- Email: [email protected]
Inhibition of low- and high-threshold Ca2+ channels of human neuroblastoma IMR32 cells by Lambert-Eaton myasthenic syndrome (LEMS) IgGs
ELSEVIER
NEUROSC LETTERS
Neuroscience Letters 181 (1994) 50-56
Inhibition of low- and high-threshold Ca2’ channels of human neuroblastoma IMR32 cells by Lambert-Eaton myasthenic syndrome (LEMS) IgGs C. Grassib, V. Magnelli”, V. Carabelli”, E. Sher”, E. Carbonea,* “Department of Anatomy and Human Physiology, Corso Raffaello 30, I-10125 Turin, Italy bInstitute of Human Physiology, UCSC, Rome, Italy ‘Department of Pharmacology, Center for Cytopharmacology, Milan. Italy Received
8 August
1994; Revised
version
received
12 September
1994; Accepted
12 September
1994
Abstract IgGs from two LEMS patients applied to human neuroblastoma IMR32 cells reduced the density of low- (LVA; T) and high-threshold (HVA; L and N) Ba” currents by different percentages: 36% (LVA) and 56% (HVA) for one and 48% and 45% for the other. A pharmacological assay of IgGs action based on the block of L-type channel by nifedipine and on the delayed activation of N-type channel by noradrenaline, indicated a preferential inhibition of the N-type current in IMR32 cells (55% and 47% for the two patients). The L-type current, contributing to approximately one-third of the total, was also depressed by LEMS IgGs but to a minor degree (49% and 30%). Except for an increase of single N-type channel inactivation, LEMS antibodies preserved the elementary properties of single HVA channels, suggesting that the macroscopic current reduction after IgGs treatment is likely due to a decrease in the number of active HVA CaZCchannels. Key words:
Ca 2+ channels; Lambert-Eaton
syndrome; Human neuroblastoma;
The Lambert-Eaton myasthenic syndrome (LEMS) is an autoimmune neuromuscular disease characterized by a decreased neurotransmitter release from nerve terminals [5]. The disease is frequently associated with smallcell lung carcinoma (SCLC) [12]. Recent reports have shown that immunoglobulins G (IgGs) purified from LEMS patient sera inhibit preferentially the voltage-dependent presynaptic Ca” channels thus reducing Ca2’dependent transmitter release following nerve activation [13] (for a review see [22]). Patch-clamp studies in bovine chromaffin and NG 108/15 hybrid cells suggested that LEMS IgGs exert a selective inhibitory action on the high-threshold (HVA) L-type Ca” channel [ 11,161. Recently, LEMS IgGs have been proposed to target also HVA P-type channels in human SCLC cells [27] and T-type channels in sensory neurons [7], suggesting a broader action of LEMS IgGs on neuronal Ca2’ chan-
*Corresponding
author.
Fax: (39) (11) 670-7708.
0304-3940/94/$7.00 0 1994 Elsevier Science Ireland SSDI 0304-3940(94)00715-2
Ltd. All rights reserved
IgG
nels. Consistent with this, IgGs purified from LEMS sera have been shown to immunoprecipitate [‘251]w-conotoxin (CTx)-GVIA-labelled N-type Ca” channels of human neuroblastoma and small-cell lung carcinoma cells [ 14,15,23,24]. w-CTx-GVIA-sensitive N-type channels are known to control hormone release from sympathetic nerve endings [9], to be localized presynaptically at the frog neuromuscular junction [20,26] and to represent a major fraction of HVA Ca*’ channels expressed by human SCLC cells [3]. Thus, the N-type channel may be both a target of LEMS IgGs and one of the antigenic stimuli for the production of the pathogenic autoantibodies. To verify this hypothesis we studied the effects of sera and purified IgGs from LEMS patients on the Ca*’ channel subtypes of partially differentiated human neuroblastoma IMR32 cells. These cells exhibits excitability and secretory properties similar to sympathetic neurons and express w-CTx-GVIA-sensitive N-type channels together with a significant fraction of L- and T-types [2]. In agreement with immunoprecipitation
C. Grassi et al. INeuroscience Letters 181 (1994) 50-56
studies [23], LEMS IgGs exert a marked depressive action on N-type channels with substantial effects also on T- and L-types, suggesting a broad inhibitory action of LEMS IgGs on multiple neuronal Ca 2÷ channel subtypes. IMR32 cells were cultured as previously described [2]. Ba 2+ currents were recorded in the patch-clamp wholecell configuration [8] using a List EPC-7 amplifier. The electrode resistance was 3-6 Mr2. External solution contained (mM): 110 NaCI, 20 BaC12, 2 MgC12, 10 HEPES and 3 x 10 -7 M TTX (pH 7.3). Internal solution was (mM): 110 CsCI, 10 TEA-C1, 2 MgC12, 10 EGTA, 8 glucose, 10 HEPES (pH 7.3). Leak and capacitive currents compensation was performed by subtracting Cd 2÷insensitive currents. Current densities were calculated by dividing the measured current by the cell capacitance. The linear cell capacitance was estimated either by the patch-clamp setting of Cs~owor by the current jump associated to a voltage ramp of 60 mV amplitude (15 ms). Single-channel experiments were performed in cellattached configuration [8]. The external solution, designated to zero membrane potential, contained (mM): 135 K-Asp, 5 EGTA, 1 MgC12, 10 HEPES. The pipette solution was (mM): 100 BaCI> 1 MgC12, 10 HEPES, 10 TEA-C1. To record single N-type channels 5/tM nifedipine (Bayer AG, Wuppertal, Germany) was added to both solutions to block L-type channels. Single L-type channels were studied in the presence of 3/IM Bay K 8644 (Bayer AG). Recordings were sampled at 10 kHz and filtered off-line at 2 kHz. Data analysis was done by AutesP software [1] (see legend Fig. 3). Whole-cell and single channel recordings were carried out at room temperature (20-22°C) on IMR32 cells at different days of differentiation (6-12 days). Cells at early days (6-8 days) contained a higher percentage of DHP-sensitive L-type channels (30-40%) with respect to fully differentiated ones (10-12 days). The latter contained mostly co-CTxGVIA-sensitive N-type channels and a minority of L-type (10-20%). Compared to previous works on differentiated IMR32 cells [2,19], this gave an overall increased percentage of DHP-sensitive currents and more variability in the proportion of L- versus N-type channels. For this reason, data from control and LEMS IgGs-treated cells were compared at the same day of differentiation. The whole-sera were obtained from three LEMS patients (B, C and M), while purified IgGs were from only two LEMS patients (D and G) and two control individuals. The IgGs were purified by a double ammonium sulphate precipitation, extensively dialyzed against phosphate buffered saline, lyophylized and kept frozen in stock solutions (20 mg/ml) [23]. Cells were incubated overnight either with sera diluted to 15-20% or with IgGs at 4 mg/ml before recording. Due to the limited amount of IgGs available the effects on N-type channels were studied through the inhibitory action of noradrenaline (NA) rather than through the irreversible
51
