- Email: [email protected]
SnCl2 reduces voltage-activated calcium channel currents of dorsal root ganglion neurons of rats
NeuroToxicology 29 (2008) 958–963
Contents lists available at ScienceDirect
NeuroToxicology
SnCl2 reduces voltage-activated calcium channel currents of dorsal root ganglion neurons of rats Anke Tomaszewski a, Dietrich Bu¨sselberg b,* a b
Institut fu¨r Physiologie, Universita¨tsklinikum Essen, Universita¨t Duisburg-Essen, Hufelandstrasse 55, 45122 Essen, Germany Texas Tech University Health Sciences Center, Paul L. Foster School of Medicine, Department of Medical Education, 4800 Alberta Avenue, El Paso, TX 79905, USA
A R T I C L E I N F O
A B S T R A C T
Article history: Received 15 January 2008 Accepted 12 February 2008 Available online 17 February 2008
Stannous dichloride (SnCl2) occurs in the environment where it has been especially enriched in aquatic ecosystems. Furthermore, it is used in food manufacturing (e.g. for stabilizing soft drinks or as an anticorrosive substance) and in nuclear medicine where it is employed as a reducing agent for technecium99m (99mTc) and therefore is applied intravenously to human beings. SnCl2 is known to have toxic effects on the nervous system which can be related to alterations of intracellular calcium homeostasis ([Ca2+]i). In this study the whole cell patch clamp technique is used on dorsal root ganglion neurons of 3-week-old ‘‘Wistar’’ rats to evaluate the effects of SnCl2 on voltageactivated calcium channel currents (ICa(V)). ICa(V) were reduced concentration-dependently by SnCl2 (1–50 mM). 1 mM SnCl2 reduced ICa(V) by 8.1 4.5% (peak current) and 19.2 8.9% (sustained current), whereas 50 mM inhibited ICa(V) by 50.6 4.3% (peak current) and 55.6 11.3% (sustained current). Sustained currents were slightly but not significantly more reduced than peak currents. The effect appeared not to be reversible. The threshold concentration was below 1 mM. The current–voltage relation did not shift which is an indication that different calcium channel subtypes were equally affected. There was a slight but not significant shift of the activation/inactivation curves towards the depolarizing direction. We conclude that voltage-gated calcium channels are affected by Sn2+ similarly to other divalent metal cations (e.g. Pb2+ or Zn2+). The reduction of ICa(V) could be related to the neurotoxic effects of SnCl2. ß 2008 Elsevier Inc. All rights reserved.
Keywords: Voltage-activated calcium channel currents DRG Heavy metal SnCl2
1. Introduction Stannous chloride (SnCl2) is widely used in food manufacturing processes (e.g. as a stabilizer in soft drinks or as a protector against corrosive processes in cans) (Assis et al., 1998a; Dantas et al., 1999) and it has been enriched especially in aquatic ecosystems (Sala´nki et al., 2000; Gyo¨ri et al., 2000). Furthermore, it is employed as a reducing agent in technecium-99m (99mTc), a substance used in nuclear medicine for brain and renal scintigraphies and is therefore applied intravenously to humans (Agha et al., 1983; Assis et al.,
* Corresponding author. Tel.: +1 915 783 1700x251. E-mail address: [email protected] (D. Bu¨sselberg). Abbreviations: Ca2+, calcium ions; [Ca2+]i, intracellular calcium concentration; DRG, dorsal root ganglion neurons; Hg2+, mercury; ICa(V), voltage-activated calcium channel currents; INa(V), voltage-activated sodium channel currents; IV, current– voltage relation; 99mTc, technecium-99m; Pb2+, lead; SnCl2, stannous chloride, tin chloride; VACCs, voltage-activated calcium channels; Zn2+, zinc. 0161-813X/$ – see front matter ß 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.neuro.2008.02.003
1998a,b; Dantas et al., 1999; Silva et al., 2002). Without any special exposure blood tin concentrations in humans are around 2–9 mg/l (17–76 mmol/l; Hamilton et al., 1972/1973). When administered to the body it accumulates and its highest concentrations are found in the kidney (15.0 1.64 mg/g wet weight; Chiba et al., 1984). After oral application of SnCl2 (a cumulative dose of 17.7 mmol/kg body weight was applied) to rats, blood serum levels of 16–60 pmol/g (16–60 nmol/l) were found (Savolainen and Valkonen, 1986). As a heavy metal SnCl2 has numerous toxic effects. In addition to its known genotoxic potential (effects occurred at concentrations of 12.5–200 mg/ml SnCl2 (Dantas et al., 1999) or 1.1 10 4 to 4.4 10 4 M SnCl2 (Assis et al., 1998a,b; Dantas et al., 1999; Silva et al., 2002)) it induces cellular inactivation (Assis et al., 1998a), inhibits the immune response and is speculated to play a role in tumor generation in the thyroid gland (Silva et al., 2002). Moreover, after intoxication with SnCl2, hematocrit and haemoglobin as well as serum iron are diminished (Fritsch et al., 1978) and heme biosynthesis is compromised (effects occur after oral application of 10–200 mg Sn/kg) (Chmielnicka et al., 1992, 1994).
