Immunoglobulins from motoneurone disease patients enhance glutamate release from rat hippocampal neurones in culture

Authors

  • Pavle R. Andjus,

    1. Biophysics Sector and INFM Unit, International School for Advanced Studies (SISSA), Via Beirut 2–4, 34014 Trieste, Italy.
    Search for more papers by this author
  • Zorica Stevic-Marinkovic,

    1. Biophysics Sector and INFM Unit, International School for Advanced Studies (SISSA), Via Beirut 2–4, 34014 Trieste, Italy.
    Search for more papers by this author
  • Enrico Cherubim

    Corresponding author
    1. Biophysics Sector and INFM Unit, International School for Advanced Studies (SISSA), Via Beirut 2–4, 34014 Trieste, Italy.
    • To whom correspondence should be addressed.

    Search for more papers by this author

  • Authors' present addresses
    P. R. Andjus: Institute of Physiology and Biochemistry, School of Biology, University of Belgrade, Studentski trg 3, P.O.B. 52, 11000 Belgrade, FR Yugoslavia.

  • Authors' present addresses
    Z. Stevic-Marinkovic: Institute of Neurology, Clinical Centre of Serbia, University of Belgrade, 11000 Belgrade, FR Yugoslavia.

  • Author's email address
    E. Cherubini: cher@sissa.it

Abstract

  • 1The whole-cell configuration of the patch-clamp technique was used to study the effects of immunoglobulins (IgGs) from patients affected by amyotrophic lateral sclerosis (ALS) on spontaneous glutamatergic currents in rat hippocampal cells in culture.
  • 2Focal application of ALS IgGs (100 μg ml−1) to hippocampal cells induced a rise in frequency but not in amplitude of spontaneous excitatory postsynaptic currents (SEPSC) which outlasted the period of IgG application. The mean frequency ratio (ALS over control) was 3.2 ± 0.6 (n. 19). No changes in frequency or amplitude of SEPSCs were observed after treatment with IgGs obtained from healthy donors (n=5) or from patients with Alzheimer's disease (n=4).
  • 3ALS IgGs also increased the frequency (by a factor of 2.0 ± 0.3) but not the amplitude of miniature excitatory postsynaptic currents (mEPSC) recorded in the presence of TTX (n=19). A rise in frequency of mEPSC was also seen in cells superfused with a calcium-free solution (n=4).
  • 4In the presence of TTX, ALS IgGs did not modify the amplitude or the shape of currents evoked by AMPA (100 μm), recorded at a holding potential of −50 mV.
  • 5It is concluded that ALS IgGs enhance both SEPSCs and mEPSCs through a presynaptic type of action. The excessive release of glutamate from nerve endings may be the cause of motoneurone death in ALS patients.

Amyotrophic lateral sclerosis (ALS) is a devastating neurological disorder which affects upper and lower motoneurones leading to muscle weakness and wasting, respiratory depression and death. There are two forms of ALS: the familial one (which constitutes only 10% of all cases) and the sporadic one which has a complex and still unknown origin. Among the different hypotheses put forward to explain the aetiopathogenesis of this disease, the excitotoxic theory has been widely accepted (Appel, 1993). According to this theory, a large and sustained increase in intracellular calcium may be the cause of motoneuronal death. An increase in [Ca2+]i would be induced by an overstimulation of glutamate receptors by excessive release of glutamate from presynaptic nerve endings. In analogy with myasthenia gravis (Ashizawa, Oshima, Ruan & Atassi, 1991) and Lambert-Eaton syndrome (Sher et al. 1989) an auto-immune process seems to be involved in this disease as suggested by the passive transfer of the disease when IgGs from ALS patients are injected into experimental animals (Appel, 1993). Possible targets for the antibodies would be voltage-dependent calcium channels. Inimunochemical studies have demonstrated the presence of antibodies against the αl subunit of the purified skeletal muscle L-type calcium channel in the serum of ALS patients (Kimura et al. 1994) as well as a correlation of antibody titres with the progression of the disease (Smith et al. 1992). However, preincubation of ALS IgGs with selective calcium channel antagonists has indicated that N- and P/Q-type channels only are responsible for ALS IgG-mediated excitotoxicity on a hybrid motoneurone cell line (Smith, Alexianu, Crawford, Nyormoi, Stefani & Appel, 1994). Moreover, P-type calcium channels have been shown to be involved in the ALS IgG-mediated increase in calcium currents in dissociated Purkinje cells (LlinÁs et al. 1993). Recently, however, the hypothesis that an auto-immune response against calcium channels plays a primary role in motoneurone death in ALS patients has been questioned (Arsac et al. 1996).

