Secretory vesicle rebound hyperacidification and increased quantal size due to prolonged methamphetamine exposure

Abstract

Acute exposure to amphetamines collapses secretory vesicle pH gradients, which increases cytosolic catecholamine levels while decreases the quantal size of catecholamine release during fusion events. Amphetamine and methamphetamine, however, are retained in tissues over long durations. We used optical and electron microscopic probes to measure the effects of long-term methamphetamine exposure on secretory vesicle pH, and amperometry and intracellular patch electrochemistry to observe the effects on neurosecretion and cytosolic catecholamines in cultured rat chromaffin cells. In contrast to acute methamphetamine effects, exposure to the drug for 6–48 h at 10 μM and higher concentrations produced a concentration-dependent rebound hyperacidification of secretory vesicles. At 5–10 μM levels, methamphetamine increased the quantal size and reinstated exocytotic catecholamine release, although very high (>100 μM) levels of the drug, while continuing to produce rebound hyperacidification, did not increase quantal size. Secretory vesicle rebound hyperacidification was temperature dependent with optimal response at ~ 37°C, was not blocked by the transcription inhibitor, puromycin, and appears to be a general compensatory response to prolonged exposure with membranophilic weak bases, including amphetamines, methylphenidate, cocaine, and ammonia. Thus, under some conditions of prolonged exposure, amphetamines and other weak bases can enhance, rather than deplete, the vesicular release of catecholamines via a compensatory response resulting in vesicle acidification.

INTRODUCTION

Amphetamine (AMPH) and methamphetamine (METH) have been prescribed to over 19 million Americans (NSDUH 2003 report) for the treatment of attention deficit hyperactivity disorder (ADHD) and other affective conditions, and are illicitly self-administered by over 25 million people worldwide (United Nations Office on Drugs and Crime 2007 report), either at very high doses by chronic drug abusers (Gettig et al. 2006), or more often at lower doses in order to alleviate fatigue and promote wakefulness. In abuse and some therapeutic settings, amphetamines are consumed several times a day, and are eliminated quite slowly with a half-life of 10 – 24 h (see Methods). Few studies, however, have examined the long-term effects of amphetamines on neurosecretion.

Acute exposure to amphetamines rapidly collapses the acidic pH gradient in secretory vesicles, redistributes stored transmitter from vesicles to the cytosol, and decreases the quantal size of stimulation-dependent exocytosis (Johnson, 1988; Sulzer and Rayport, 1990; Sulzer et al., 1992b; Jones et al., 1998; Mundorf et al., 1999; Mosharov et al., 2003). Additionally, amphetamines release catecholamines in a stimulation-independent manner by inducing reverse transport from the cytosol to the extracellular milieu via actions on plasma membrane uptake transporters (Fischer and Cho, 1979; Sulzer et al., 1995b; Sonders et al., 1997; Jones et al., 1998; Kahlig et al., 2005). Together with additional properties of these drugs, including an inhibition of DA catabolism and an activation of its synthesis (Sulzer et al. 2005), the net effect of acute METH or AMPH is to reduce stimulation-evoked catecholaminergic neurotransmission and induce stimulation-independent, non-exocytotic transmitter release.

Here we compare the impact of acute and prolonged METH exposure on evoked adrenaline release by examining secretory vesicle pH, cytosolic catecholamine levels, and quantal secretory vesicle transmission in adrenal chromaffin cells. We show that, in contrast to acute METH effects that inhibit evoked catecholamine secretion, long-term drug treatment elicits a delayed reacidification of secretory vesicles, resulting in increased quantal size at pharmacologically relevant METH concentrations. The data thus indicate that prolonged amphetamine exposure promotes an adaptation that we term rebound hyperacidification which reverses the decrease in quantal size, and augments rather than inhibits stimulation-dependent secretory catecholaminergic neurotransmission.

MATERIALS AND METHODS

Chemicals and Cell Culture

Chemicals were purchased from Sigma-Aldrich (St. Louis, MO) unless otherwise stated. Animal protocols were approved by the Columbia University Institutional Animal Care and Use Committees. Rat-derived postnatal chromaffin cells were isolated and cultured in a 5% CO₂ incubator at 37°C as previously described (Mosharov et al. 2003). All experiments were conducted between days 3–4 post-plating. Stock solutions of METH (100 mM) and other weak bases were added directly to the culture medium for the time indicated and washed out immediately prior to the measurements. For electron microscopy, the dishes were prepared with Aclar coverslips but were otherwise treated identically.

Amperometric recordings

Solutions used for amperometric recordings were as follows. The bath saline (pH 7.4) contained (in mM): 128 NaCl, 2 KCl, 1 NaH₂PO₄, 2 MgCl₂, 1.2 CaCl₂, 10 glucose, 10 HEPES-KOH. Secretagogue solution was the same except 90 NaCl and 40 KCl. Secretagogue was applied by local perfusion through a pressurized glass micropipette (Picospritzer, General Valve Co., Fairfield, NJ) for 5 sec at ~10 μm from the cell. A 5 μm diameter carbon fiber electrode held at +700 mV was pressed against the cell surface and catecholamine oxidation was monitored as amperometric current spikes. The current was filtered using a 4-pole 5 kHz Bessel filter built into an Axopatch 200B amplifier (Axon Instruments, Foster City, CA) and sampled at 25 kHz (ITC-18, Instrutech, Great Neck, NY). After secretagogue application, the amperometric current was recorded for 60 sec and the data were analyzed using a locally written routine (see (Mosharov and Sulzer 2005) for details) in IGOR Pro (WaveMetrics, Lake Oswego, OR). The current was digitally filtered using a low-pass binomial 600 Hz filter and the root mean square of the noise (RMS noise) on the first derivative of the current (dI/dt) was measured in a segment of the trace that did not contain spikes. dI/dt was then used to detect amperometric events that were 4.5-fold larger than the RMS noise. To improve the quality of spike detection, the current was additionally filtered using low-pass binomial 150 Hz filter before taking dI/dt (Mosharov and Sulzer 2005). For each amperometric spike, the following shape characteristics were determined: quantal size (Q, pC), amplitude (I_max, pA), duration at half-height (t_1/2, ms), rise-time between 25% and 75% of I_max excluding the pre-spike foot (t_rise, ms), and the incline of the rising phase (slope, pA/ms). The falling phase of a spike was characterized by two time constants (τ₁ and τ₂, ms) of the double-exponential fit of the current between 25% of I_max and spike's end. The number of catecholamine molecules released from individual vesicles (Q_N) was calculated as Q_N = Q/(n·F), where F = 96,485 C/mole is Faraday's constant and n = 2 is the number of electrons donated by each catecholamine molecule (Bruns and Jahn 1995). Spikes with I_max smaller than 3 pA and traces with fewer than 10 amperometric spikes were excluded from the analysis.

