moderate inhibition of both SERCA and plasma membrane Ca2+ ATPase (PMCA) pumps occurs fairly rapidly in anoxia and corresponds to decline in cytosolic ATP levels [34, 35]. These factors probably account for the slower recovery of the Ca2+ transient observed in anoxic–aglycaemic–acidosis compared to that seen in acidosis alone. Indeed, we also noted that anoxia both enhanced the [Ca2+] i response to capsaicin and prolonged [Ca2+] i recovery following capsaicin removal (Fig. 11c). Given the prominent role for mitochondria in Ca2+ buffering in these cells [2, 63, 79, 83, 103], one might also anticipate that reduced mitochondrial Ca2+ uptake should contribute to the enhanced Ca2+ transient. In previous studies, however, we have found that although relatively brief anoxia does cause partial mitochondrial depolarisation, this does not appear to affect the capacity of mitochondria to take up Ca2+ [34, 35]. In summary, the causes of the enhancement of the acid evoked Ca2+ transient in anoxia and aglycaemia are likely to be complex, involving the modulation of both Ca2+-influx and Ca2+-efflux pathways and are in need of further investigation.

Even in the presence of anoxia and aglycaemia, it is notable that the Ca2+-influx phase of the response to acidosis is still relatively short-lived (t 1/2 = 43 s; Fig. 11). There are a number of possible explanations for this: Firstly, intrinsic inactivation leads to rapid (seconds) decline in ASIC activity [42, 78], secondly, elevated intracellular calcium leads to inactivation of TRPV1, and finally, declining ATP levels may contribute to loss of TRPV1 activity [91, 94, 112]. What remains after the Ca2+ transient has subsided is a sustained elevation of [Ca2+] i which is largely independent of Ca2+ influx. Aglycaemia and anoxia alone can also induce a sustained increase in [Ca2+] i . The response to the combination of all three ‘stimuli’ is greater than that observed in response to any single stimulus given in isolation (see Fig. 12). Thus, the effects of anoxia, aglycaemia and acidosis on intracellular Ca2+ regulation appear to summate. The cause of the rise in [Ca2+] i in response to anoxia has been described previously and includes inhibition of both SERCA and PMCA as well as enhanced leak of Ca2+ from internal (ER) Ca2+ stores [34]. The cause of the sustained rise in response to acidosis has not yet been investigated in full, but our preliminary data indicate that it is likely to involve changes to ER calcium handling and hint at a possible role for the recently discovered G protein-coupled proton receptors.

Conclusions

In this study, we have only investigated events likely to occur early in ischaemia; more prolonged ischaemic conditions are predicted to further disturb Ca2+ homeostasis as endogenous glycolytic fuel reserves are depleted, ATP levels collapse and Ca2+-extrusion mechanisms fail completely [34, 35]. What is interesting to note, however, is that even before this stage is reached, anoxia and aglycaemia have a marked influence over cellular Ca2+ homeostasis enhancing both the calcium influx and the sustained rise in [Ca2+] i in response to acidosis. Even under anoxic and aglycaemic conditions, however, the acid evoked Ca2+ influx, mediated predominantly via TRPV1, remains relatively short-lived.

Important questions that remain to be addressed are how might the Ca2+-signalling events described here be further influenced by other mediators released in ischaemia (e.g. bradykinin and adenosine), could these support, enhance or prolong the Ca2+-influx phase or is Ca2+ influx only relevant in the very early stages of ischaemia? If the latter, what are the functional consequences of the sustained (Ca2+-influx independent) rise in [Ca2+] i in these cells, can it promote continued release of CGRP or is this only linked to the Ca2+-influx phase? Finding answers to these questions is important if we are to gain a better understanding of the role played by acid-sensing cation channels, calcium stores and calcium signalling in general in sensory neurone function under ischaemic conditions.