A. Tomaszewski, D. Bu¨sselberg / NeuroToxicology 29 (2008) 958–963
Stannous chloride also affects the nervous system. By generating reactive oxygen species it causes oxidative stress, involved in both aging and neurological diseases (Silva et al., 2002). Furthermore behavioral abnormalities which have a possible neuronal basis have been observed in living organisms intoxicated with SnCl2 (Gyo¨ri et al., 2000), and it has been shown that SnCl2 produces stimulation or depression of the central nervous system (Silva et al., 2002). In addition, the release of neurotransmitters in nerve terminals is accelerated by SnCl2 and thus neuromuscular transmission is facilitated (Silva et al., 2002). Further investigation revealed that SnCl2 causes a decrease of the acetylcholine-induced inward current of neurons of mollusc Lymnaea stagnalis L. (Gyo¨ri et al., 2000; Sala´nki et al., 2000) which is a direct action upon the neuronal membrane, is important in modulating synaptic transmission and facilitates the release of neurotransmitters from nerve terminals (Hattori and Maehashi, 1993; Salanki et al., 1998). This last effect of SnCl2 has been related to a modulation of the calcium homeostasis by Hattori and Maehashi (1989). Calcium ions (Ca2+) are present in the intra- and extra-cellular space and their concentration is strictly regulated (Berridge, 1997; Garcı´a et al., 2006). Changes of the intracellular calcium concentration ([Ca2+]i) are necessary in processes like muscle contraction, cell proliferation, activation of oncogenes and apoptosis (Berridge, 1997; Garcı´a et al., 2006). Neuronal cells are highly sensitive to slight changes of the [Ca2+]i and thus are able to use these modifications of [Ca2+]i as a second messenger and a means of signal transmission (Garcı´a et al., 2006). For example calcium entry into the cell triggers neurotransmitter release at the synaptic membrane. While slight modifications of the [Ca2+]i are physiological, greater changes of the [Ca2+]i may result in malfunction of the cell and lead to acute or chronic cell injuries or cell death (Berridge, 1997). There are several ways for a cell to regulate the [Ca2+]i. Either Ca2+ is removed from the cell via active transport mechanisms or it is moved to internal Ca2+ stores. Both mechanisms result in a decrease of the [Ca2+]i. In order to increase the [Ca2+]i, Ca2+ can be released from internal stores or enter the cell via Ca2+ selective pores located in the cell membrane. These can be receptormediated channels like the NMDA-receptor channel complex or voltage-activated channels. Therefore, it is possible that a malfunction of voltage-activated calcium channels results in changes of [Ca2+]i which may lead to a disruption of intracellular signalling and thus to cellular dysfunction (Bu¨sselberg, 2004). SnCl2 and its effect on [Ca2+]i have been investigated by Hattori and colleagues who found an increase of the [Ca2+]i which was related to SnCl2. They conclude that voltage-activated calcium channels (VACCs) (N-type (Hattori and Maehashi, 1992) or L-type (Hattori et al., 2001)) are involved, yet as far as we know modulations of currents through voltage-activated calcium channel currents (ICa(V)) by SnCl2 have never been studied directly. Divalent metal ions like lead (Pb2+), zinc (Zn2+), mercury (Hg2+), methyl mercury and trimethyl lead are known to reduce the current of voltage-activated ion channels of mammalian neurons. This is considered to be a possible explanation of the cognitive dysfunctions caused by heavy metals (Bu¨sselberg, 1995, 2004; Bu¨sselberg et al., 1994; Florea and Bu¨sselberg, 2005, 2006). Furthermore, Needleman and colleagues showed that metal ions like Pb2+ alter cognitive functions like learning and memory directly (Needleman, 1979; Needleman et al., 1979; Needleman and Bellinger, 1991). As SnCl2, another divalent metal ion is neurotoxic; too, it is likely to cause similar effects on neuronal cells which might e.g. alter the current of voltage-activated channels, too. Dorsal root ganglion neurons (DRG) are ‘‘strategically well placed’’ to cause a wide range of neuronal deficits when harmed, as
959
a malfunction can affect cellular metabolism as well as axonal transport (Gregg et al., 1992). In addition, they are not protected by the blood–brain-barrier. This study was designed to evaluate the effect of SnCl2 on VACCs of neuronal cells. Therefore, we have examined its effect on (ICa(V)) of dorsal root ganglion neurons of rats using the whole cell patch clamp method. 2. Methods 2.1. Cell culture preparation Short-term cell cultures (cells were kept in culture for a period less than 24 h) of dorsal root ganglion neurons from 3-week-old ‘Wistar’ rats were prepared as described previously (Tomaszewski and Bu¨sselberg, 2007). Shortly after anesthetizing the animals with approximately 5 ml isoflurane (CuraMed Pharma GmbH, Karlsruhe, Germany) and in the absence of pain reflexes, the rats were decapitated; the spinal column was removed and opened from the dorsal side. Thereafter the spinal cord was taken out; the DRGs were collected and placed in ice cooled F12 media (Sigma, Taufkirchen, Germany). After cutting off the spinal nerves and opening the ganglion capsules under optical control, the ganglia were transferred into a medium containing 0.9 ml F12 and 0.1 ml collagenase (2612.5 U ml 1 Sigma Type II) to be incubated for a period of 45 min in a humidified atmosphere containing 5% CO2 at a temperature of 37 8C. The ganglia were washed three times in 1.5 ml F12 medium to remove the collagenase. Afterwards they were trypsinized (2525 U Trypsin per ml F12 medium, Sigma Type IX) for 2 min under the same conditions. Trypsin was removed by washing the cells three times with F12 medium leaving a final volume of 0.7 ml. By triturating, using a fire-polished Pasteur pipette (tip diameter 150 mM) the neurons were released from the capsules. 