Intracellular Patch Electrochemistry (IPE)

Measurements of cytosolic catecholamine concentrations were performed as previously described (Mosharov et al. 2003). Briefly, the concentration of cytosolic catecholamine was determined using a cyclic voltammetric mode of electrochemical detection; 300 mV/ms ramps of voltage were applied over 100 ms intervals to the carbon fibre electrode positioned inside the patch pipette. After achieving the whole cell configuration, cytosolic catecholamines were monitored as an oxidation wave and their concentration was calculated using the calibration curves. The initial cytosolic catecholamine concentration was estimated using images of the cell and the patch pipette to account for molecule dilution within the pipette tip.

When a CFE is employed in cyclic voltammetric (CV) mode of detection, a unique red-ox profile (voltammogram) provides a means to distinguish catechols from other oxidazable cytosolic metabolites. The electrode in CV mode is 15–20-fold more sensitive to catecholamines (dopamine, norepinephrine and epinephrine) than to other catechol species present within the cytosol, thus allowing one to discriminate against the presence of L-DOPA, DOPAC and DOPEG (Mosharov et al. 2003).

Measurements of secretory vesicle acidity using acridine orange (AO)

At each time point of incubation with METH, chromaffin cells were rinsed once with drug-free Dulbecco's PBS pH 7.4 and treated with 10 nM AO for 10 minutes at 37°C. Cultures were rinsed with 3 ml PBS twice and series of images were taken from randomly chosen cells over the following 5 minutes using 480 nm excitation and 535 nm emission filter set for FITC/AO (Chroma Technology Co., Rockingham, VT). Images were acquired with a cooled CCD Star I camera (Photometrics, Tuscon, AZ) operated by IPLab software v. 3.1.1. (BD Biosciences Bioimaging, Rockville, MD). The pixel intensities of all images were equally normalized in 8-bit (BYTE) monochrome image range (0–255). The pixel intensity value was averaged from the total cell area excluding the nucleus using `Segmentation' and `Quantify Segments' modes in IPLab software. Average pixel intensity of the background staining was quantified from the cell-free area of the image and subtracted from the average value of cell-specific AO staining. For each time point the intensity of AO staining is shown as the average of the means from at least 30 individual cells ± SEM.

Ratiometric pH measurements with LysoSensor Yellow/Blue DND-160 (LSYB) in situ

Cells were incubated with 5 μM LSYB (Invitrogen Co., Carlsbad, CA) in CO₂-conditioned culture medium for 5 minutes at 37°C, washed with dye-free PBS and imaged on inverted fluorescent microscope using 40^X Plan-Neofluar oil objective (Carl Zeiss MicroImaging Inc.) and two customized filter sets. Both filter sets shared the same 365 nm excitation filter that corresponded to the isobestic point of LSYB excitation spectrum. The emission filters were 445 nm (“blue”) and 510 nm (“yellow”), corresponding to the maxima of pH-sensitive peaks of LSYB emission spectrum (Diwu et al. 1999). Images were acquired using Star I camera operated by IPLab software within 10 min following LSYB addition. Following background subtraction and accounting for bleaching, the mean pixel values from the cell body excluding the nucleus were calculated as described above for AO. Yellow to blue ratio of fluorescence intensities was further used to calculate vesicular pH employing the exponential fit parameters from the calibration curve (Fig. 2c). To calibrate LSYB emission ratios for a range of pH values in situ, vesicles were saturated with LSYB as described, and then the extracellular media for LSYB-filled cells was replaced with a buffer of known pH in the presence of monensin, a Na+/H+ antiporter, and nigericin, a K+/ H+ antiporter, which act to bring the cytosol and chromaffin vesicles to the pH of the surrounding media (Holopainen et al. 2001). Cells were incubated with 5 μM LSYB in CO₂-conditioned culture medium for 20 minutes at 37°C and washed with dye-free PBS at RT. Next, the saline was exchanged to a buffer of known pH (values from 3.5 to 6.5), 10 μM monensin and 10 μM nigericin were applied, and images were acquired within the following 15 min and processed as above (Supplementary Fig. 1). Although the leakage of the dye from vesicles to the cytosol cannot be avoided during the calibration procedure, this was unlikely to affect the resulting LSYB calibration curve, because the pH in the vesicles and the cytosol was equilibrated to that of the extracellular buffer. Yellow/Blue values from triplicate groups of 10–20 individual cells were averaged and used to build the calibration curve. Ratios ≥ 3 standard deviations from the mean were excluded from analysis. To follow the rapid changes in vesicular pH induced by an exposure to weak bases, the cells were pre-incubated with 5 μM LSYB for 5 min, the dye was washed out with PBS, and immediately after the addition of a weak base, Yellow and Blue images were acquired from 10 randomly chosen cells. The ratios from 10 dishes were averaged and pH values calculated from the calibration curve as described.