og-CTx-GVIA block. Both effects are specific for N-type channels [4,6,19,25] but the former had the advantage of being reversible. NA could be tested on many cells from the same culture dish, while co-CTx-GVIA required one dish for each attempt and thus a large amount of IgGs to obtain data with statistical significance. Data are given as mean + S.E.M. for n = number of cells. Statistical significance (P) was calculated using the Student's /-test. IMR32 cells are characterized by the presence of different HVA channel subtypes which have been separated in DHP-sensitive (L-type) and co-CTx-sensitive (N-type). The cells also express a sizeable density of LVA (T-type) channels [2]. Fig. 1 shows representative whole-cell Ba2+ current densities recorded from control IgG- (A) and LEMS IgG-treated cells (B) between - 3 0 and +40 mV from - 90 mV holding potential. A short prepulse to - 30 mV was used to both estimate the size of LVA currents and separate their contribution from the HVA currents that activate positive to -20 mV during subsequent depolarizations. LEMS IgGs from patient D and G (from here on indicated as IgGs(D) and IgGs(G)) reduced the size of both HVA and LVA Ba 2+ current density. IgGs(G) depressed more the HVA (56 + 7%, n = 37) than the LVA (36 + 10%, n = 29) while IgGs(D) had a more or less similar depressive effect on both current density (45 + 7%, n = 30 and 48 + 10%, n = 11, respectively) (Fig. 1C). Similar effects were observed also on cells treated with three LEMS sera. HVA and LVA current densities depression by LEMS IgGs occurred without significantly affecting their voltage-dependency and time course of activation (insets in Fig. 1B and D). Current density-to-voltage (1/V) relationships from control (o in Fig. 1D) and IgGs-treated cells (o) could be simply scaled by a factor of 2.13 to nearly overlap (dashed line). Also the time to peak (tp, IIl,[~in Fig. 1B2 inset) and the time constant of inactivation vs. voltage (Z'h,O,O) showed little changes between control (o,t2) and IgGs-treated cells (e,w). Percentage of inactivation was also little affected by LEMS IgGs. At +10 mV, HVA currents in control and LEMS IgG(G)-treated cells inactivated by 30 + 2% (n = 31) and 36 + 2% (n = 29) after 90 ms, respectively. Thus, IgGs inhibition of neuronal macroscopic Ba2+ currents occurred without significantly affecting their voltage-dependent activation-inactivation kinetics. To define LEMS IgGs specificity for the HVA channel subtypes we tested the blocking action of nifedipine (5 gM) on the HVA current density at +10 mV in cells treated with either control or LEMS IgGs (Fig. 2A). This gave a good estimate of the fraction of L- (nifedipinesensitive) and N-type channels (nifedipine-resistant) persisting after IgG-treatment. Compared to the blocking action on control cells, nifedipine was more effective on LEMS IgG-treated cells. Thus, in the case of IgGs(G), the block by nifedipine increased from 31 + 3% (n = 25,
52
C. Grassiet aL / Neuroscienee Letters 181 (1994) 50 56
PatientD
A1 -30.~ ~-,,,r ~
A2
ControlIgGs -20mY ~
,
~*
+
C ~
~
~ 1 o o ~ .,-~.
0
HVA
LVA
u PatientG •~ lO0g
o
• ControlIgGs [] LEMSIgGs
Z
B1
LEMSIgGs(D)
o .w -40
LEMSIgGslO} " ControlIgGs ~1.._ ,- 10ms
,w 0
~h
4 E ~_~ 00
~
60
.9-° 20 ms
,-t
".-_ 3 V ~ +2'0 m +40
trol IgGs o LEMS IgGs pA/pF "
Fig. 1. Inhibition of HVA and LVA Ba =+ currents by LEMS IgGs. A~,A2: whole-cell Ba 2+ current densities recorded from a IMR32 cell treated with control IgGs at the potential indicated. Test depolarizations were preceded by a 20 ms depolarization to - 30 mV to minimize the contribution of LVA channels. Vh - 8 0 mV. B~,B2: same as in panel A but from a IMR32 cell treated with LEMS IgGs(D). Insets: (Left) LVA current densities at - 30 mV from the same cells of panel A and B. Notice the similar activation-inactivation kinetics of the two traces. (Right) Time to peak (tp I,[]) and inactivation time constant (rh e,o) versus voltage from 12 control (o,[]) and 12 LEMS IgGs-treated cells (e,m). C: percentage inhibition of HVA and LVA current densities induced by IgGs(D) (empty bars, top) and IgGs(G) (empty bars, bottom) with respect to control IgGs (filled bars). Statistical significance of filled and empty bars data was < 0.01 in all four cases. D : / / V relationships of HVA current densities derived from 8 control (o) and 10 IgGs treated cells (o) (4 from patient D and 6 from patient G) using the protocol of panel A. The dashed line overlapping the control I / V was obtained by scaling the LEMS I / V curve by × 2.13.
control) to 41 + 5% (n = 28, LEMS) (P < 0.1) and for IgGs(D) from 45 + 3% (n = 28, control) to 55 _+ 5% (n = 17, LEMS) (P < 0.05) (Fig. 2A2, 2A3). This clearly suggests that, compared to control, LEMS IgG-treated cells contained a larger fraction of L-type channels. In addition, the comparable action of nifedipine on control and IgG-treated cells indicates that LEMS IgGs do not inhibit only N-type channels but also exert a significant depression on L-types. The percentage of L-type channel inhibition estimated by comparing nifedipine block in control and LEMS IgG-treated cells gave average reductions of 30% for IgGs(D) and 49% for IgGs(G). Similarly for the nifedipine-resistant current component (N-type) we estimated average inhibitions of 47% (IgGs(D)) and 55% (IgGs(G)). Nifedipine action on LEMS IgGstreated cells was tested at different membrane potentials and found comparable to that at +10 mV (Fig. 2B).
There was, however, a slight positive shift between control (o in Fig.2B) and nifedipine-resistant IIV curves (o) that gave origin to a nifedipine-sensitive I/V relationship shifted by approximately 10 mV to the left (dashed line). This derives, very likely, from the block of L-type channels that activate at more negative potentials with respect to N-type channels (see [18]). The inhibition of N-type channels by LEMS IgGs was further confirmed by testing the action of noradrenaline (NA) on N-type channels in IMR32 cells [19]. Like other neurotransmitters (see [25]), NA selectively depresses the size of N-type currents at +10 mV and prolongs their activation time course with little effects on neuronal L-type channels [4]. NA is thus a useful tool to identify N-type channels in excitable cells [18]. In 22 out of 25 IMR32 cells treated with control IgGs, 20 /.tM NA caused the characteristic prolongation and marked inhi-
C Grassi et aL/Neuroscience Letters 181 (1994) 50-56
bition of N-type currents during pulses to +10 mV (61 + 5% at the peak of control current) (Fig. 2C1). NA inhibition was more attenuated at higher potentials and could be facilitated by double-pulse protocols with preconditioning steps to +70 mV [19] (not shown). On the contrary, NA action was drastically attenuated in cells treated with IgGs(G). In 17 out of 22 cells, NA depression was significantly smaller (10 + 3%) and showed little effects on the activation kinetics (see Fig. 2C2). In the remaining cells NA action resembled that at control but was more attenuated (45 + 8% at the peak). There was also a marginal double-pulse induced facilitation under these conditions (not shown), suggesting a marked inhibition of N-type Ca 2÷ channels in LEMS IgGs-treated cells. To assay whether LEMS IgGs affected either the number or the elementary properties of HVA channel subtypes, we studied the activity of single N-type channels in cell-attached patches of LEMS IgGs-treated cells. Separation of N- from L-type channels was achieved using pipette and bath solutions containing 5 # M nifedipine.