50 ml of this suspension were placed in the middle of each Petri dish (3 cm, ‘Easy Grip’, Falcon, Gra¨tling-Lochau, Germany), and the cells were incubated for at least 2 h in which time they adhered to the substrate. One ml F12 medium with 10% horse serum and nerve growth factor was then added to each dish. Afterwards the cells were put into the incubator until they were used the next day. 2.2. Recording of voltage-activated calcium channel currents (ICa(V)) The whole cell patch clamp technique was used to record voltage-activated channel currents of DRG neurons. For this a HEKA EPC9 amplifier with the ‘‘EPC screen’’ software (HEKA Instruments, Port Washington, NY, USA) was used. Microelectrodes, consisting of borosilicate glass with filament (o.d.: 1.5 mm and i.d.: 0.7 mm; Biomedical Instruments (BMI), Zo¨llnitz, Germany), were pulled with an electrode puller (Sutter, model P-87, Sutter Instruments, Navato, USA or a DMZ-Universal Puller, Zeitzlnstrumente GmbH Mu¨nchen). Electrodes were fire polished to a final resistance of 3–5 MV using a Narashige microforge (MF-830, Narashige Instrument Laboratory, Tokyo, Japan). An external solution (containing 1 mM MgCl2, 10 mM HEPES, 0.001 mM TTX, 10 mM Glucose, 10 mM BaCl2 and 130 mM TEA-chloride with a pH of 7.3) to separate voltage-activated calcium channel currents was prepared and glass electrodes were filled with the adequate internal solution (consisting of 140 mM CsCl, 4mM MgCl2, 10 mM HEPES, 10 mM EGTA and 2 mM Na-ATP with an pH of 7.2). Both (the external as well as the internal) solutions are nominally calcium free and barium ions were used as charge carriers. After the formation of a giga-V seal and establishment of the whole cell configuration the cells were clamped at 80 mV. By depolarizing the cells to 0 mV for 75 ms, voltage-activated calcium channel
960
A. Tomaszewski, D. Bu¨sselberg / NeuroToxicology 29 (2008) 958–963
currents (ICa(V)) were elicited. Depolarization steps were started at 60 mV and the depolarizations were increased in intervals of 10 mV until a maximum depolarization of 40 mV was reached in order to record a current–voltage relation (IV) of ICa(V) (inset in Fig. 2). Activation and inactivation curves were generated by firstly hyperpolarizing the cell to 90 mV for a period of 75 ms, secondly depolarizing the cell to 80 mV for 75 ms and thirdly depolarizing to 0 mV for 75 ms before returning to the resting membrane potential of 80 mV. The second depolarization step was increased by +10 mV until 0 mV was reached (insets Fig. 3). All data were sampled and stored on hard disc (sample rate: 25 data points/ms). Tin chloride (SnCl2) (Tin (II) chloride dihydrate; Sigma–Aldrich, Steinem, Germany) was freshly dissolved (from a stock solution of 1 mM) in the external solution and was applied by using a bath application system with a continuous flow (flow rate: 5 ml/min). 2.3. Data analysis All currents were leak corrected using a p/3 protocol, where the neuron was hyperpolarized three times one third of the proposed voltage depolarization. The resulting currents were added and the sum was subtracted from the current which was elicited by the
depolarization. This guarantees that only the voltage-activated currents (but not passive ‘‘leak’’-currents) are taken into account. Peak currents were taken during the first 10 ms of each depolarization whereas sustained currents were measured between 60 and 70 ms of each depolarization. A minimum of five single successive peak/sustained currents ( 80 to 0 mV) which were taken under control conditions were averaged. All successive currents were expressed as a percentage of these currents to allow a comparison of single experiments. A ‘‘steady state’’ was defined as a situation in which the peak current amplitude did not decline for more than 5% within 40 s. Before calculating time courses, current–voltage relationships or concentration response curves, the currents were corrected for linear rundown. Therefore, the average reduction of the current within a defined time interval was estimated under control conditions and the calculated time-dependent rundown was added to each current. To calculate IV relationships, currents were expressed as a percentage of the maximum current of the control IV. The inactivation curves were calculated by defining the degree of depolarization at which the smallest current (either under control conditions or after SnCl2 was applied) was registered and the cell was completely inactivated as 100%. All following currents are expressed as reduction in percent of this current.