Fig. 2.

In situ measurement of vesicular pH in chromaffin secretory vesicles using LSYB. (a) Maximal projection two-photon images of LSYB-stained chromaffin cell. Two emission channels 445 nm (Blue) and 510 nm (Yellow) were pseudo-colored as green and red, correspondingly. Note the punctate LSYB staining with negligible signal from the cytosol. Scale bar is 10 μm. (b) Examples of LSYB epifluorescence images taken at 510 nm (Y) and 445 nm (B) from untreated cells and cells exposed to 100 μM METH for 4 min or 24 h; λ_ex = 365 nm. Scale bar is 10 μm. (c) Calibration curve for ratiometric measurement of vesicular pH. (d) Time course of vesicular pH changes in chromaffin cells acutely treated with 100 μM METH, 100 μM chloroquine and 10 mM ammonium chloride. N = 10 cells at each experimental group. (e) Vesicular proton concentration and pH in cells treated with different METH concentrations for 24 h. All treatment groups were significantly different from untreated cells (p < 0.005 by one-way ANOVA). Dotted line and shadowed box represent the mean ± SEM in untreated cells.

Two-photon images of LSYB-stained chromaffin cells were acquired on Leica DM6000B microscope equipped with a 63×0.9 water-immersion objective. Excitation wavelength was at 750 nm and two emission channels had bandwidths of 400–450nm and 520–600 nm.

Staining with monodancylcadaverine (MDC)

Cells were labeled with MDC as described (Larsen et al., 2002). Briefly, cells were incubated with 50 μM MDC in CO₂ - conditioned culture medium for 20 minutes at 37°C. After extensive wash with drug-free conditioned media, cells were incubated in CO₂ incubator for 10 minutes, washed twice with PBS and imaged on inverted fluorescent microscope using 40^X Plan-Neofluar oil objective. Live cell images were taken with a customized filter set with 335 nm excitation and 525 nm emission filters (Chroma Technology Co., Rockingham, VT). Pixel intensity values were compared as described for AO.

Immuno-gold electron microscopy of DAMP in chromaffin vesicles

3-(2,4 dinitroanilino)-3'-amino-N-methyl dipropylamine (DAMP) is a weak base that is trapped in acidic compartments and can be detected in tissues with antibodies to dinitrophenol (DNP) (Anderson et al. 1984). DAMP immunoreactivity was detected in chromaffin cells as previously described (Pothos et al. 2002). Briefly, the cultures were treated with 100 μM METH from 1 h or 24 h before adding 50 μM DAMP (Molecular Probes, Eugene, OR, USA) for 30 min. Following a 5 min rinse in PBS, cells were fixed with 4% paraformaldehyde, 0.1% glutaraldehyde, 3% sucrose in 0.1 M phosphate buffer, pH 7.4 for 1 h at room temperature. The fixed cells were washed twice for 5 min in PBS and mordanted with 0.25% tannic acid in 100 mM sodium phosphate buffer containing 3.5% sucrose for 1 h at 4°C. Specimens were then quenched with 50 mM NH₄Cl, washed with 100 mM sodium maleate containing 4% sucrose (pH 6.2) and stained en bloc with maleate-buffered 2% uranyl acetate. The stained cells were dehydrated to 70% ethanol at 4°C; the temperature was then lowered to −20°C for dehydration to 90% ethanol, clearing and embedding in LR Gold Resin (London Resin Co). The LR Gold was polymerized with ultraviolet light at −20°C. Thin sections were picked up on Ni grids that had previously been layered with Formvar (Electron Microscopy Sciences, Hatfield, PA). The sections on the grids were blocked with 10% donkey serum with TBST (10 mM Tris [pH 7.2], 500 mM NaCl, 0.05% Tween 20) for 30 min at 20°C. Sections were incubated overnight with rabbit polyclonal antibodies to DNP (1:200; Sigma-Aldrich). Grids were subsequently washed in blocking solution and incubated for 2 h at room temperature with donkey anti-rabbit antibodies coupled to 18 nm particles of colloidal gold (1:40; Jackson ImmunoResearch, West Grove, PA). The sections were washed in PBS, postfixed with 2.5% glutaraldehyde and stained with aqueous 2% osmium, uranyl acetate and lead citrate. Sections were examined with a JEOL 1200 EX electron microscope.

Estimation of vesicular pH gradients in DAMP-containing organelles

The internal pH of cellular organelles at the ultrastructural level was measured by determining the partition of DAMP, as previously published (Anderson et al. 1984; Orci et al. 1994). For each micrograph, the density of immunogold particles was measured in an area outside the cell (D_extracell), the nucleus (D_nucleus), a lysosome (D_lysosome), and in twenty randomly chosen LDCVs (D_vesicle = total area / total # of gold particles). Intravesicular and lysosomal pHs were then estimated as:

$$\mathit{pH}=7.0-\text{log}\left(\frac{{\mathrm{D}}_{\text{vesicle or lysosome}}-{\mathrm{D}}_{\text{extracell}}}{{\mathrm{D}}_{\text{nucleus}}-{\mathrm{D}}_{\text{extracell}}}\right)$$

and expressed as an average of the mean pH values from 10–15 cells in each group (Fig. 4b). Alternatively, intravesicular pH of individual vesicles was calculated using the same formula and distributions of vesicle pH for each group was plotted as cumulative histograms in Excel (Fig. 4c).