A1 uJ
Control IgGs
53
The identity of single N-type channel activity was ensured also by the voltage range of activation (> - 10 mV in 100 mM Ba >) [6], the unitary conductance (18-20 pS) and the appearance of prolonged channel bursting activity followed by sweeps of relatively fast inactivation [I 7]) (Fig. 3Al). Voltage-dependency of activation, unitary conductance and open-closed time distributions of single N-type channels were not significantly affected by LEMS IgGs (Fig. 3A2,B,C). Single channel conductances of 20.8 + 3.7 and 18.2 + 5.2 pS (P < 0.2) were derived from the open-channel 1/11 relationships of control (O) and LEMS IgG-treated cells (I) (Fig. 3B). Open time histograms were fitted in both cases with a single exponential function with ro = 0.5 ms and 0.6 ms while closed time histograms were fitted with double exponential functions with similar fast and slow time constants [1] (see figure legend to Fig. 3C). Reconstituted N-type currents at +10 mV obtained by summing 45 current traces in control and IgG-treated cells were characterized by a fast activation phase followed by an incomplete inactivation (Fig. 3D). Inactivation developed with com-
A2 LEMSIgGs ID)
A3
LEMSIgGs (GI
+10mV Nife
It_
o_
~a-lO ms lOms
B -40
LEMSIgGslD) , , 0 ,mV.+iO
C1 ~,~
Control IgGs +lOmV ,,~
C2
LEMSIgGs (G) +10
"L o 10 ms -10 \
/- • Control o Nife 5pM pA/pF
Fig. 2. Effects of nifedipine (5 #M) and noradrenaline (20 #M) on control and LEMS IgGs-treated cells. A: representative block of HVA current densities at +10 mV by nifedipine from cells treated with control IgGs (A 0, IgGs(D) (A2) and IgGs(G) (A3). Each panel includes three traces recorded before (C), during (Nife) and after (R) nifedipine application. Vh -- 80 mV. B: current density vs. voltage relationships obtained before (0) and during (0) nifedipine application from 10 LEMS IgG(D)-treated cells. The dashed line represents the difference between control (0) and nifedipine-resistant (o) currents. C: representative action of NA on HVA current densities at +10 mV in control (C]) and LEMS IgG(G)-treated cells (C2). Recordings were before (C), during (NA) and after (R) noradrenaline application. Vh - 8 0 mV.
54
C. Grassi et al./Neuroscience Letters 181 (1994) 50-56
A1
mV -lO
Control .L
B
LEMS
,~, . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
-40
L
•
'I,T-'°
i
0 mV +40 .
I
.
I
.
i
.
I
......T v ........r"
-1.0 +10 ÷20 ~
~ n J pA. -3.O
Control
÷30
D
C1
6ooI 11I
$ "6
0
Control
Control %=lms .
10
.
.
.
.
30
.
50
+0oot( LEMS
JllPhi__ +o=O m
.
O
70 ms
•
•
0
I
½
'
4
' ms
~
. ~r"~ "~IT "I' .11llVlr
Z=40 ms
-(3
%=3ms
I
10
ao
•
so
200t 0
,o I0 ms
LEMS ii
70
ms
LEMS
--
'
;
'
ms
Fig. 3. LEMS IgGs do not affect single channel kinetic properties. A: cell-attached current recordings of single HVA channels obtained from a control cell (A]) and a LEMS IgG(D)-treated cell (A2). The selected traces were recorded at the indicated potentials from Vh - 9 0 mV in the presence of 5 # M nifedipine inside and outside the pipette. A comparable channel activity was detectable in both cases. B: unitary 1/V relationship for the channel recordings shown in panel A. Conductances of 20.8 + 3.7 pS, n = 7 for control (I) and 18.2 + 5.2 pS, n = 7 for LEMS-treated cell (I) were obtained by linear regression to the open-channel current amplitude vs. voltage. Open-channel amplitudes at different voltages were determined by manually fitting with a cursor well-resolved channel openings. Channel openings (or closure) were detected when more than one sampling point crossed the threshold level set at 50% of the unitary current. Idealized records obtained by an iterative computer program were used to construct open and closed times histograms [1]. C: closed (CO and open (C2) time distributions obtained from cell-attached patches in control and in LEMS IgG-treated cells. Experimental data were fitted by double (closed times) or single (open times) exponential functions. Mean values were as indicated• D: reconstituted whole-currents obtained from control and LEMS IgG-treated cells. Currents are sums of 45 traces showing single channel activity with average Po of 0.14 and 0.2 in control and IgG-treated patches. Inactivation time course was fitted by a single exponential curve yielding similar inactivation time constants (r) of 40 and 43.5 ms, respectively•
parable decaying time constants (40 vs. 43.5 ms) but was more pronounced after IgG treatment. This effect was not further investigated and could derive from the low signal-to-noise ratio due to the low probability of channel openings (Po 0.14 and 0.2 in control and LEMS IgGtreated cells) and the limited number of summed traces. Similarly to the N-type, also the elementary properties of L-type channels recorded in the presence of Bay K 8644 (3 #M) at more negative potentials (-40 to - 1 0 mV) were little affected by LEMS IgGs (not shown). We conclude therefore that LEMS IgGs act on neuronal Ca 2÷ channels mainly by down-regulating the number of functioning channels [11]. The marked inhibitory action of LEMS IgGs on the three Ca 2+ channel subtypes expressed by IMR32 cells (N, L and T) is in agreement with recent patch-clamp
reports on the broad action of LEMS IgGs on non-Ltype Ca 2+ channels (P- and T-types) [7,27] and with immunoprecipitation studies showing high affinity binding of LEMS IgGs to [125I]go-CTx GVIA labelled N-type channels [23]. The original proposal of a specific action on neuronal L-type channels [11,16] must then be reconsidered. On the other hand, the involvement of Ca 2÷ channels other than L in LEMS is suggested by a number of impelling pieces of evidence. First, neurotransmitter release in mammalian and amphibian neuromuscular junctions is controlled by presynaptic Ca 2+ channels that are insensitive to DHPs (N, P or Q-types) [21,28]. Thus, impairment of the CaZ+-dependent ACh release at the end-plate should involve some non-L- rather than L-type channels. Second, catecholamine release from sympathetic neurons is mainly controlled by N-type chan-
C. Grassi et al./Neuroscience Letters 181 (1994) 50-56
nels [9]. Down-regulation of N-type channels as reported here may then account for the decreased efficiency of sympathetic and parasympathetic nerve terminals and for the characteristic disturbances of the autonomic system observed in LEMS patients [14,22]. A third reason to expect a broad specificity of LEMS IgGs on neuronal Ca 2+ channels derives from recent observations that SCLC cells express L- as well as N-, P- and other non-Ltype channels [3,10]. If the voltage-activated Ca 2+ channels of these neuroendocrine cells represent the antigenic stimulus for LEMS autoantibody production [12], it is reasonable to expect a number of pathogenic IgGs against a multitude of Ca 2+ channel proteins and, possibly, against other classes of membrane ion channels. The question is, therefore, whether LEMS IgGs are unspecific Ca 2+ channel inhibitors or possess some tissue- or species-specificity. Indeed, LEMS IgGs are shown to be ineffective against Ca 2+ channels of rodent cardiac cells, insect skeletal muscle (see [22]) and crustacean neurons (Sher, Richmond and Cooke, unpublished data). Thus, LEMS autoantibodies target well-defined Ca 2+ channel subtypes and can be still considered useful probes for them. A final interesting observation concerns the different percent of inhibition that IgGs from different LEMS patients exert on Ca 2+ channels. Different patients may develop IgGs with different specificity towards Ca 2+ channel subtypes with a consequent variability in the clinical manifestation of the disease. While autoantibodies against N- and P-type channels are expected to be responsible for the deficit of neurotransmitter release, still unknown are the physiological and neurological consequences for the down-regulation of neuronal L-and T-type channel. We wish to thank Drs. F. Clementi, J.M. Cooke, M. Passafaro, A. Pollo and J.M. Richmond for stimulating discussions. The financial support of Telethon-Italy (Grant 386 to E.C.) is gratefully acknowledged. [l] Carbone, E. and Lux, H.D., Single low-voltage-activated calcium channels in chick and rat sensory neurones, J. Physiol., 386 (1987) 571~o01. [2] Carbone, E., Sher, E. and Clementi, F., Ca currents in human neuroblastoma IMR32 cells: kinetics, permeability and pharmacology, Pfliigers Arch., 416 (1990) 170-179. [3] Codignola, A., Tarroni, E, Clementi, F., Pollo, A., Lovallo, M., Carbone, E. and Sher, E., Calcium channel subtypes controlling serotonin release from human small cell lung carcinoma cell lines, J. Biol. Chem., 268 (1993) 26240-26247. [4] Cox, D.H. and Dunlap, K., Pharmacological discrimination of N-type from L-type current and its selective modulation by transmitters, J. Neurosci., 12 (1992) 906-914. [5] Elmqvist, D. and Lambert, E.H., Detailed analysis of neuromuscular transmission in a patient with the myasthenic syndrome sometimes associated with bronchogenic carcinoma, Mayo Clin. Proc., 43 (1968) 689-713. [6] Elmslie, K.S., Kammermeier, EJ. and Jones, S.W., Reevaluation of Ca 2+channel types and their modulation in bullfrog sympathetic neurons, Neuron, 13 (1994) 217-228.