In the present experiments the effects of ALS IgGs on spontaneous release of glutamate have been tested on hippocampal cells in culture. It is reported that ALS IgGs induce a significant increase in frequency but not in amplitude of spontaneous and miniature glutamatergic currents through a mechanism that at least in part is independent of external calcium. Part of this work has been reported in a preliminary form (Andjus, Spergel & Cherubini, 1996).

METHODS

Cell cultures

Primary cell cultures were prepared from rat hippocampal pyramidal neurones according to the method described by Malgaroh & Tsien (1992) with slight modifications. Hippocampi were dissected free from 2–4-day-old postnatal animals previously anaesthetized with an intraperitoneal injection of urethane (2 g kg−1). Dissection and dissociation was performed in the presence of 100 μm kynurenic acid and 25 μm 2-amino-5-phosphonovaleric acid (Tocris Cookson). The isolated tissue was quickly sliced and digested with trypsin in the presence of DNAse. After stopping the digestion with trypsin inhibitor, cells were triturated in dissection medium containing DNAse. After two successive centrifugations at 40 g the resuspended cell pellet was distributed to 12 mm Nunc Petri dishes previously coated with polyornithine (20mgml−1) and Matrigel (2% (w/v), Collaborative Research; Bedford, MA, USA), each one containing 2 ml of modified minimal essential medium supplemented with fetal calf serum (Gibco). Two days after plating, the culture medium was supplemented with 1–5 μm cytosine-β-d-arabinofuranoside (Sigma), which inhibits the growth of glial cells. The culture medium was exchanged every 2–3 days. Experiments were performed on large (∼20μm diameter) pyramidal-like neurones between 6 and 18 days in culture.

Electrophysiological recordings

The whole-cell configuration of the patch-clamp technique was used to record spontaneous synaptic currents. Microelectrodes were pulled from thin-walled borosilicate capillaries (Hilgenberg, Malsfeld, Germany) and fire-polished. The internal solution contained (mm): potassium gluconate, 110; NaCl, 10; MgCl2, 5; EGTA, 0.6; Na2-ATP, 2; Hepes-KOH, 49; pH 7.2. After fining the pipette with the above solution the resistance was 5–8 MΩ. The external solution contained (mm): KCl, 3.5; NaCl, 132; MgCl2, 1; CaCl2, 2; glucose, 20; Hepes-NaOH, 10; pH 7.4.

To standardize the membrane potential level at which synaptic currents were measured, cells were voltage clamped using an EPC-7 amplifier (List Medical, Darmstadt, Germany) driven by an Atari computer (Acquire software, Instrutech Co, Elmont, NY, USA). To avoid contamination with inhibitory GABA-mediated synaptic currents, cells were clamped at −50 mV, a value that in our experimental conditions is similar to the reversal potential for chloride ions. Series resistance was monitored throughout the experiment with small voltage pulses. Tetrodotoxin (TTX, 1 μm, from Sigma) and 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX 20 μm, from Tocris Cookson) were bath applied, (R,S)-α-amino-3-hydroxy-5-methyl-4-isoxadepropionate (AMPA, 100 μm, from Tocris Cookson) was applied by gravity with a rapid perfusion system (Bio-Logic RSC 200, Claix, France). With this system, a complete exchange of the solution around the patched cell was achieved in less than 30 ms.