Fig. 4.

Immunoelectron microscopy of chromaffin granules in METH-treated cells. (a) Representative micrographs of DAMP-stained untreated cells and cells treated with 100 μM METH for 1 h or 24 h. The scale bar is 500 nm. Insets show lysosomes at the same scale. (b) Quantification of vesicular and lysosomal pH (see Materials and Methods). * - p<0.05 from untreated cells. (c) Cumulative distribution of pH values of individual secretory vesicles in untreated and METH-treated chromaffin cells (N = 212 vesicles in control, 146 in 1h METH, and 224 in 24h METH). All curves are different from each other with p<0.001 by Kolmogorov-Smirnov test.

Estimation of pharmacologically relevant METH concentration

METH and AMPH are not discriminated by humans, are similarly retained in tissues, and have nearly identical mechanisms of action and abilities to induce catecholamine release (Sulzer et al. 2005). These drugs are self-administered in typical doses of 250 – 500 mg by occasional users and at higher doses by chronic abusers (personal communication, Perry N. Halkitis, New York University). There is little data on the kinetics of METH and AMPH retention in the body after such large doses, but case reports showed plasma drugs concentrations of 30–100 μM and higher (Cravey and Baselt, 1968; Orrenius and Maehly, 1970; Jones and Holmgren, 2005) (see also (Talloczy et al. 2008) for the discussion of relevant plasma levels in drug addicts). Studies on the pharmacokinetics of METH levels in healthy adult human volunteers show that following a single oral 10–30 mg dose (0.14 – .43 mg/kg), plasma METH concentration reaches 0.1 – 0.7 μM; amphetamines are eliminated quite slowly by excretion and cytochrome P450 metabolism with the half-life of ~ 12 h (Perez-Reyes et al., 1991; Cook et al., 1992; Cho et al., 2001; Harris et al., 2003; Schepers et al., 2003; Kim et al., 2004). Because METH doses administered to patients with ADHD are usually ~2-fold higher (5 – 40 mg single doses; 0.17 – 1.33 mg/kg), and because amphetamines are retained in the brain and kidneys at >10-fold higher levels than in the plasma (Melega et al., 1995; Riviere et al., 2000), 5–10 μM METH concentrations and incubation times used in our experiments were assumed to be close to those in the brain following clinical administration.

Relationship between quantal size and vesicle pH

Catecholamine gradients across the secretory vesicles membrane depends on the electrochemical proton gradient:

$$\mathit{Log}\left(\frac{{\left[\mathit{CA}\right]}_{\mathit{vesicle}}}{{\left[\mathit{CA}\right]}_{\mathit{cytosol}}}\right)=\frac{\Delta \Psi \cdot F}{R\cdot T}+2\cdot \Delta \mathit{pH}$$

, where [CA]vesicle and [CA]cytosol are catecholamine concentrations inside and outside the vesicle, ΔΨ is electrical gradient, F is Faraday's constant, R is the universal gas constant, T is the absolute temperature, and ΔpH is the difference in pH between the vesicle and the cytosol. Assuming that electric gradient and pH in the cytosol are constant, a linear correlation is expected between [CA]_vesicle and [CA]_cytosol, and between [CA]_vesicle and the square of the proton concentration in the vesicle, [H⁺]_vesicle:

$${\left[\mathit{CA}\right]}_{\mathit{vesicle}}=A\cdot {\left[\mathit{CA}\right]}_{\mathit{cytosol}}\cdot {\left[H^{+}\right]}_{\mathit{vesicle}}^2,\text{where}A={10}^{\frac{\Delta \Psi \cdot F}{R\cdot T}}∕{\left[H^{+}\right]}_{\mathit{cytosol}}^2$$

Statistical analysis

Unless stated otherwise, data were analyzed by one-way ANOVA followed by Tukey's post-hoc test. For the normal probability plots (Fig. 7c) the data are presented as Log(Q_N), the distribution of which is close to normal (Finnegan and et al. 1996). Occasional large events with log quantal sizes > 3 standard deviations greater than the geometric mean of Log(Q_N) were excluded from the analysis. Data on all graphs are presented as mean ± SEM.

Fig. 7.

Effects of 10 μM METH on exocytotic quantal size in chromaffin cells. (a) Representative amperometric recordings of quantal catecholamine release from untreated cells and cells exposed to 10 μM METH for 1 h or 24 h. (b) Time dependence of quantal sizes in cells incubated with 10 μM METH. After 24 h of treatment, the cells were washed with drug-free media and incubated for additional 6 h before the amperometric recordings (grey bar). Data presented as percentages of changes relative to matched untreated sister cultures at the same day. * - p < 0.05 compared to untreated cells (see also Table 1). (c) Normal probability plots of the Log-transformed Q_N from untreated cells (N = 6,522 events) and cells treated with 10 μM METH for 1 h (N = 963 events), 24 h (N = 3,045) or 48 h (N = 955).

RESULTS

Prolonged METH causes rebound hyperacidification of chromaffin granules in situ

To investigate the effect of methamphetamine treatment on secretory vesicle acidity, we labeled large dense core vesicles (LDCV) in METH-treated chromaffin cells with acridine orange (AO), a fluorescent vital dye that has been used to study the acidification of secretory organelles, including synaptic vesicles (Tabb et al. 1992) and adrenal chromaffin granules (Steyer et al., 1997; Pothos et al., 2002).