55
[7] Garcia, K.D., Beam, K.G., Walrond, J.P. and Mynlieff, M., LEMS serum reduces both LVA and HVA calcium current in DRG neurons, Soc. Neurosci. Abstr., 19 (1993) 198. [8] Hamill, O.P., Marty, A., Neher, E., Sakmann, B. and Sigworth, F.J., Improved patch-clamp techniques for high-resolution current recording from cells and cell-free membrane patches, Pfliigers Arch., 391 (1981) 85-100. [9] Hirning, L.D., Fox, A.E, McCleskey, E.W., Olivera, B.M., Thayer, S.A., Miller, R.J. and Tsien, R.W., Dominant role of N-type Ca 2+ channels in evoked release of norepinephrine from sympathetic neurons, Science, 239 (1988) 57-61. [10] Johnston, I., Lang, B., Leys, K. and Newson-Davis, J., Heterogeneity of calcium channel autoantibodies detected using a small-cell lang cancer line derived from a Lambert-Eaton myasthenic syndrome patient, Neurology, 44 (1994) 334-338. [1 l] Kim, Y.I. and Neher, E., IgG from patients with Lambert-Eaton syndrome blocks voltage-dependent calcium channels, Science, 239 (1988) 405-408. [12] Lambert, E.H., Rooke, E.D., Eaton, L.M. and Hodgson, C.H., Myasthenic syndrome occasionally associated with bronchial neoplasm: neurophysiologic studies. In H.R. Viets (Ed.), Myasthenia Gravis, 1961, pp. 362-410. [13] Lang, B., Johnston, I., Leys, K., Elrington, G., Marqueze, B., Leveque, C., Martin-Moutot, N., Seagar, M., Hoshino, T., Takahashi, M., Sugimori, M., Cherksey, B.D., Llin~ts, R. and Newsom-Davis, J., Autoantibody specificities in Lambert-Eaton myasthenic syndrome, Ann. NY Acad. Sci., 681 (1993) 382-393. [14] Lang, B., Newsom-Davis, J., Peers, J., Prior, C. and Wray, D.W., The effect of myasthenic syndrome antibody on presynaptic calcium channels in the mouse, J. Physiol., 360 (1987) 257-270. [15] Lennon, V.A. and Lambert, E.H., Autoantibodies bind solubilized Ca channel-to-conotoxin complexes from small cell lung carcinoma: a diagnostic aid for Lambert-Eaton myasthenic syndrome, Mayo Clin. Proc., 64 (1989) 1498-1504. [16] Peers, J., Lang, B., Newsom-Davis, J. and Wray, D.W., Selective action of myasthenic syndrome antibodies on calcium channels in a rodent neuroblastoma x glioma cell line, J. Physiol., 421 (1990) 293-308. [17] Plummer, M.R. and Hess, E, Reversible uncoupling of inactivation in N-type calcium channels, Nature, 351 (1991) 653~59. [18] Polio, A., Lovallo, M., Biancardi, E., Sher, E., Socci, C. and Carbone, E., Sensitivity to dihydropyridines, to-conotoxin and noradrenaline reveals multiple high-voltage-activated Ca 2+ channels in rat insulinoma and human pancreatic fl-cells, Pfliigers Arch., 423 (1993) 462-471. [19] Pollo, A., Lovallo, M., Sher, E. and Carbone, E., Voltage-dependent noradrenergic modulation ofto-conotoxin-sensitive Ca 2÷channels in human neuroblastoma IMR32 cells, Pfliigers Arch., 422 (1992) 75-83. [20] Robitaille, R., Adler, E.M. and Charlton M.E, Strategic location of calcium channels at transmitter release sites of frog neuromuscular synapses, Neuron, 5 (1990) 773-779. [21] Sano, K., Enomoto, K. and Maeno, T., Effects of synthetic c0-conotoxin, a new type of Ca 2+ antagonist, on frog and mouse neuromuscular transmission, Eur. J. Pharmacol., 141 (1987) 235241. [22] Sher, E., Carbone, E. and Clementi, F., Neuronal calcium channels as target for Lambert-Eaton myasthenic syndrome autoantibodies, Ann. NY Acad. Sci., 681 (1993) 373-381. [23] Sher, E., Gotti, C., Canal, N., Scoppetta, C., Piccolo, G., Evoli, A. and Clementi, F., Specificity of calcium channel autoantibodies in Lambert-Eaton myasthenic syndrome, Lancet, 2 (1989) 640643. [24] Sher, E., Pandiella, A. and Clementi, F., Voltage-operated calcium channels in small cell lung carcinoma cell lines: pharmacological, functional and immunological properties, Cancer Res., 5 (1990) 3892-3896.
56
C Grassi et al./Neuroscience Letters 181 (1994) 5 ~ 5 6
[25] Swandulla, D., Carbone, E. and Lux, H.D., Do calcium channel classification account for neuronal calcium channel diversity? Trends Neurosci., 14 (1991) 46-51. [26] Torri-Tarelli, F., Passafaro, M., Clementi, F. and Sher, E., Presynaptic localization of w-conotoxin-sensitive calcium channels at the frog neuromuscular junction, Brain Res., 547 (1991) 331-334. [27] Viglione, M. P. and Kim, Y.I., Lambert-Eaton syndrome serum
inhibits P-type Ca channels in small-cell lung cancer cells, Soc. Neurosci. Abstr., 19 (1993) 703. [28] Wessler, I., Dooley, D.J., Werhand, J. and Schlemmer, F., Differential effects of calcium channel antagonists (to-conotoxin GVIA, nifedipine, verapamil) on the electrically-evoked release of 3H]acetylcholine from the myenteric plexus, phrenic nerve and neocortex of rats, Arch. Pharmacol., 341 (1990) 288-294.