Data acquisition and analysis

Data were sampled at 5 kHz and stored on video tape for later transfer to a PMAC computer for analysis. Recordings were retrieved from tape and filtered with a Butterworth filter (Krohn-Hite 3202, Avon, MA, USA) at a cut-off frequency of 1 kHz and digitized at the rate of 2 kHz using an ITC–16 A/D converter (Instrutech Co). The synaptic events were detected using the AxoGraph 3.5.5 program (Axon Instruments), which uses a detection algorithm based on a sliding template. The template did not induce any bias in the sampling of the events since it was moved along the data trace one point at a time and was optimally scaled to fit the data at each position. The detection criterion was calculated from the template scaling factor and from how closely the scaled template fitted the data. Slow, presumably incompletely clamped events, were excluded from the present analysis. The threshold for detection was set at 4 times the standard deviation of the baseline noise. Using the same program the decay time constants of averaged SEPSCs or mEPSCs were taken from a mono-exponential fit of the decay phase. For each neurone, data samples obtained during 3–5 min recordings after treatment with ALS IgGs were compared with control samples (baseline activity or control IgGs) of similar duration and statistically analysed for differences with the Mann-Whitney U test. Individual data were expressed as mean amplitude or frequency ratio (treatment/control). Pooled data values from different cells are presented as the mean ±s.e.m. The Student's paired t test was used for comparison of paired data.

Immunoglobulins

Seven ALS patients (two male, five female, aged 56.6 ± 6.2 years with 1.1 ± 0.2 years illness duration; one had been treated for 10 days with prednisone, 50 mg per day, see Table 1), one Alzheimer's disease patient (female, aged 73 years, 1 year illness duration) and four healthy donors (four males; aged 47.5 + 4.9 years) provided the sera from which IgGs were obtained. IgGs were isolated using affinity chromatography (protein A-sepharose). For three patients (see Table 1), IgGs were prepared using sera previously heated to 56 °C for 30 min before elution to inactivate the complement. Since no statistical differences were found between the effects of IgGs prepared with heated-inactivated or native sera, experimental data were pooled together. Elution was performed with 1 m acetic acid. The first elution peak contained the IgG-free fraction, while the second peak included IgGs. Samples were dialysed, lyophilized, resuspended in Hank's balanced salt solution (Sigma) without Ca2+ and Mg2+ (pH 7.4) and frozen until used. Aliquots of diluted IgGs (0.1 mg ml−1 in standard external solution), kept frozen until used, were focally applied by pressure (515mmHg, 30s duration; Pneumatic Pressure System PPS-2, Medical Systems Corp., Grenvale, NY, USA) through a glass pipette (tip diameter of 2–5 μm) placed 100 μm from the patched cell. Usually, by means of two adjacent pipettes, treatment with control IgG was followed after 10–15 min by ALS IgG application on the same cell.

Table 1. Rise in frequency but not in amplitude of SEPSCs and mEPSCs induced by IgGs obtained from different ALS patients
PatientSexAge(years)Duration(months)Frequency ratioAmplitude ratio
  1. * IgGs isolated from heat-inactivated sera. †Patient treated with prednisone (50 mg day−1). For each patient, the upper values in the Frequency and Amplitude ratio columns refer to SEPSCs whereas the lower values refer to mEPSCs. In parentheses are the number of cells showing significant effects out of total number of neurones tested. The effects of IgGs from Patient 7 were compared with those of IgGs obtained from an Alzheimer's patient.