As previously reported (Sulzer and Rayport 1990), acute (1 h) exposure of cultured chromaffin cells to METH resulted in a concentration-dependent decrease in vesicle-associated AO fluorescence, indicating a collapse of the vesicular proton gradient (data not shown). In contrast, we observed a gradual increase in AO fluorescence after prolonged exposure to METH, suggesting that chromaffin vesicles in drug-treated cells became more acidic than those in untreated cells (Fig. 1), a response we label rebound hyperacidification. The effect was reversible because 3–6 hours after washout of METH (Fig. 1, white arrow) vesicular pH returned to initial levels. Remarkably, chromaffin granules in cells exposed to the second METH dose after drug washout acidified more rapidly then during the first exposure. Thus, the cellular response that produced rebound hyperacidification, once induced, remained active even after METH removal.

Fig. 1.

Time course of METH-induced vesicle acidification. Time course of vesicle pH changes assessed by the retention of AO in cells treated with 10 μM and 100 μM METH. The upper bar shows the time scale of two consecutive METH additions (solid arrows) and a drug washout (open arrow). AO was added for 10 min before each measurement. Data point represent mean fluorescence intensities ± SEM from N = 30 – 40 cells. Dotted line and shadowed box represent the mean ± SEM in control cells at time zero.

To study the changes in vesicle acidity in more detail, we employed LysoSensor Yellow/Blue DND-160 (LSYB), a fluorescent pH probe that allows quantitative measurements of vesicular acidity. LSYB has two pH-sensitive fluorescence emission peaks and the ratio of the fluorescence intensities at their maxima can be calibrated to determine pH within acidic organelles in living cells (Diwu et al. 1999). As was seen with AO, LSYB rapidly accumulated in chromaffin cells with punctate distribution that filled almost the entire cytosol, consistent with the distribution of the secretory vesicles (Fig. 2a,b). We found that LSYB could be used to estimate intravesicular pH in a range of < 3.5 to ~ 6 (Fig. 2c; Supplementary Fig. 1), as was reported earlier (Holopainen et al. 2001).

Intravesicular pH in untreated chromaffin cells, determined from LSYB ratiometric measurements, was 5.4 ± 0.1, a value similar to estimates made by alternative methods, including electron spin resonance, uptake of radiolabeled tracers by isolated vesicles, and quantitative immunocytochemistry in situ (Pollard et al., 1979; Njus et al., 1986; Johnson, 1988; Pothos et al., 2002). METH exposure alkalinized the vesicles within minutes, as did ammonium chloride and chloroquine, two other broadly used amphiphilic weak bases (Fig. 2b,d). As was seen with AO, vesicles accumulated more LSYB in their lumen after prolonged METH treatment and changes in Y/B ratio confirmed the enhanced vesicle acidity. The accumulation of LSYB was abolished by pretreatment of cells with a monensin/nigericin cocktail (data not shown), indicating that changes in the fluorescence of the dye were indeed caused by a corresponding change in vesicular pH, rather than by an interaction between LSYB and METH. There was a linear relationship between the logarithm of the METH concentration and intravesiclular pH, which reached apparent values lower than 4.0 following the highest doses of METH (Fig. 2e).

To investigate whether the effect of METH on organelle acidification was specific to the drug, we examined the internal pH of secretory vesicles in cells that were treated for 24 h with weak base psychostimulants other than METH. While AMPH was as potent as METH in inducing vesicle hyperacidification, identical concentrations of methylphenidate and cocaine were significantly less efficient than METH (Fig. 3a). The magnitude of drug-induced hyperacidification was related reciprocally to the pKa of the psychostimulants examined (Fig. 3b), suggesting that the effect on intravesicular pH was dependent on the weak base properties of the drugs. However, ammonia, an inorganic weak base that, like METH, has a pKa of 9.2 but is less efficient than METH at alkalinizing vesicles, was also relatively inefficient at producing rebound vesicle hyperacidification (Fig. 3c). This observation highlights the importance of other factors, including the abilities of various weak bases to cross plasma and vesicle membranes either due to the lipophilicity of the compounds or their facilitated diffusion/active transport.

Fig. 3.

Dependence of vesicle acidity on weak-base properties of various drugs. (a) Vesicle pH in chromaffin cells treated for 24 h with drugs at 10 μm (grey bars) and 100 μM (black bars) concentrations. N = 30 cells in each group, * - p < 0.05 compared to untreated cells. (b) Dependence of vesicular pH on pKa values of the drugs. All weak bases were used at 100 μM concentrations for 24 h. Acidity of vesicles in cells treated with 50 μM puromycin (pKa = 7.2) for 6 h is shown as the leftmost point. The values of pKa for other drugs were form the online Hazardous Substances Data Bank (http://toxnet.nlm.nih.gov/cgi-bin/sis/htmlgen?HSDB): AMPH – 10.1, METH – 9.9, MPh – 8.8, cocaine – 8.6. (c) Dose-dependent accumulation of AO in vesicles of chromaffin cell treated with ammonium chloride for 24 h. Dotted line and shadowed box represent the mean ± SEM in cells treated with 100 µM METH for 24 h.

Ultrastructural analysis of METH-treated chromaffin granules

The intracytoplasmic compartments responsible for the trapping of AO and LSYB cannot be identified with certainty by fluorescence microscopy. We therefore studied the partitioning of DAMP, another weak base, which also accumulates in acidic compartments, but can be fixed in situ and visualized by electron microscopic immunocytochemistry (Anderson et al., 1984; Pothos et al., 2002). Intracellular DAMP was detected immunocytochemically by post-embedding labeling with immunogold. The number of gold particles within an organelle reflects the intracellular concentration of DAMP, which is in turn proportional to the proton gradient across the membrane that separates the interior of the organelle from the cytosol. By quantifying the density of gold particles, therefore, the ΔH⁺ could be estimated. The density of particles in nuclei was also determined and used as a reference to estimate the pH within other organelles, assuming that the intranuclear pH 7.0 was unaffected by the treatments. The density of gold particles in lysosomes served as a positive control. Lysosomes were defined as acidic, large, pleomorphic, perinuclear organelles, whereas secretory vesicles had a uniform shape, were <300 nm in diamter and contained a round dense core. The number of lysosomes was considerably smaller than the number of LDCVs (~ 1 to 50 in Controls).