NEUROSC LETTERS
Neuroscience Letters 181 (1994) 50-56
Inhibition of low- and high-threshold Ca2’ channels of human neuroblastoma IMR32 cells by Lambert-Eaton myasthenic syndrome (LEMS) IgGs C. Grassib, V. Magnelli”, V. Carabelli”, E. Sher”, E. Carbonea,* “Department of Anatomy and Human Physiology, Corso Raffaello 30, I-10125 Turin, Italy bInstitute of Human Physiology, UCSC, Rome, Italy ‘Department of Pharmacology, Center for Cytopharmacology, Milan. Italy Received
8 August
1994; Revised
version
received
12 September
1994; Accepted
12 September
1994
Abstract IgGs from two LEMS patients applied to human neuroblastoma IMR32 cells reduced the density of low- (LVA; T) and high-threshold (HVA; L and N) Ba” currents by different percentages: 36% (LVA) and 56% (HVA) for one and 48% and 45% for the other. A pharmacological assay of IgGs action based on the block of L-type channel by nifedipine and on the delayed activation of N-type channel by noradrenaline, indicated a preferential inhibition of the N-type current in IMR32 cells (55% and 47% for the two patients). The L-type current, contributing to approximately one-third of the total, was also depressed by LEMS IgGs but to a minor degree (49% and 30%). Except for an increase of single N-type channel inactivation, LEMS antibodies preserved the elementary properties of single HVA channels, suggesting that the macroscopic current reduction after IgGs treatment is likely due to a decrease in the number of active HVA CaZCchannels. Key words:
Ca 2+ channels; Lambert-Eaton
syndrome; Human neuroblastoma;
The Lambert-Eaton myasthenic syndrome (LEMS) is an autoimmune neuromuscular disease characterized by a decreased neurotransmitter release from nerve terminals [5]. The disease is frequently associated with smallcell lung carcinoma (SCLC) [12]. Recent reports have shown that immunoglobulins G (IgGs) purified from LEMS patient sera inhibit preferentially the voltage-dependent presynaptic Ca” channels thus reducing Ca2’dependent transmitter release following nerve activation [13] (for a review see [22]). Patch-clamp studies in bovine chromaffin and NG 108/15 hybrid cells suggested that LEMS IgGs exert a selective inhibitory action on the high-threshold (HVA) L-type Ca” channel [ 11,161. Recently, LEMS IgGs have been proposed to target also HVA P-type channels in human SCLC cells [27] and T-type channels in sensory neurons [7], suggesting a broader action of LEMS IgGs on neuronal Ca2’ chan-
*Corresponding
author.
Fax: (39) (11) 670-7708.
0304-3940/94/$7.00 0 1994 Elsevier Science Ireland SSDI 0304-3940(94)00715-2
Ltd. All rights reserved
IgG
nels. Consistent with this, IgGs purified from LEMS sera have been shown to immunoprecipitate [‘251]w-conotoxin (CTx)-GVIA-labelled N-type Ca” channels of human neuroblastoma and small-cell lung carcinoma cells [ 14,15,23,24]. w-CTx-GVIA-sensitive N-type channels are known to control hormone release from sympathetic nerve endings [9], to be localized presynaptically at the frog neuromuscular junction [20,26] and to represent a major fraction of HVA Ca*’ channels expressed by human SCLC cells [3]. Thus, the N-type channel may be both a target of LEMS IgGs and one of the antigenic stimuli for the production of the pathogenic autoantibodies. To verify this hypothesis we studied the effects of sera and purified IgGs from LEMS patients on the Ca*’ channel subtypes of partially differentiated human neuroblastoma IMR32 cells. These cells exhibits excitability and secretory properties similar to sympathetic neurons and express w-CTx-GVIA-sensitive N-type channels together with a significant fraction of L- and T-types [2]. In agreement with immunoprecipitation
C. Grassi et al. INeuroscience Letters 181 (1994) 50-56
studies [23], LEMS IgGs exert a marked depressive action on N-type channels with substantial effects also on T- and L-types, suggesting a broad inhibitory action of LEMS IgGs on multiple neuronal Ca 2÷ channel subtypes. IMR32 cells were cultured as previously described [2]. Ba 2+ currents were recorded in the patch-clamp wholecell configuration [8] using a List EPC-7 amplifier. The electrode resistance was 3-6 Mr2. External solution contained (mM): 110 NaCI, 20 BaC12, 2 MgC12, 10 HEPES and 3 x 10 -7 M TTX (pH 7.3). Internal solution was (mM): 110 CsCI, 10 TEA-C1, 2 MgC12, 10 EGTA, 8 glucose, 10 HEPES (pH 7.3). Leak and capacitive currents compensation was performed by subtracting Cd 2÷insensitive currents. Current densities were calculated by dividing the measured current by the cell capacitance. The linear cell capacitance was estimated either by the patch-clamp setting of Cs~owor by the current jump associated to a voltage ramp of 60 mV amplitude (15 ms). Single-channel experiments were performed in cellattached configuration [8]. The external solution, designated to zero membrane potential, contained (mM): 135 K-Asp, 5 EGTA, 1 MgC12, 10 HEPES. The pipette solution was (mM): 100 BaCI> 1 MgC12, 10 HEPES, 10 TEA-C1. To record single N-type channels 5/tM nifedipine (Bayer AG, Wuppertal, Germany) was added to both solutions to block L-type channels. Single L-type channels were studied in the presence of 3/IM Bay K 8644 (Bayer AG). Recordings were sampled at 10 kHz and filtered off-line at 2 kHz. Data analysis was done by AutesP software [1] (see legend Fig. 3). Whole-cell and single channel recordings were carried out at room temperature (20-22°C) on IMR32 cells at different days of differentiation (6-12 days). Cells at early days (6-8 days) contained a higher percentage of DHP-sensitive L-type channels (30-40%) with respect to fully differentiated ones (10-12 days). The latter contained mostly co-CTxGVIA-sensitive N-type channels and a minority of L-type (10-20%). Compared to previous works on differentiated IMR32 cells [2,19], this gave an overall increased percentage of DHP-sensitive currents and more variability in the proportion of L- versus N-type channels. For this reason, data from control and LEMS IgGs-treated cells were compared at the same day of differentiation. The whole-sera were obtained from three LEMS patients (B, C and M), while purified IgGs were from only two LEMS patients (D and G) and two control individuals. The IgGs were purified by a double ammonium sulphate precipitation, extensively dialyzed against phosphate buffered saline, lyophylized and kept frozen in stock solutions (20 mg/ml) [23]. Cells were incubated overnight either with sera diluted to 15-20% or with IgGs at 4 mg/ml before recording. Due to the limited amount of IgGs available the effects on N-type channels were studied through the inhibitory action of noradrenaline (NA) rather than through the irreversible
51