1*F3482.38 ± 0.45 (6/7)1.11 ± 0.13 (1/7)
    1.86 ± 0.30 (5/6)1.03 ± 0.08 (2/6)
2*F52101.96 ± 0.49 (3/7)1.15 ± 0.10 (2/7)
    1.41 ± 0.16 (7/10)1.10 ± 0.06 (3/10)
3*M47163.85 ± 2.04 (3/4)1.07 ± 0.20 (1/4)
    2.19 ± 0.94 (3/3)1.22 ± 0.38 (1/3)
4F77123.47 ± 1.41 (5/5)0.57 ± 0.10 (0/5)
    3.72 ± 1.87 (2/2)0.96 ± 0.02 (0/2)
5†M799
    1.66 ± 0.28 (2/2)0.96 ± 0.02 (0/2)
6F60242.21 ± 0.89 (2/3)2.04 ± 1.04 (1/3)
    0.94 (0/1)0.91 (0/1)
7F4761.21 ± 0.14 (1/2)0.89 ± 0.29 (0/2)
    1.54 (1/1)1.02 (0/1)

RESULTS

ALS IgGs increase the frequency of spontaneous glutamatergic currents

Spontaneous synaptic currents were recorded in the whole-cell configuration of the patch-clamp technique from hippocampal neurones in culture. At a holding potential (Vh) of −50 mV they were inward. Under our recording conditions, the reversal potential for inhibitory Cl-mediated GABAergic currents (−49 mV) was close to Vh and therefore inwardly directed synaptic activity could be ascribed only to spontaneous excitatory postsynaptic currents (SEPSCs). Moreover, SEPSCs were reversibly blocked by the selective AMPA receptor antagonist CNQX (20μM, Fig. 1A), suggesting that they were mediated by glutamate acting on the non-NMDA type of ionotropic glutamate receptors. They could be recorded for periods of up to 1 h without run down. Usually, after a control pre-treatment period lasting 5–10 min, IgGs from healthy donors or ALS patients were applied by pressure. In our culture conditions, spontaneous activity was often very high with synaptic currents occurring in bursts. We selected, therefore, only those cells in which spontaneous activity was relatively low. IgGs from healthy donors (100 μg ml−1) did not modify the frequency or amplitude of spontaneous glutamatergic synaptic activity. SEPSC frequency was 1.0 ± 0.2 Hz (n= 6) in control and 1.2 ± 0.2 Hz (n= 6) after treatment with IgGs from healthy donors. The amplitude of SEPSC also did not significantly (P > 0.05) differ between these two experimental groups. At −50 mV, amplitude values ranged from 5 to 120pA (on average 36 ± 17 pA, n= 6 or 28 ± 10 pA, n= 6, in control or after application of IgGs from healthy donors, respectively). In both cases current amplitude distribution histograms were skewed towards lower values. Since amplitude and frequency values of SEPSCs in control and after application of healthy donor IgGs did not significantly differ, data from these two experimental groups were pooled together. As shown in the example of Fig. 1A and C, local pressure application of ALS IgGs (100 μg ml−1, for 30 s) significantly (P < 0.05) increased the frequency of SEPSCs from 1.5 to 6.8 Hz. Similar results were found in 19/26 cells with ALS IgGs from six different patients. The effect became apparent 3–5 min after ALS IgG application and reached a peak 8–12 min later (see Fig. 2A). The frequency of spontaneous synaptic events then slowly declined to a steady state value still higher than in the control. A second application of ALS IgGs, 15–20 min after the first one was ineffective (n= 4). The potentiated synaptic activity was suppressed by CNQX (20 μm) and reappeared after wash at a frequency still higher than in control (Fig. 2B). A full recovery to initial control frequency values was never obtained for at least 20 min after ALS IgGs application. No changes in cell input resistance or baseline membrane current were detected during or after ALS IgG treatment. Overall, the mean frequency ratio (ALS over control) was 3.2 ± 0.6 (P < 0.01, Student's paired t test applied to individual raw data). The mean frequency ratios of ALS IgGs over control (2.9 ± 0.6, n= 14) and ALS over healthy donor IgGs (3.9 ± 1.3, n= 5) were not significantly different (P > 0.05). The effects of ALS IgGs on the amplitude of SEPSCs were rather variable (n=19 cells) comprising an increase (21% of cells), a decrease (26% of cells) or no change (53% of cells). Overall the mean amplitude ratio (ALS over controls) was 1.2 ± 0.2 (n= 19). A summary of the data on SEPSCs are represented in Fig. 3 and Table 1. In addition, the decay time constant values of SEPSCs obtained in three neurones (from two patients) before or after ALS IgG treatment were similar (6.98 ± 0.44 or 6.93 ± 0.46 ms, respectively). In 55% of the cases, a transient rise in frequency of SEPSCs (3- to 5-fold with respect to the baseline activity) was observed during pressure application of ALS IgGs, but not of IgGs from healthy donors. Occasionaly (10% of the cases) this activity was superimposed on slow inward currents (ranging from 50 to 150 pA).