Fig. 1. Raw traces of ICa(V) generated by depolarizing the cell from the resting potential of 80 mV to 0 mV. Currents are shown before (black line) and after (grey line) the application of SnCl2 in the concentrations of (A) 1 mM, (C) 5 mM and (E) 10 mM. (B, D and F) show the reduction of ICa(V) peak currents (black diamonds) and sustained currents (white diamonds) caused by SnCl2 in the course of time for SnCl2 1 mM (B), 5 mM (D) and 10 mM (F). The grey line indicates the time during which SnCl2 has been applied. Data are shown with S.D.
A. Tomaszewski, D. Bu¨sselberg / NeuroToxicology 29 (2008) 958–963
961
Activation curves were created following a similar procedure. The degree of depolarization resulting in the largest current (under control conditions or after the application of SnCl2) was elicited and defined as 100%. All other currents are expressed as a percentage of this current. Data are given as means S.D. ‘‘P-values’’ were calculated with student’s t-test (double sided, unpaired or paired as required, type 1) (***P < 0.001; **P < 0.005; *P < 0.01) using Excel software. 3. Results Voltage-activated calcium channel currents were generated by depolarizing the neuron from the resting potential of 80 mV to 0 mV. A total of 33 neurons were included in this study. SnCl2 (10 mM) was applied to 22 cells, 5 cells (50 mM), 3 cells (5 mM) and 3 cells (1 mM). ICa(V) were sensitive to the application of SnCl2. Parts A, C and E of Fig. 1 represent raw traces of ICa(V) before and after the application of SnCl2. The peak currents of 1 nA (A), 800 pA (C) or 800–900 pA (E) were elicited 3 1 ms (A), 8 3 ms (C) or 5 2 ms (E) after starting the depolarization of the cell. The currents taken under control conditions (black line) declined by up to one third of the peak current during the 75 ms of depolarization. The application of SnCl2 in the concentrations of A, 1 mM; C, 5 mM; E, 10 mM resulted in a reduction of the ICa(V) (grey line). After the application of 1 mM SnCl2 the ICa(V) was comparable to control (8.1 4.5% reduction (peak current) and 19.2 8.9% (sustained current), respectively) as illustrated in Fig. 1B. A steady state was reached after 200 s. SnCl2 (5 mM) reduced the peak (sustained) currents significantly (P = 0.009, * (peak); P = 0.004, ** (sustained)) by
Fig. 2. (A) Current–voltage relationship of ICa(V) peak currents before (black diamonds) and after (white diamonds) the application of SnCl2 10 mM. Currents are expressed as a percentage of the maximum current of the control IV. Data are shown with S.D. N = 9. Inset: current–voltage relationship. Depolarization steps were started at 60 mV and increased by 10 mV (indicated by the broken lines) until a maximal depolarization of +40 mV was reached. Each depolarization lasted 75 ms. (B) Reduction of ICa(V) peak currents taken after the application of 10 mM SnCl2 in percent of the control IV.
Fig. 3. Activation (black) and inactivation (grey) curves taken before (solid line) and after (broken line) the application of SnCl2 10 mM. Currents are expressed as a percentage of the maximum current of the control IV. Data are shown with S.D. N = 8 (activation curves); N = 6 (inactivation curves). Insets: Waveforms of the recording protocol used to measure activation/inactivation curves (a) and IV relationships (b). A: voltage-activated channel currents were elicited by hyperpolarizing the cell to 90 mV. The second depolarization step was increased (indicated by the broken lines) by 10 mV until a final depolarization to 0 mV was reached. Each depolarization/hyperpolarization was carried out for the duration of 75 ms.