og-CTx-GVIA block. Both effects are specific for N-type channels [4,6,19,25] but the former had the advantage of being reversible. NA could be tested on many cells from the same culture dish, while co-CTx-GVIA required one dish for each attempt and thus a large amount of IgGs to obtain data with statistical significance. Data are given as mean + S.E.M. for n = number of cells. Statistical significance (P) was calculated using the Student's /-test. IMR32 cells are characterized by the presence of different HVA channel subtypes which have been separated in DHP-sensitive (L-type) and co-CTx-sensitive (N-type). The cells also express a sizeable density of LVA (T-type) channels [2]. Fig. 1 shows representative whole-cell Ba2+ current densities recorded from control IgG- (A) and LEMS IgG-treated cells (B) between - 3 0 and +40 mV from - 90 mV holding potential. A short prepulse to - 30 mV was used to both estimate the size of LVA currents and separate their contribution from the HVA currents that activate positive to -20 mV during subsequent depolarizations. LEMS IgGs from patient D and G (from here on indicated as IgGs(D) and IgGs(G)) reduced the size of both HVA and LVA Ba 2+ current density. IgGs(G) depressed more the HVA (56 + 7%, n = 37) than the LVA (36 + 10%, n = 29) while IgGs(D) had a more or less similar depressive effect on both current density (45 + 7%, n = 30 and 48 + 10%, n = 11, respectively) (Fig. 1C). Similar effects were observed also on cells treated with three LEMS sera. HVA and LVA current densities depression by LEMS IgGs occurred without significantly affecting their voltage-dependency and time course of activation (insets in Fig. 1B and D). Current density-to-voltage (1/V) relationships from control (o in Fig. 1D) and IgGs-treated cells (o) could be simply scaled by a factor of 2.13 to nearly overlap (dashed line). Also the time to peak (tp, IIl,[~in Fig. 1B2 inset) and the time constant of inactivation vs. voltage (Z'h,O,O) showed little changes between control (o,t2) and IgGs-treated cells (e,w). Percentage of inactivation was also little affected by LEMS IgGs. At +10 mV, HVA currents in control and LEMS IgG(G)-treated cells inactivated by 30 + 2% (n = 31) and 36 + 2% (n = 29) after 90 ms, respectively. Thus, IgGs inhibition of neuronal macroscopic Ba2+ currents occurred without significantly affecting their voltage-dependent activation-inactivation kinetics. To define LEMS IgGs specificity for the HVA channel subtypes we tested the blocking action of nifedipine (5 gM) on the HVA current density at +10 mV in cells treated with either control or LEMS IgGs (Fig. 2A). This gave a good estimate of the fraction of L- (nifedipinesensitive) and N-type channels (nifedipine-resistant) persisting after IgG-treatment. Compared to the blocking action on control cells, nifedipine was more effective on LEMS IgG-treated cells. Thus, in the case of IgGs(G), the block by nifedipine increased from 31 + 3% (n = 25,
52
C. Grassiet aL / Neuroscienee Letters 181 (1994) 50 56
PatientD
A1 -30.~ ~-,,,r ~
A2
ControlIgGs -20mY ~
,
~*
+
C ~
~
~ 1 o o ~ .,-~.
0
HVA
LVA
u PatientG •~ lO0g
o
• ControlIgGs [] LEMSIgGs
Z
B1
LEMSIgGs(D)
o .w -40
LEMSIgGslO} " ControlIgGs ~1.._ ,- 10ms
,w 0
~h
4 E ~_~ 00
~
60
.9-° 20 ms
,-t
".-_ 3 V ~ +2'0 m +40
trol IgGs o LEMS IgGs pA/pF "
Fig. 1. Inhibition of HVA and LVA Ba =+ currents by LEMS IgGs. A~,A2: whole-cell Ba 2+ current densities recorded from a IMR32 cell treated with control IgGs at the potential indicated. Test depolarizations were preceded by a 20 ms depolarization to - 30 mV to minimize the contribution of LVA channels. Vh - 8 0 mV. B~,B2: same as in panel A but from a IMR32 cell treated with LEMS IgGs(D). Insets: (Left) LVA current densities at - 30 mV from the same cells of panel A and B. Notice the similar activation-inactivation kinetics of the two traces. (Right) Time to peak (tp I,[]) and inactivation time constant (rh e,o) versus voltage from 12 control (o,[]) and 12 LEMS IgGs-treated cells (e,m). C: percentage inhibition of HVA and LVA current densities induced by IgGs(D) (empty bars, top) and IgGs(G) (empty bars, bottom) with respect to control IgGs (filled bars). Statistical significance of filled and empty bars data was < 0.01 in all four cases. D : / / V relationships of HVA current densities derived from 8 control (o) and 10 IgGs treated cells (o) (4 from patient D and 6 from patient G) using the protocol of panel A. The dashed line overlapping the control I / V was obtained by scaling the LEMS I / V curve by × 2.13.
control) to 41 + 5% (n = 28, LEMS) (P < 0.1) and for IgGs(D) from 45 + 3% (n = 28, control) to 55 _+ 5% (n = 17, LEMS) (P < 0.05) (Fig. 2A2, 2A3). This clearly suggests that, compared to control, LEMS IgG-treated cells contained a larger fraction of L-type channels. In addition, the comparable action of nifedipine on control and IgG-treated cells indicates that LEMS IgGs do not inhibit only N-type channels but also exert a significant depression on L-types. The percentage of L-type channel inhibition estimated by comparing nifedipine block in control and LEMS IgG-treated cells gave average reductions of 30% for IgGs(D) and 49% for IgGs(G). Similarly for the nifedipine-resistant current component (N-type) we estimated average inhibitions of 47% (IgGs(D)) and 55% (IgGs(G)). Nifedipine action on LEMS IgGstreated cells was tested at different membrane potentials and found comparable to that at +10 mV (Fig. 2B).
There was, however, a slight positive shift between control (o in Fig.2B) and nifedipine-resistant IIV curves (o) that gave origin to a nifedipine-sensitive I/V relationship shifted by approximately 10 mV to the left (dashed line). This derives, very likely, from the block of L-type channels that activate at more negative potentials with respect to N-type channels (see [18]). The inhibition of N-type channels by LEMS IgGs was further confirmed by testing the action of noradrenaline (NA) on N-type channels in IMR32 cells [19]. Like other neurotransmitters (see [25]), NA selectively depresses the size of N-type currents at +10 mV and prolongs their activation time course with little effects on neuronal L-type channels [4]. NA is thus a useful tool to identify N-type channels in excitable cells [18]. In 22 out of 25 IMR32 cells treated with control IgGs, 20 /.tM NA caused the characteristic prolongation and marked inhi-
C Grassi et aL/Neuroscience Letters 181 (1994) 50-56
bition of N-type currents during pulses to +10 mV (61 + 5% at the peak of control current) (Fig. 2C1). NA inhibition was more attenuated at higher potentials and could be facilitated by double-pulse protocols with preconditioning steps to +70 mV [19] (not shown). On the contrary, NA action was drastically attenuated in cells treated with IgGs(G). In 17 out of 22 cells, NA depression was significantly smaller (10 + 3%) and showed little effects on the activation kinetics (see Fig. 2C2). In the remaining cells NA action resembled that at control but was more attenuated (45 + 8% at the peak). There was also a marginal double-pulse induced facilitation under these conditions (not shown), suggesting a marked inhibition of N-type Ca 2÷ channels in LEMS IgGs-treated cells. To assay whether LEMS IgGs affected either the number or the elementary properties of HVA channel subtypes, we studied the activity of single N-type channels in cell-attached patches of LEMS IgGs-treated cells. Separation of N- from L-type channels was achieved using pipette and bath solutions containing 5 # M nifedipine.