Figure 1.

SEPSC amplitude and frequency distributions after treatments with control IgGs and ALS IgGs

A, current traces of spontaneous synaptic activity taken from the same cell after treatment with IgGs from a healthy donor (top trace) or ALS IgGs (middle trace). CNQX (20 μm; bottom trace) completely blocked spontaneous synaptic events. Currents underlying action potentials (indicated by dots) were not acquired by the event-catching algorithm. B, amplitude histograms (2 pA bins) of SEPSCs after treatment with control IgGs from a healthy donor or ALS IgGs from the cell shown in A. The mean amplitude values were 14.6 ± 1.1 and 22.6 ± 0.6 pA, respectively. In the control 208 events occurred, while after ALS IgGs 955 events were measured during the equivalent sampling time (140 s). C, cumulative distributions of inter-event intervals of SEPSCs after control IgGs (thin line) or ALS IgGs (thick line), taken from the same recording samples as in B. The mean SEPSC frequency rise was from 1.5 to 6.8 Hz. All data are from a single neurone.

Figure 2.

Time course of spontaneous synaptic activity before or after ALS IgGs

A, synaptic events from three different cells, detected in 20 min samples and sorted according to their time of occurrence into 1 min bins, before and after focal application (horizontal bar) of ALS IgGs obtained from two patients. For each bin, number of events have been averaged and normalized to the mean number of events occurring before ALS IgG administration. The interrupted horizontal line indicates the mean number of events per bin before ALS IgG administration. B, the synaptic events from a single cell, detected in 5 consecutive recording samples (each lasting ∼ll min) and sorted according to their time of occurrence into 1 min bins, before, during (horizontal bars) and after focal application of IgGs from a healthy donor, from an ALS patient or superfusion with CNQX (20 μm). The interrupted horizontal line indicates the mean number of events per bin prior to ALS IgG administration. Note the return of potentiated synaptic activity after washout of CNQX.

Figure 3.

ALS IgGs induce a rise in frequency but not in amplitude of SEPSCs and mEPSCs

Pooled data of the effects of ALS IgGs on the frequency (□) and amplitude ( inline image) of SEPSCs and mEPSCs. Data are expressed as the mean ratio of ALS IgGs over control (100 %). The data on SEPSC were taken from 19 cells treated with ALS IgGs (from 5 patients). Controls consisted of treatment with IgGs from two healthy donors (n= 5) or no IgG treatment (n= 14). The mEPSC data were obtained from 19 cells treated with ALS IgGs (from 6 patients). Controls consisted of treatments with IgGs from 1 healthy donor (n= 4) or no IgG treatment (n= 15). Error bars indicate s.e.m. The data were statistically analysed by Student's paired ? test on raw data; *P < 0.01.

ALS IgGs increase the frequency of spontaneous miniature glutamatergic currents

Miniature glutamatergic currents (mEPSC) were recorded in the presence of TTX (1 μm). Miniature events occurred at a frequency of i.1 ± 0.2 Hz (n= 22). A clear loss of higher amplitude events was produced by TTX since the mean amplitude values varied from 11 to 62 pA (on average 25 ± 3pA, n=22). Amplitude distribution histograms were skewed towards lower values (Fig. 4B). Focal application of ALS IgGs (obtained from six patients) induced a 2.0 ± 0.3-fold rise (P < 0.01, paired t test on individual raw data) in frequency (Fig. 4C) in 19/24 cells when compared with controls. The effect occurred with a similar time course as in the case of spontaneous activity. No significant changes (P > 0.05) in mEPSC amplitude were observed after application of ALS IgGs (Fig. 4B and Table 1). The overall mean amplitude ratio (ALS over control) was 1.0 ± 0.1 (n=19). The values of the decay time constant before or after ALS IgGs were similar to those observed for SEPSCs: 6.5 ± 0.4 and 6.2 ± 0.6 ms, before and after ALS IgGs, respectively (obtained from two patients, n= 4: see also Fig.4D). A summary of the data on mEPSCs are represented in Fig. 3 and Table 1.