As expected, DAMP accumulated in LDCVs in both untreated and METH-treated cells (Fig. 4a). The average vesicular and lysosomal pH in untreated cells was 5.7 ± 0.1 and 5.2 ± 0.1, respectively; acute exposure to METH (1 h, 100 µM) alkalinized both organelles to near neutral pHs (Fig. 4b,c). Confirming the data obtained with AO and LSYB, vesicles in cells treated with METH for 24 h trapped significantly more DAMP than vesicles of untreated cells, reaching a pH that was equivalent to that of lysosomes (pH 5.1 ± 0.1); this decrease in intravesicular pH represented a 4-fold increase in the intravesicular proton concentration over that found in vesicles of untreated cells. Neither acute, nor prolonged METH exposure significantly altered the profile of individual secretory vesicles (mean area of 16,600 ± 1300 nm² in control vs. 15,400 ± 1600 nm² and 13,500 ± 1100 nm² in cells treated with 100 µM METH for 1h or 24h, correspondingly; mean ± SEM). While the degree of pH decrease estimated from DAMP partitioning was lower than that from LSYB measurements (see Discussion), the results confirm that long-term METH treatment induced rebound hyperacidification in secretory vesicles. The effect of prolonged METH treatment was specific for secretory vesicles, because lysosomes in cells treated with the drug for 24 h recovered their acidity to control levels, but did not hyperacidify (Fig. 4a,b).

To determine whether rebound hyperacidification resulted from changes in the expression levels of enzymes and transporters, we measured intravesicular pH in cells that were incubated with METH in the presence of puromycin to block protein translation (Fig. 5a). Puromycin treatment did not prevent vesicle hyperacidification, indicating that the phenomenon does not require de novo protein synthesis. To corroborate the involvement of an enzymatic component in the maintenance and alteration of vesicle acidity, we measured the temperature dependence of intravesicular pH in cultured chromaffin cells. The curve relating the dependence of pH to temperature, either in the presence or in the absence of METH, was saddle-shaped with the maximal proton gradient achieved at ~37°C (Fig. 5b). Such a relationship is common for mammalian enzymes, as increased enzymatic activity at higher temperatures is overcome by heat-induced protein denaturation and deactivation due to a cellular heat-shock response. However, when METH-specific rebound hyperacidification was displayed as the difference between intravesicular pH in untreated and drug-treated cells, the response increased steadily with temperature, reaching the highest difference at 42°C (Fig. 5c). We conclude that enzymatic processes involved in the induction of METH-related alterations in cellular homeostasis were not inactivated even at 42°C, similar to thermostable lysosomal enzymes (Keech and Wills 1979) or Na⁺-K⁺-ATPase (Burdon et al. 1984).

Fig. 5.

Involvement of enzymatic component in METH-induced vesicle hyperacidification. (a) Effect of translation inhibitor, puromycin, on METH-induced hyperacidification of chromaffin secretory vesicles. Cells were treated with 100 μM METH for 6 h with or without 50 μM puromycin, followed by pH measurements with LSYB. N = 30 – 40 cells, * - p < 0.005 compared to untreated cells. (b) Temperature dependence of vesicle pH in control cells and those treated with 100 –M METH for 6 hours. N = 40 cells in each group. (c) Temperature dependence of METH-induced vesicle hyperacidification calculated as the difference between the two curves on (b).

Changes in cytosolic and vesicular catecholamine concentrations after prolonged METH

To investigate whether METH-induced rebound hyperacidification affected the vesicular content of catecholamines, we used amperometric recordings with carbon fiber electrodes to determine the number of catecholamine molecules released per secretory vesicle during exocytosis. The relationship between the concentration of METH and quantal size in cells treated with METH for 24 h was bell-shaped. At low-micromolar METH concentrations (5 – 10 μM) the amount of catecholamine released per vesicle was significantly increased, while at higher concentrations (100 – 500 μM) that are often employed in studies of METH-induced neurotoxicity in vitro, quantal size was decreased by as much as 50% (Fig. 6a).

Fig. 6.

METH concentration dependences of vesicular and cytosolic catecholamine contents. (a) Quantal size vs. METH concentration in cells treated with the drug for 24 h. Data are presented as the means ± SEM of the median Q_N values from individual cells (N = 97 cells in control and 11 – 43 in METH-treated groups). (b) Cytosolic catecholamine concentrations in cells incubated with METH for 24 h (N = 22 – 46 cells). * - p < 0.05 compared to untreated cells. Dotted line and shadowed box represent the mean ± SEM in untreated cells.

Vesicular catecholamine is generally thought to be at equilibrium with cytosolic transmitter concentration (Leviel et al., 1989; Chien et al., 1990). To investigate the discrepancy between unexpectedly low quantal sizes observed in hyperacidified vesicles following prolonged exposure to METH, we checked whether cytosolic catecholamines were affected by long exposures to the highest drug levels (100–500 μM) that are known to upregulate macroautophagy, a pathway that can lead to a bulk degradation of cytosolic content (Larsen et al., 2002). Using intracellular patch electrochemistry (IPE), we have previously shown that even though free cytosolic transmitter concentration increases several-fold in chromaffin cells acutely treated with 50 μM AMPH, it returns to initial levels after approximately 30 min of exposure (Mosharov et al. 2003). IPE measurements in cells treated with high-micromolar METH concentrations for 24h showed that cytosolic catecholamines were indeed significantly lower than those found in untreated cells (Fig. 6b), which may underlie the reduction in quantal size at high, possibly cytotoxic METH concentrations (Fig. 6a). Macroautophagy, which can be induced by prolonged METH treatment (Larsen et al., 2002; Talloczy et al., 2008) did not play a significant role in the regulation of quantal catecholamine release, as seen by a lack of effect by macroautophagy blockade (Supplementary Fig. 2).