A1 uJ
Control IgGs
53
The identity of single N-type channel activity was ensured also by the voltage range of activation (> - 10 mV in 100 mM Ba >) [6], the unitary conductance (18-20 pS) and the appearance of prolonged channel bursting activity followed by sweeps of relatively fast inactivation [I 7]) (Fig. 3Al). Voltage-dependency of activation, unitary conductance and open-closed time distributions of single N-type channels were not significantly affected by LEMS IgGs (Fig. 3A2,B,C). Single channel conductances of 20.8 + 3.7 and 18.2 + 5.2 pS (P < 0.2) were derived from the open-channel 1/11 relationships of control (O) and LEMS IgG-treated cells (I) (Fig. 3B). Open time histograms were fitted in both cases with a single exponential function with ro = 0.5 ms and 0.6 ms while closed time histograms were fitted with double exponential functions with similar fast and slow time constants [1] (see figure legend to Fig. 3C). Reconstituted N-type currents at +10 mV obtained by summing 45 current traces in control and IgG-treated cells were characterized by a fast activation phase followed by an incomplete inactivation (Fig. 3D). Inactivation developed with com-
A2 LEMSIgGs ID)
A3
LEMSIgGs (GI
+10mV Nife
It_
o_
~a-lO ms lOms
B -40
LEMSIgGslD) , , 0 ,mV.+iO
C1 ~,~
Control IgGs +lOmV ,,~
C2
LEMSIgGs (G) +10
"L o 10 ms -10 \
/- • Control o Nife 5pM pA/pF
Fig. 2. Effects of nifedipine (5 #M) and noradrenaline (20 #M) on control and LEMS IgGs-treated cells. A: representative block of HVA current densities at +10 mV by nifedipine from cells treated with control IgGs (A 0, IgGs(D) (A2) and IgGs(G) (A3). Each panel includes three traces recorded before (C), during (Nife) and after (R) nifedipine application. Vh -- 80 mV. B: current density vs. voltage relationships obtained before (0) and during (0) nifedipine application from 10 LEMS IgG(D)-treated cells. The dashed line represents the difference between control (0) and nifedipine-resistant (o) currents. C: representative action of NA on HVA current densities at +10 mV in control (C]) and LEMS IgG(G)-treated cells (C2). Recordings were before (C), during (NA) and after (R) noradrenaline application. Vh - 8 0 mV.
54
C. Grassi et al./Neuroscience Letters 181 (1994) 50-56
A1
mV -lO
Control .L
B
LEMS
,~, . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
-40
L
•
'I,T-'°
i
0 mV +40 .
I
.
I
.
i
.
I
......T v ........r"
-1.0 +10 ÷20 ~
~ n J pA. -3.O
Control
÷30
D
C1
6ooI 11I
$ "6
0
Control
Control %=lms .
10
.
.
.
.
30
.
50
+0oot( LEMS
JllPhi__ +o=O m
.
O
70 ms
•
•
0
I
½
'
4
' ms
~
. ~r"~ "~IT "I' .11llVlr
Z=40 ms
-(3
%=3ms
I
10
ao
•
so
200t 0
,o I0 ms
LEMS ii
70
ms
LEMS
--
'
;
'
ms
Fig. 3. LEMS IgGs do not affect single channel kinetic properties. A: cell-attached current recordings of single HVA channels obtained from a control cell (A]) and a LEMS IgG(D)-treated cell (A2). The selected traces were recorded at the indicated potentials from Vh - 9 0 mV in the presence of 5 # M nifedipine inside and outside the pipette. A comparable channel activity was detectable in both cases. B: unitary 1/V relationship for the channel recordings shown in panel A. Conductances of 20.8 + 3.7 pS, n = 7 for control (I) and 18.2 + 5.2 pS, n = 7 for LEMS-treated cell (I) were obtained by linear regression to the open-channel current amplitude vs. voltage. Open-channel amplitudes at different voltages were determined by manually fitting with a cursor well-resolved channel openings. Channel openings (or closure) were detected when more than one sampling point crossed the threshold level set at 50% of the unitary current. Idealized records obtained by an iterative computer program were used to construct open and closed times histograms [1]. C: closed (CO and open (C2) time distributions obtained from cell-attached patches in control and in LEMS IgG-treated cells. Experimental data were fitted by double (closed times) or single (open times) exponential functions. Mean values were as indicated• D: reconstituted whole-currents obtained from control and LEMS IgG-treated cells. Currents are sums of 45 traces showing single channel activity with average Po of 0.14 and 0.2 in control and IgG-treated patches. Inactivation time course was fitted by a single exponential curve yielding similar inactivation time constants (r) of 40 and 43.5 ms, respectively•
parable decaying time constants (40 vs. 43.5 ms) but was more pronounced after IgG treatment. This effect was not further investigated and could derive from the low signal-to-noise ratio due to the low probability of channel openings (Po 0.14 and 0.2 in control and LEMS IgGtreated cells) and the limited number of summed traces. Similarly to the N-type, also the elementary properties of L-type channels recorded in the presence of Bay K 8644 (3 #M) at more negative potentials (-40 to - 1 0 mV) were little affected by LEMS IgGs (not shown). We conclude therefore that LEMS IgGs act on neuronal Ca 2÷ channels mainly by down-regulating the number of functioning channels [11]. The marked inhibitory action of LEMS IgGs on the three Ca 2+ channel subtypes expressed by IMR32 cells (N, L and T) is in agreement with recent patch-clamp
reports on the broad action of LEMS IgGs on non-Ltype Ca 2+ channels (P- and T-types) [7,27] and with immunoprecipitation studies showing high affinity binding of LEMS IgGs to [125I]go-CTx GVIA labelled N-type channels [23]. The original proposal of a specific action on neuronal L-type channels [11,16] must then be reconsidered. On the other hand, the involvement of Ca 2÷ channels other than L in LEMS is suggested by a number of impelling pieces of evidence. First, neurotransmitter release in mammalian and amphibian neuromuscular junctions is controlled by presynaptic Ca 2+ channels that are insensitive to DHPs (N, P or Q-types) [21,28]. Thus, impairment of the CaZ+-dependent ACh release at the end-plate should involve some non-L- rather than L-type channels. Second, catecholamine release from sympathetic neurons is mainly controlled by N-type chan-
C. Grassi et al./Neuroscience Letters 181 (1994) 50-56
nels [9]. Down-regulation of N-type channels as reported here may then account for the decreased efficiency of sympathetic and parasympathetic nerve terminals and for the characteristic disturbances of the autonomic system observed in LEMS patients [14,22]. A third reason to expect a broad specificity of LEMS IgGs on neuronal Ca 2+ channels derives from recent observations that SCLC cells express L- as well as N-, P- and other non-Ltype channels [3,10]. If the voltage-activated Ca 2+ channels of these neuroendocrine cells represent the antigenic stimulus for LEMS autoantibody production [12], it is reasonable to expect a number of pathogenic IgGs against a multitude of Ca 2+ channel proteins and, possibly, against other classes of membrane ion channels. The question is, therefore, whether LEMS IgGs are unspecific Ca 2+ channel inhibitors or possess some tissue- or species-specificity. Indeed, LEMS IgGs are shown to be ineffective against Ca 2+ channels of rodent cardiac cells, insect skeletal muscle (see [22]) and crustacean neurons (Sher, Richmond and Cooke, unpublished data). Thus, LEMS autoantibodies target well-defined Ca 2+ channel subtypes and can be still considered useful probes for them. A final interesting observation concerns the different percent of inhibition that IgGs from different LEMS patients exert on Ca 2+ channels. Different patients may develop IgGs with different specificity towards Ca 2+ channel subtypes with a consequent variability in the clinical manifestation of the disease. While autoantibodies against N- and P-type channels are expected to be responsible for the deficit of neurotransmitter release, still unknown are the physiological and neurological consequences for the down-regulation of neuronal L-and T-type channel. We wish to thank Drs. F. Clementi, J.M. Cooke, M. Passafaro, A. Pollo and J.M. Richmond for stimulating discussions. The financial support of Telethon-Italy (Grant 386 to E.C.) is gratefully acknowledged. [l] Carbone, E. and Lux, H.D., Single low-voltage-activated calcium channels in chick and rat sensory neurones, J. Physiol., 386 (1987) 571~o01. [2] Carbone, E., Sher, E. and Clementi, F., Ca currents in human neuroblastoma IMR32 cells: kinetics, permeability and pharmacology, Pfliigers Arch., 416 (1990) 170-179. [3] Codignola, A., Tarroni, E, Clementi, F., Pollo, A., Lovallo, M., Carbone, E. and Sher, E., Calcium channel subtypes controlling serotonin release from human small cell lung carcinoma cell lines, J. Biol. Chem., 268 (1993) 26240-26247. [4] Cox, D.H. and Dunlap, K., Pharmacological discrimination of N-type from L-type current and its selective modulation by transmitters, J. Neurosci., 12 (1992) 906-914. [5] Elmqvist, D. and Lambert, E.H., Detailed analysis of neuromuscular transmission in a patient with the myasthenic syndrome sometimes associated with bronchogenic carcinoma, Mayo Clin. Proc., 43 (1968) 689-713. [6] Elmslie, K.S., Kammermeier, EJ. and Jones, S.W., Reevaluation of Ca 2+channel types and their modulation in bullfrog sympathetic neurons, Neuron, 13 (1994) 217-228.