Figure 4.

Amplitude and frequency distribution of mEPSCs after treatments with normal and ALS IgGs

A, current traces of mEPSCs recorded on the same neurone in the presence of TTX (1 μm) after treatment with IgGs from a healthy subject (top trace) or ALS IgGs (bottom trace). B, amplitude histograms (2 pA bins) of mEPSCs after treatment with control IgGs (top panel) or ALS IgGs (bottom panel). The mean amplitude values were 16.1 ± 1.1 and 17.3 ± 0.7 pA, respectively. For the same sampling time (2 min) 228 events occurred in control and 570 after ALS IgGs. C, cumulative distributions of inter-event intervals of mEPSCs after control (thin line) or ALS IgGs (thick line), taken from the same recording samples as in B. The mean mEPSC frequency rise was from 1.9 to 4.8 Hz. D, averaged mEPSC traces from control (89 events; thin line) and after ALS IgGs (87 events; thick line). The decay phases could be fitted with single exponentials having time constants of 6.21 and 5.88 ms, respectively. All data were obtained from a single neurone.

In order to evaluate the role of extracellular calcium in ALS IgG effects, immunoglobulins from healthy donors or ALS patients were applied to cells superfused with an extracellular solution in which calcium was substituted with magnesium. In these conditions the mean frequency of mEPSC was also augmented by a factor of 2.05 ± 0.5 (n= 4) when compared with controls (Fig. 5).

Figure 5.

Effect of ALS IgGs on the frequency of mEPSCs recorded in the absence of extracellular Ca2+

A, current traces of mEPSCs recorded in a nominally Ca2+-free extracellular medium (Ca2+ was substituted with Mg2+) after treatment with IgGs from a healthy donor (top trace) or ALS IgGs (bottom trace). B, cumulative distribution of niter-event intervals of mEPSOs taken from the same neurone after administration of IgGs from a healthy donor (thin line) or after subsequent application of ALS IgGs (thick line). The sampling time in both cases was the same (140 s).

ALS IgGs did not modify AMPA-evoked currents

In order to study whether ALS IgGs could affect AMPA-mediated postsynaptic currents, AMPA (100 μm) was applied directly to eight neurones (holding potential of −50 mV) in the presence of TTX (1 μm). AMPA induced inward currents that, after an initial peak, rapidly declined to a steady state value. No significant (P > 0.05) changes in peak amplitude or steady state currents were observed when AMPA was applied 5–8 min after ALS IgGs. The ratio between peak amplitude of AMPA–evoked currents before and after ALS IgGs was 1.04 ± 0.04 while between steady state currents it was 0.93 ± 0.04 (Fig. 6).

Figure 6.

ALS IgGs did not modify AMPA-evoked currents

Whole-cell currents evoked by application of AMPA (100 μm, horizontal bars) in the presence of TTX, before (left panel) or 8min after ALS IgGs treatment (right panel). Note the presence of miniature excitatory postsynaptic currents superimposed on the trace recorded after ALS IgGs treatment.