15.5 5.4% (23.8 5.5%) (Fig. 1D). The effect of SnCl2 (10 mM) was even more pronounced. Peak (sustained) currents were reduced significantly (P = 1.8 10 8, *** (peak); P = 1.2 10 7; *** (sustained)) by 27.9 6.7% (40.3 8.5%). A steady state was reached after 200 s. Since after eliminating SnCl2 from the bath solution there was hardly any recovery, the effect of SnCl2 on ICa(V) appeared not to be reversible (Fig. 1F). Gyo¨ri et al. (2000) tested the effect of SnCl2 on acetylcholineinduced and voltage-dependent inward currents of the mollusc, L. stagnalis L. In addition, they found a slight shift to the left of the current–voltage relation of voltage-activated sodium channels caused by SnCl2 (10 mM). To estimate whether ICa(V) are effected similarly by SnCl2 we measured the current–voltage dependency of ICa(V) before (Fig. 2A, black diamonds) and after (white diamonds) the application of SnCl2 (10 mM). The reduction of the current after the application of SnCl2 expressed as a percentage of the current measured under control conditions is shown in part B of Fig. 2. ICa(V) were reduced equally by 30–40% over the whole voltage range tested. There was no indication of a shift of the current–voltage relation. To test whether the activation or the inactivation of voltagegated calcium channels are influenced by SnCl2 (10 mM), activation as well as inactivation protocols were applied. This is illustrated in Fig. 3 where the activation (black) and the inactivation (grey) curves of VACCs before (solid line) and after (broken line) the
Fig. 4. Concentration dependency of ICa(V) peak currents (black diamonds) and sustained currents (white diamonds) after the application of SnCl2 in the concentrations of 1, 5, 10 and 50 mM. P-values are indicated by asterisks (***P < 0.0001; **P < 0.005; *P < 0.01).
962
A. Tomaszewski, D. Bu¨sselberg / NeuroToxicology 29 (2008) 958–963
application of SnCl2 are shown. The degree of inactivation decreased with an increasing pre-depolarization of the neurons. With a pre-depolarization of 30 mV 25% of the channels were inactivated and with a pre-depolarization of 20 mV the VACCs were not inactivated. Voltage-activated calcium channels were 50% activated at a depolarization to 30 mV under control conditions while at 20 mV a maximal depolarization was observed. After SnCl2 (10 mM) was applied, the activation as well as the inactivation curves were shifted slightly, but not significantly to more positive depolarization steps. The concentration dependency of the effect of SnCl2 on ICa(V) is shown in Fig. 4. As 1 mM SnCl2 reduced the ICa(V) peak current by 8.1 4.5% the threshold concentration is about 1 mM. The highest concentration tested in this study (50 mM) reduced the ICa(V) significantly (P = 0.0008; *** (peak), P = 0.0009; *** (sustained)) by 50.6 4.3% (peak) and by 55.6 11.3% (sustained), respectively. As after either oral or intraperitoneal admissions of SnCl2 or lead to rats the blood levels of SnCl2 have been less than or 1/10 of the blood lead levels (Zareba and Chmielnicka, 1989) and as lead causes a total blockage of ICa(V) at concentrations about 1 mM (Bu¨sselberg et al., 1994) we abstained from testing higher concentrations of SnCl2. Although the sustained current of all concentrations tested was more reduced than the peak current this difference was not significant for any of the concentrations tested. 4. Discussion We found that voltage-activated calcium channel currents were reduced by SnCl2 depending on the concentration applied. The threshold concentration was about 1 mM and 50 mM SnCl2 reduced the ICa(V) peak currents by 50.6 4.3%. This effect appeared to be irreversible in the course of our experiments. Seemingly contradictory to our findings, earlier research has revealed that SnCl2 results in an increased [Ca2+]i (Hattori and Maehashi, 1989, 1990; Hattori et al., 2001). As mentioned above there are two possibilities for a cell to raise its [Ca2+]i. Calcium could enter from the extracellular space or be released from internal stores. Hattori and Maehashi (1989, 1990) studied extracellularly recorded potentials of motor nerve terminals of the bullfrog. They found that apart from raising the [Ca2+]i, SnCl2 increased the prolonged positive deflection of this potential, which they suspect is related to calcium entry. By studying the interactions of SnCl2 with calcium channel blockers they found that the SnCl2-induced increase of the [Ca2+]i can be reduced by specific calcium channel blockers. As CdCl2 or v-conotoxin proved to be more effective than NiCl2 or nifedipine