Next, we investigated the changes in amperometric quantal events induced by pharmacologically relevant low-micromolar METH doses (see Methods) in further detail. In agreement with previous reports (Sulzer et al., 1992a; Sulzer et al., 1995b; Mundorf et al., 1999), acute 1 h exposure to 10 μM METH decreased the amplitudes of amperometric events (I_max) and the number of catecholamine molecules exocytosed from individual vesicle fusion events (Q_N) (Fig. 7a,b and Table 1). Analysis of quantal size distributions using a normal probability plot (Sulzer and Pothos 2000) revealed that acute exposure to METH produced a downward shift of the population of quantal sizes (Fig. 7c).

Table 1.

Comparative characteristics of amperometric events recorded from chromaffin cells treated with 10 μM METH. The median values of each parameter were first calculated for amperometric events recorded from individual cells and then the average of the medians was computed. The data are presented as percentages of changes in amperometric spike characteristics relative to matched controls recorded from sister cultures at the same day (N = 85 cells for control, 30 cells for 1h METH, 72 cells for 24h METH and 15 cells for 48 h METH). For absolute values of spike parameters in control cells and those treated with 10 μM METH for 24 h see Supplementary Table 1.

	Inter-spike interval	Events/stimulation	t1/2	Imax	QN	trise (25–75%)	Rise slope	Fall τ1	Fall τ2
1h METH	112 ± 17	106 ± 11	136 ± 9*#	38 ± 3*	59 ± 4*	168 ± 15*	26 ± 5*	216 ± 16*	164 ± 26*
24h METH	111 ± 16	111 ± 7	102 ± 8	143 ± 9*	126 ± 5*	98 ± 5	203 ± 26*	95 ± 9	102 ± 6
48h METH	91 ± 13	90 ± 3	94 ± 14	148 ± 29*	146 ± 17*	102 ± 9	173 ± 69*	82 ± 14	114 ± 19

^#Although t_1/2 was significantly increased in cells acutely treated with METH, this might be due to several reasons, including alteration of the fusion pore kinetics or the composition of the intravesicular matrix. It is also possible that smaller amperometric spikes are more affected by the noise and thereby have longer t_1/2 due to artifact (Mosharov and Sulzer 2005).

^*p < 0.05 versus untreated cells.

In marked contrast to the acute effects of METH, 24 h or 48 h exposure to 10 μM METH increased quantal size by 26 % and 48 %, respectively, also enlarging the average I_max of amperometric events (Fig. 7b, Table 1). No significant difference was detected between cytosolic catecholamine levels in untreated cells and cells incubated with 10 μM METH (Fig. 6b). Quantal sizes from cells treated for one or two days displayed a uniform upward shift on the normal probability plot, indicating that transmitter release was increased by the same percentage in all secretory vesicles, regardless of their pre-treatment catecholamine content (Fig. 7c). These data demonstrate that at low, pharmacologically relevant METH concentrations and exposure times, vesicular quantal size can be augmented by amphetamines, in contrast to the predictions from the weak base model of amphetamine action but consistent with the effect of secretory vesicle rebound hyperacidification.

DISCUSSION

Neurosecretory vesicles can modulate their transmitter content and quantal size even after maturation (Sulzer and Pothos, 2000; Edwards, 2007). One of the regulatory mechanisms that can alter the amount of stored vesicular catecholamine is an adjustment of intravesicular pH gradient created by the vH⁺-ATPase (Johnson 1988). Studies on isolated chromaffin vesicles have established that they reach a stable pH of ~5.5 (Winkler and Westhead 1980), which provides a driving force for the transport of catecholamines into vesicles against an adverse ~10⁵ concentration gradient of transmitter between the vesicle interior and the cytosol (Johnson 1988). When the acidic intravesicular pH is collapsed by the vH⁺-ATPase inhibitor bafilomycin A (Zhou et al., 2000; Pothos et al., 2002) or following acute exposure to weak bases such as chloroquine, ammonia or amphetamines (Van der Kloot, 1987; Sulzer et al., 1995a; Mundorf et al., 1999; Camacho et al., 2006), there is a decrease in the number of catecholamine molecules released per exocytotic event. Using a combination of optical and electrochemical techniques, we confirm that acute METH exposure alkalinized chromaffin vesicles and decreases exocytotic quantal size, but further identified a previously unsuspected rebound vesicle hyperacidification that eventually overcompensated for the initial collapse of the transmembrane proton gradient.