Alzheimer's disease IgGs did not affect SEPSCs or mEPSCs

In order to see whether the observed effects on the frequency of SEPSCs or mEPSCs were specific for immuno-globulins of patients affected by motoneurone disease, IgGs from a patient affected by another neurodegenerative disorder, Alzheimer's disease, were used. In seven additional experiments the effects of IgGs from the patient suffering from Alzheimer's disease were tested on SEPSCs (n= 4) or mEPSCs (n= 3). In both cases, no significant (P > 0.05) rise in frequency of synaptic currents was observed. The mean frequency ratio (Alzheimer's over controls) was 0.96 ± 0.06 for SEPSCs and 0.92 ± 0.13 for mEPSCs. In one case, a significant drop (by 31 %, P < 0.01) in frequency of mEPSC was found. After applying Alzheimer's IgGs, ALS IgGs still significantly (P < 0.05) increased the frequency of SEPSC in 2/2 cells (mean frequency ratio = 1.35 and 1.25). Moreover, ALS IgGs also increased the frequency of mEPSC (by a factor of 1.54, Fig. 7) when applied after Alzheimer's IgGs.

Figure 7.

ALS IgGs induce a rise in frequency of mEPSCs when compared with IgGs from Alzheimer's disease patients

A, current traces of mEPSCs obtained from the same neurone in the presence of TTX (1 μm) after treatment with IgGs from an Alzheimer's disease patient (top trace) or ALS IgGs (bottom trace). B, the top graph indicates cumulative distribution of niter-event intervals of mEPSCs after Alzheimer IgGs (thin line) or ALS IgGs (thick line). The mean frequency of events significantly (P < 0.01) rose from 0.9 to 1.4 Hz, respectively. The bottom graph represents cumulative amplitude distribution of mEPSC taken from the same recording samples. No significant (P > 0.05) differences were observed between Alzheimer and ALS IgGs (mean amplitudes were 6.2 and 6.4 pA, respectively). All recording samples had the same duration (170 s) and were taken from the same cell as in A. Both IgG samples were pre-heated at 56 °C for 30 min to eliminate possible complement contamination.

DISCUSSION

The main finding of the present study is that focal application of immunoglobulins from ALS patients to hippocampal neurones in culture substantially increases electrophysiologically monitored glutamate release from presynaptic nerve endings. A degree of specificity can be ascribed to the IgG serum fraction from ALS patients since no changes in frequency of the SEPSCs or mEPSCs was detected with IgGs from healthy donors or from a patient affected by Alzheimer's disease. Moreover, as expected for an antibody-mediated effect, the potentiation of glutamate release by ALS IgG developed slowly and it persisted after the glutamate receptor antagonism by CNQX was washed out. One limitation of the present study may be the use of hippocampal neurones which are normally not affected by motoneurone disease. Since sparing of hippocampal cells in ALS patients might be due to their inadequate exposure to ALS IgGs, hippocampal neurones in culture may still be considered as a useful experimental model for studying release mechanisms and excitotoxicity. The complement was not involved in ALS IgG effects since similar results were obtained with IgGs from patients whose sera were heated for 30 min at 56 °C before isolation of IgGs with affinity chromatography. Interestingly, in a few cases, an immediate effect of ALS IgGs on the steady current was observed. This effect was not seen when IgGs from healthy donors were used, ruling out the possibility that it was dependent on a pressure artefact or some factor present in the vehicle solution of the IgGs. Although further work is required to clarify the mechanisms underlying this effect, the possibility that antibodies from ALS patients can interact directly with postsynaptic glutamate receptors cannot be excluded. A similar type of mechanism has been recently reported for sera of patients with active Rasmussen's encephalitis, an intractable paediatric form of epilepsy (Twyman, Gahring, Spiess & Rogers, 1995).