they concluded that this calcium entry into motor nerve terminals is linked to ICa(V) through N-type calcium channels. Furthermore, after proving that SnCl2 has similar effects on amphibians as on mammals (Hattori and Maehashi, 1993), Hattori et al. (2001) studied the effect of SnCl2 on the [Ca2+]i of osteoblastic MC3T3-E1 cells. They found, that in combination with a high extracellular concentration of potassium (10–100 mM) the [Ca2+]i was increased by SnCl2 (100 mM). This increase was inhibited by blockers of the L-type VACC (nicardipine, verapamil, and diltiazem) but not by N-type blockers (v-conotoxin GVIA). Further investigation revealed that SnCl2 applied without a high extracellular concentration of potassium (10–100 mM) did not change the [Ca2+]i while the rise of [Ca2+]i was inhibited by using a Ca2+-free medium or nifedipine. Hattori et al. (2001) conclude that calcium enters the cells through VACCs of the L-type although they also offer another possible explanation of the increase of the [Ca2+]i in suggesting that this increase is due to a release of Ca2+ from internal Ca2+ stores. A metal-triggered Ca2+ release from the calcium stores is also described for other metal compounds, e.g. Levesque et al. (1992),
who studied the effect of methyl mercury on nerve terminal mitochondria. They found that Ca2+ is released from internal stores under the influence of the heavy metal. In the light of our findings this seems to be the more plausible explanation and could explain the contradictory results. Calcium channels can be inhibited by either calcium or modulations of the membrane potential. In this study short depolarization pulses (75 ms) were used with Ba2+ as a charge carrier which minimizes the calcium-induced calcium channel inactivation as well as the voltage-dependent calcium channel inactivation. If, as is shown in this study, the VACCs are inhibited by SnCl2 but the calcium-induced calcium channel inactivation is prevented similarly, the [Ca2+]i could still be increased despite the inhibition of VACCs. In this case Ca2+ would not enter the cell through ICa(V) but Ca2+ influx would follow long-duration depolarizing pulses. As SnCl2 increased the prolonged positive deflection of extracellularly recorded potentials of motor nerve terminals of the bullfrog (Hattori and Maehashi, 1989, 1990) this could explain the contradictory results. On the contrary, the findings that the increase of [Ca2+]i can be inhibited by either a Ca2+-free medium or by blockers of calcium channels indicate that calcium enters the cell from the extracellular space. Yet in our study we found a clear reduction of ICa(V) after the application of SnCl2. One possible explanation of these differences is that Ca2+ enters the cell from the extracellular space not through voltage-activated channels but different calcium selective channels. Unfortunately this is not in agreement with the finding that the co-application of highly specific L- and N-type channel blockers inhibits the increase of the [Ca2+]i. Another possibility to explain the contradictory effects caused by SnCl2 is that Ca2+ enters the cell via voltage-activated calcium channels but not through the L or T-type. In our study the sustained currents of ICa(V) were slightly (but not significantly) more reduced by SnCl2 than peak currents of ICa(V). In addition to that the largest current elicited in the current–voltage relations both under control conditions and after the application of SnCl2 was generated by a depolarization from 80 to 20 or 10 mV. There was no indication of a shift of the current–voltage relation. Furthermore, there was no significant difference in activation/inactivation curves taken before and after the application of SnCl2. Therefore, we have no proof that sub-types (e.g. L-, N-or T-type) of VACCs are influenced differently as was proposed by Hattori and Maehashi (1990, 1991) and Hattori et al. (2001). Nevertheless, not all studies on SnCl2 are contradictory to our findings. Gyo¨ri et al. (2000), who compared the effect of SnCl2 on voltage-activated sodium channel currents (INa(V)) of the mollusc L. stagnalis L. to acetylcholine-induced currents, found that these currents were reduced by SnCl2 with an effective threshold concentration of 0.1 mM. Furthermore, there is a possibility that SnCl2 causes distinct effects in different concentrations. Hattori et al. (2001) applied SnCl2 in relatively high concentrations (about 0.1 mM) and found an increase of the [Ca2+]i, while we used SnCl2 in concentrations between 1 and 50 mM which reduced