The ability of secretory vesicles to undergo rebound hyperacidification was shown by live optical measurements of intravesicular acidity with AO and LSYB and by the quantitative immunoelectron microscopic measurement of DAMP. Under the same conditions of exposure to METH, the effect of rebound hyperacidification estimated in living cells by ratiometric measurements with LSYB (ΔpH 1.2) was larger than that estimated ultrastructurally with DAMP (ΔpH 0.6) . Both approaches, however, are based on the partition of lipophilic weak bases according to the intracellular pH gradients and are susceptible to experimental artifacts that can result in different apparent pH values. Intravesicular pH estimated from the partitioning of DAMP would be increased if DAMP were to diffuse out of vesicles during fixation. A slight efflux of DAMP out of vesicles might be expected to occur unless its coupling to protein by fixation was so rapid as to be almost instantaneous; the immunocytochemical detection of DAMP accumulation, therefore, probably underestimates the degree of intravesicular acidity. Indeed, our current DAMP-derived estimates in untreated cells (pH 5.7) were slightly more basic than those determined by alternative methods (pH 5.5) (Pollard et al., 1979; Njus et al., 1986; Johnson, 1988) or by LSYB (pH 5.4). On the other hand, while LSYB does not require fixation, it is possible that LSYB spectrum may change when the dye accumulates to a high concentration inside of vesicles, as occurs with AO (Mason and McCaffery 1964; Lee and Forte 1980), or that other factors affect its accumulation or calibration. If the LSYB measurements are accurate, the intravesicular acidity below pH 4.0 in cells treated with 100–500 μM METH (Fig. 2e) would be unexpectedly low for mammalian vesicles, although secretory organelles in plants and some algae equipped with essentially the same vH⁺-ATPase maintain a pH below 2.0 (Ziegler et al., 1995; Muller et al., 1996). While it is not clear which approach more accurately reflects intravesicular pH, both methods were consistent in demonstrating rebound hyperacidification of secretory vesicles.

Both the initial pH collapse and the rebound hyperacidification depended similarly on the METH concentration, suggesting that one might be a delayed consequence of the other. The process was reversible, independent of de novo protein synthesis, and occurred at a higher rate upon re-exposure to METH. Rebound hyperacidification, moreover, occurred with a variety of weak bases, and may result from the cytosolic acidification and decreased vesicular/cytosolic pH concentration gradient well known to follow weak base withdrawal, the latter of which is activated by a Ca²⁺ -dependent compensatory increase in proton uptake (Burns et al. 1991). This mechanism is consistent with recent findings showing that intravesicular Ca²⁺ is displaced into the cytosol in cells acutely exposed to weak bases (Mundorf et al. 1999), that Ca²⁺ entry causes transient acidification of chromaffin vesicles (Camacho et al. 2006), and that stimulation-dependent chromaffin vesicle acidification is dependent on Ca²⁺ entry (Pothos et al. 2002). We cannot, however, rule out the participation of additional mechanisms, including delivery and assembly of V₀ and V₁ subunits of the vH⁺-ATPase, a process described for human and porcine epithelial cells and yeast cells (Kane and Smardon, 2003; Sautin et al., 2005; Beyenbach and Wieczorek, 2006), a compensation of the ion gradients across the vesicle membrane (Grabe and Oster, 2001; Demaurex, 2002; Li et al., 2002), or an alteration in the permeability of vesicular membranes to protons, the so-called `proton leak' (Diwu et al., 1999a; Grabe and Oster, 2001; Demaurex, 2002).

An important functional consequence of METH-induced vesicle rebound hyperacidification is the corresponding alteration of quantal size of stimulation-dependent exocytosis. At low METH doses, when cytosolic catecholamine concentration was apparently unaffected, there was a nearly linear relationship between quantal size and the square of the intravesicular proton concentration (Fig. 8), in agreement with predictions from studies on isolated chromaffin vesicles (Johnson 1988; Maycox et al. 1990) (see Methods). Increased quantal size after prolonged METH treatment was uniform for all vesicles, as indicated by an upward shift of the entire quantal size distribution when displayed as a normal probability plot.

Fig. 8.

Relationship between the square of the vesicle proton concentration and quantal size. The values for Q_N are the averages of cell medians from untreated cells and those treated with 10 μM METH for 24 and 48 h. The square of the vesicular proton concentration, which has been found in empirical studies to be directly related to vesicular catecholamine concentration (see Methods), was calculated from ratiometric pH measurements as: [H]² = (10^−pH)².

The effects of METH on quantal size, however, were different at the highest drug concentrations. Although there was a substantial rebound hyperacidification at these levels, the quantal size was decreased. Possible reasons for the bell-shaped dependence of quantal size on the METH concentration include the effects of METH on catecholamine vesicular uptake transporter (Brown et al. 2000; Partilla et al. 2006), and the ability of high METH concentrations to induce neurotoxicity associated with an autophagic stress response (Larsen et al. 2002). The latter possibility is supported by the decrease in cytosolic catecholamine observed exclusively at higher drug levels that are employed in METH neurotoxicity studies (Bennett et al., 1993; Stumm et al., 1999; Larsen et al., 2002; Jimenez et al., 2004). Therefore, at high METH doses vesicular catecholamine uptake appears to be limited not by the acidity of the vesicles, but by the availability of the cytosolic catecholamine.

Overall, we found that at METH concentrations (5–10 μM) and incubation times (24–48 h) that may be close to those found in drug abusers and ADHD patients (see Methods), the net output of all processes affecting intracellular catecholamine homeostasis was to increase stimulation-dependent catecholamine release. Thus, while acute METH reduces the signal from evoked, stimulation-dependent release, prolonged METH renormalizes and could even enhance the stimulation-dependent exocytotic catecholamine release. As evoked catecholamine release is thought to reflect the incentive value of external stimuli to attention and working memory (Waelti et al. 2001) and as amphetamines are used to treat fatigue and attention deficit disorders, an ability of prolonged low levels of amphetamines administration to increase quantal size could contribute to their therapeutic effects.

The abbreviations used are

AO

acridine orange

CV

cyclic voltammetry

DAMP

3-(2,4 dinitroanilino)-3'-amino-N-methyl dipropylamine

LSYB

LysoSensor Yellow/Blue DND-160

METH

methamphetamine

MDC

monodancylcadaverine

I_max

spike amplitude

Q

quantal size (pQ)

Q_N

quantal size (number of molecules)

t_1/2

duration at half-height