The effect of ALS IgG on spontaneous synaptic currents was mainly presynaptic in origin since changes in frequency of SEPSCs or mEPSCs were not associated with modifications in current amplitude, decay time constant or holding current. Moreover, the observation that AMPA-evoked currents were not affected by ALS IgGs argues against a postsynaptic type of action. An increase in the rate of spontaneous acetylcholine release or in the quantal content of evoked cholinergic currents was found at the neuromuscular junction of mice acutely or chronically exposed to ALS IgGs (Uchitel, Appel, Crawford & Szczupak, 1988; Uchitel, Scornik, Protti, Pumberg, Alvarez & Appel, 1992) or in animal models of ALS (Appel, Engelhardt, Garcia & Stefani, 1991). It is worth noting that in the above mentioned studies with acute application of ALS IgGs to murine muscle, the potentiating effect was studied only after 4 to 5 h incubation and, in comparison with the present study, higher concentrations of immunoglobulins were used. It has been suggested that the enhanced acetylcholine release by ALS IgGs was due to a sustained increase in cytoplasmic calcium concentration in presynaptic nerve endings through voltage-dependent calcium channels (Uchitel et al. 1988; LlinÁs et al. 1993; Mosier et al. 1995). Likewise, the increase in action potential–dependent spontaneous glutamatergic release (SEPSC) found in the present experiments might be due to a rise in intracellular calcium directly through voltage-activated presynaptic calcium channels or indirectly via reduction in potassium conductances of presynaptic nerve endings leading to broadening of action potentials. Excessive levels of glutamate in the vicinity of motoneurones may lead to neuronal death through an excitotoxic action (Plaitakis, 1990; Krieger, Lanius, Pelech & Shaw, 1996). Recent work from our laboratory, using confocal laser scanning microscopy, has demonstrated on the same culture preparation that ALS IgGs depress depolarization-induced [Ca2+]i rise at the somatic level, an effect blocked by the P/Q-type calcium channel antagonist ?-agatoxin IVA (Andjus, Khirough, Nistri & Cherubini, 1996). Although we cannot exclude the possibility that voltage-dependent calcium channels responsible for transmitter release are different from those present on the soma, a reduction in somatic [Ca2+]i might still enhance the postsynaptic effects of glutamate by reducing the calcium-dependent desensitization of NMDA-sensitive glutamate receptors (Tong, Sheperd & Jahr, 1995). This hypothesis, however, seems unlikely since, in the present experiments, at a membrane potential of -50 mV only AMPA receptors mediated excitatory currents as demonstrated by their complete block by CNQX. Furthermore, doubts about the presence, in ALS patients, of antibodies against voltage-dependent calcium channels have recently been raised (Arsac et al. 1996; Nyormoi, 1996). It is therefore possible that the enhanced release of glutamate may constitute the primary effect of ALS IgG. An excessive production of glutamate could in turn cause the observed reduction in calcium entry through voltage-dependent calcium channels (Andjus et al. 1996). Such a mechanism, at least for high voltage-activated calcium channels, has already been reported (Cherubini & Nistri, 1991). In keeping with this view, the present data clearly show that ALS IgGs were able to increase the frequency of miniature glutamatergic currents recorded in the presence of TTX. Each miniature current is thought to result from the spontaneous release of a quantal packet of glutamate from nerve terminals following fusion of synaptic vesicles with the presynaptic membrane, an effect which is independent of extracellular calcium (Thompson, Capogna & Scanziani, 1993). ALS IgGs were still able to increase the frequency of miniature glutamatergic currents when extracellular calcium was omitted from the external solution. This indicates that immunoglobulins from ALS patients interfere with calcium-independent processes such as phosphorylation of vesicle-associated proteins (Greengard, Valtorta, Czernik & Benefenati, 1993). Protein phosphorylation could occur as a consequence of increased protein kinase C activity as recently reported for ALS patients (Krieger et al. 1996). Finally we cannot exclude the possibility that ALS IgGs impair glutamate membrane transport leading to excessive concentration of extracellular glutamate and to activation of glutamatergic non–NMDA receptor types (Rothstein, 1995). This would result in an increased [Ca2+]i and cell death due to activation of cytosolic proteases and other intracellular messengers.

Acknowledgements

We are grateful to Dr A. Nistri for helpful discussions during the course of this work. Immunoglobuhn fractions were kindly provided by Dr P. Annunziata (Istituto di Scienze Neurologiche, Facoltá di Medicina e Chirurgia, Universitá di Siena) and Dr S. Apostolski (Institute of Neurology, University Clinical Center, Belgrade). We are grateful to Dr Lj Dimitrijevic for the IgG preparation. This work was supported by a grant (no. 823) from the Telethon Foundation.

Ancillary