the calcium entry through ICa(V) and therefore should lower the [Ca2+]i. Moreover, it has been shown, that other divalent metal ions such as Zn2+ or Pb2+ also reduce voltage-activated channel currents (Bu¨sselberg et al., 1994, 1998; Bu¨sselberg, 1995). Apart from Pb2+, Zn2+ has been found to block the ICa(V) completely (more than 80%), but both metals differ in their effective concentration range: while 1 mM of Pb2+ was sufficient, 150–200 mM Zn2+ were needed to accomplish the same effects. The threshold concentration for Pb2+ was even below 0.1 mM (Pb2+). The hill coefficient of both metals was about 1 which suggests that the ICa(V) were directly blocked (Bu¨sselberg et al., 1994). As SnCl2 is also a divalent metal ion which affects human beings in a similar way as Pb2+ and Zn2+, it is
A. Tomaszewski, D. Bu¨sselberg / NeuroToxicology 29 (2008) 958–963
possible that it follows a similar mechanism of reaction at the channel side. Therefore, a reduction of ICa(V) by SnCl2 is reasonable since as a divalent cation it also could compete with Ca2+ at the calcium binding sites at or within any calcium conducting pore. 4.1. Toxicity of SnCl2 As SnCl2 occurs in our environment (especially in aquatic ecosystems (Sala´nki et al., 2000; Gyo¨ri et al., 2000)), is used in food manufacturing, and applied intravenously during radiological diagnostic procedures, human beings are likely to have contact with SnCl2. In various studies it has been shown that SnCl2 can be absorbed by mammal after oral as well as after intraperitoneal administration. It accumulates in internal organs such as brain, liver, bones and kidney in which the highest concentration (15.0 1.6 mg/ g wet weight after the administration of 5 mM SnCl2/(kg day) for a period of 6 days (Chiba et al., 1984)) is found (Chiba et al., 1984; Fritsch et al., 1978; Hayashi et al., 1994; Savolainen and Valkonen, 1986; Yamaguchi et al., 1982; Zareba and Chmielnicka, 1989). As described above SnCl2 causes behavioural abnormalities (Gyo¨ri et al., 2000) and produces stimulation or depression of the nervous system (Silva et al., 2002). Lead, another divalent metal ion might have similar toxic effects. Lead interferes with processes like learning and memory (Needleman, 1979; Needleman et al., 1979; Needleman and Bellinger, 1991). Furthermore, the SnCl2-induced effect on the blood system as described by Chmielnicka et al. (1993) has the same mechanism of toxicity that lead has. The toxic effect of lead has been partly related to changes of the [Ca2+]i (Bu¨sselberg et al., 1994; Bu¨sselberg, 2004; Florea and Bu¨sselberg, 2005), but this is definitely not the only explanation of lead toxicity as it has been shown to interact with other critical pathways as well (e.g. it activates PKC, calmodulin CAMKII and synaptogamin). In this study we found that SnCl2 causes a reduction of ICa(V) depending on the concentration, which could be at least partly responsible for the toxic effects caused by SnCl2. Acknowledgements We thank Frank Splettstoesser and Kirsten Go¨pelt for supporting our work with great enthusiasm and excellent technical assistance and Dr. W. Michael King for critical reading of the manuscript. References Agha NH, Al-Hilli M, Karim HM, Al-Hissoni MH, Jassim MN. Instant 99mTc-labelled glucoheptonate kit for kidney imaging. Nuklearmedizin 1983;22(5):246–50. Assis ML, Caceres MR, De Mattos JC, Caldeira-de-Arau´jo A, Bernardo-Filho M. Cellular inactivation induced by a radiopharmaceutical kit: role of stannous chloride. Toxicol Lett 1998a;99(3):199–205. Assis ML, Neto JB, Souza JE, Caldeira-de-Arau´jo A, Bernardo-Filho M. Stannous chloride and the glucoheptonic acid effect: study of a kit used in nuclear medicine. Cancer Lett 1998b;130(1–2):127–31. Berridge MJ. Elementary and global aspects of calcium signalling. J Physiol 1997; 499(2):291–306. Bu¨sselberg D, Platt B, Michael D, Carpenter DO, Haas HL. Mammalian voltage-activated calcium channel currents are blocked by Pb2+, Zn2+, and Al3+. J Neurophysiol 1994;71(4):1491–7. Bu¨sselberg D. Calcium channels as target sites of heavy metals. Toxicol Lett 1995;82/ 83:255–62. Bu¨sselberg D, Schirrmacher K, Domann R, Wiemann M. Lead interferes with calcium entry through membrane pores. Fresenius J Anal Chem 1998;361(4):372–6. Bu¨sselberg D. Actions of metals on membrane channels, calcium homeostasis and synaptic plasticity. In: Hirner AV, Emons H, editors. Organometal and metalloid specism in the environment: analysis, distribution, processes and toxicological evaluation. Wien, New York: Springer; 2004. p. 259–81.