55
[7] Garcia, K.D., Beam, K.G., Walrond, J.P. and Mynlieff, M., LEMS serum reduces both LVA and HVA calcium current in DRG neurons, Soc. Neurosci. Abstr., 19 (1993) 198. [8] Hamill, O.P., Marty, A., Neher, E., Sakmann, B. and Sigworth, F.J., Improved patch-clamp techniques for high-resolution current recording from cells and cell-free membrane patches, Pfliigers Arch., 391 (1981) 85-100. [9] Hirning, L.D., Fox, A.E, McCleskey, E.W., Olivera, B.M., Thayer, S.A., Miller, R.J. and Tsien, R.W., Dominant role of N-type Ca 2+ channels in evoked release of norepinephrine from sympathetic neurons, Science, 239 (1988) 57-61. [10] Johnston, I., Lang, B., Leys, K. and Newson-Davis, J., Heterogeneity of calcium channel autoantibodies detected using a small-cell lang cancer line derived from a Lambert-Eaton myasthenic syndrome patient, Neurology, 44 (1994) 334-338. [1 l] Kim, Y.I. and Neher, E., IgG from patients with Lambert-Eaton syndrome blocks voltage-dependent calcium channels, Science, 239 (1988) 405-408. [12] Lambert, E.H., Rooke, E.D., Eaton, L.M. and Hodgson, C.H., Myasthenic syndrome occasionally associated with bronchial neoplasm: neurophysiologic studies. In H.R. Viets (Ed.), Myasthenia Gravis, 1961, pp. 362-410. [13] Lang, B., Johnston, I., Leys, K., Elrington, G., Marqueze, B., Leveque, C., Martin-Moutot, N., Seagar, M., Hoshino, T., Takahashi, M., Sugimori, M., Cherksey, B.D., Llin~ts, R. and Newsom-Davis, J., Autoantibody specificities in Lambert-Eaton myasthenic syndrome, Ann. NY Acad. Sci., 681 (1993) 382-393. [14] Lang, B., Newsom-Davis, J., Peers, J., Prior, C. and Wray, D.W., The effect of myasthenic syndrome antibody on presynaptic calcium channels in the mouse, J. Physiol., 360 (1987) 257-270. [15] Lennon, V.A. and Lambert, E.H., Autoantibodies bind solubilized Ca channel-to-conotoxin complexes from small cell lung carcinoma: a diagnostic aid for Lambert-Eaton myasthenic syndrome, Mayo Clin. Proc., 64 (1989) 1498-1504. [16] Peers, J., Lang, B., Newsom-Davis, J. and Wray, D.W., Selective action of myasthenic syndrome antibodies on calcium channels in a rodent neuroblastoma x glioma cell line, J. Physiol., 421 (1990) 293-308. [17] Plummer, M.R. and Hess, E, Reversible uncoupling of inactivation in N-type calcium channels, Nature, 351 (1991) 653~59. [18] Polio, A., Lovallo, M., Biancardi, E., Sher, E., Socci, C. and Carbone, E., Sensitivity to dihydropyridines, to-conotoxin and noradrenaline reveals multiple high-voltage-activated Ca 2+ channels in rat insulinoma and human pancreatic fl-cells, Pfliigers Arch., 423 (1993) 462-471. [19] Pollo, A., Lovallo, M., Sher, E. and Carbone, E., Voltage-dependent noradrenergic modulation ofto-conotoxin-sensitive Ca 2÷channels in human neuroblastoma IMR32 cells, Pfliigers Arch., 422 (1992) 75-83. [20] Robitaille, R., Adler, E.M. and Charlton M.E, Strategic location of calcium channels at transmitter release sites of frog neuromuscular synapses, Neuron, 5 (1990) 773-779. [21] Sano, K., Enomoto, K. and Maeno, T., Effects of synthetic c0-conotoxin, a new type of Ca 2+ antagonist, on frog and mouse neuromuscular transmission, Eur. J. Pharmacol., 141 (1987) 235241. [22] Sher, E., Carbone, E. and Clementi, F., Neuronal calcium channels as target for Lambert-Eaton myasthenic syndrome autoantibodies, Ann. NY Acad. Sci., 681 (1993) 373-381. [23] Sher, E., Gotti, C., Canal, N., Scoppetta, C., Piccolo, G., Evoli, A. and Clementi, F., Specificity of calcium channel autoantibodies in Lambert-Eaton myasthenic syndrome, Lancet, 2 (1989) 640643. [24] Sher, E., Pandiella, A. and Clementi, F., Voltage-operated calcium channels in small cell lung carcinoma cell lines: pharmacological, functional and immunological properties, Cancer Res., 5 (1990) 3892-3896.
56
C Grassi et al./Neuroscience Letters 181 (1994) 5 ~ 5 6
[25] Swandulla, D., Carbone, E. and Lux, H.D., Do calcium channel classification account for neuronal calcium channel diversity? Trends Neurosci., 14 (1991) 46-51. [26] Torri-Tarelli, F., Passafaro, M., Clementi, F. and Sher, E., Presynaptic localization of w-conotoxin-sensitive calcium channels at the frog neuromuscular junction, Brain Res., 547 (1991) 331-334. [27] Viglione, M. P. and Kim, Y.I., Lambert-Eaton syndrome serum
inhibits P-type Ca channels in small-cell lung cancer cells, Soc. Neurosci. Abstr., 19 (1993) 703. [28] Wessler, I., Dooley, D.J., Werhand, J. and Schlemmer, F., Differential effects of calcium channel antagonists (to-conotoxin GVIA, nifedipine, verapamil) on the electrically-evoked release of 3H]acetylcholine from the myenteric plexus, phrenic nerve and neocortex of rats, Arch. Pharmacol., 341 (1990) 288-294.