963
Chiba M, Ogihara K, Inaba Y, Nishima T, Kikuchi M. The organ distribution of tin and the effect of tin on concentrations of several essential elements in rabbit. Toxicology 1984;31(1):23–32. Chmielnicka J, Zareba G, Grabowska U. Protective effect of zinc on heme biosynthesis disturbances in rabbits after administration per os of tin. Ecotoxicol Environ Saf 1992;24(3):266–74. Chmielnicka J, Zareba G, Polkowska-Kulesza E, Najder M, Korycka A. Comparison of tin and lead toxic action on erythropoietic system in blood and bone marrow of rabbits. Biol Trace Elem Res 1993;36(1):73–87. Chmielnicka J, Zareba G, Nasiadek M. Combined effect of tin and lead on heme biosynthesis in rats. Ecotoxicol Environ Saf 1994;29(2):165–73. Dantas FJ, Moraes Mo. de Mattos JC, Bezerra RJ, Carvalho EF, Filho MB, et al. Stannous chloride mediates single strand breaks in plasmid DNA through reactive oxygen species formation. Toxicol Lett 1999;110(3):129–36. Florea A-M, Bu¨sselberg D. Toxic effects of metals: modulation of intracellular calcium homeostasis. Materialwiss Werkst 2005;36:757–60. Florea A-M, Bu¨sselberg D. Occurrence, use and potential toxic effects of metals and metal compounds. Biometals 2006;19:419–27. Fritsch P, de saint-Blanquat G, Derache R. Nutritional and toxicological impacts of administered inorganic tin for six months in rats (author’s trans.). Toxicol Eur Res 1978;1(4):253–61. Garcı´a AG, Garcı´a-De-Diego AM, Gandia L, Borges R, Garcı´a-Sancho J. Calcium signalling and exocytosis in adrenal chromaffin cells. Physiol Rev 2006;86:1093–131. Gregg RW, Molepo JM, Monpetit VA, Mikael NZ, Redmond D, Gadia M, et al. Cisplatin neurotoxicity: the relationship between dosage, time, and platinum concentration in neurologic tissues, and morphologic evidence of toxicity. J Clin Oncol 1992; 10(5):795–803. Gyo¨ri J, Platoshyn O, Carpenter DO, Sala´nki J. Effect of inorganic and organic tin compounds on ACh- and voltage-activated Na currents. Cell Mol Neurobiol 2000;20(5):591–604. Hamilton El, Minski MJ, Cleary JY. The concentration and distribution of some stable elements in healthy human tissues from the United Kingdom. An environmental study. Sci Total Environ 1972/1973;1:341–74. Hattori T, Maehashi H. Stannous chloride-induced increase in calcium entry into motor nerve terminals of the frog. Eur J Pharmacol 1989;166(3):527–30. Hattori T, Maehashi H. Evidence for enhancement of calcium inward current by stannous chloride at frog motor nerve terminals. Res Commun Chem Pathol Pharmacol 1990;69(3):369–72. Hattori T, Maehashi H. Activation of N-type calcium channels by stannous chloride at frog motor nerve terminals. Res Commun Chem Pathol Pharmacol 1991;74(1): 125–8. Hattori T, Maehashi H. Interaction between stannous chloride and calcium channel blockers in frog neuromuscular transmission. Res Commun Chem Pathol Pharmacol 1992;75(2):243–6. Hattori T, Maehashi H. Facilitation of transmitter release from mouse motor nerve terminals by stannous chloride. Res Commun Chem Pathol Pharmacol 1993;82(1): 121–4. Hattori T, Maehashi H, Miyazawa T, Naito M. Potentiation by stannous chloride of calcium entry into osteoblastic MC3T3-E1 cells through voltage-dependent L-type calcium channels. Cell Calcium 2001;30(1):67–72. Hayashi R, Shima S, Taniwaki H, Nagaoka K, Kurita H. Comparative study on placental transfer and maternal distribution of organic and inorganic tin following oral administration in pregnant rats. Nippon Eiseigaku Zasshi 1994;49(5):869–76. Levesque PC, hare MF, Atchison WD. Inhibition of mitochondrial Ca2+ release diminishes the effectiveness of methyl mercury to release acetylcholine from synaptosomes. Toxicol Appl Pharmacol 1992;115(1):11–20. Needleman HL. Lead levels and children’s psychologic performance [letter]. N Engl J Med 1979;301:163. Needleman HL, Gunnoe C, Leviton A, Reed R, Persie H, Maher C, et al. Deficits in psychologic and classroom performance of children with elevated dentine lead levels. N Engl J Med 1979;300:689–95. Needleman HL, Bellinger D. The health effects of low level exposure to lead. Annu Rev Public Health 1991;12:111–40. Salanki J, Gyeri F, Platokhin A, Rozsa KS. Neurotoxicity of the environmental pollutants: tin (Sn2+) effects on currents induced by acetylcholine in the snail neurons. Ross Fiziol Zh Im I M Sechenova 1998;84(10):1061–73. Sala´nki Y, D’eri Y, Platokhin A, Sh-Ro´zsa K. The neurotoxicity of environmental pollutants: the effects of tin (Sn2+) on acetylcholine-induced currents in greater pond snail neurons. Neurosci Behav Physiol 2000;30(1):63–73. Savolainen H, Valkonen S. Dose-dependent brain tin concentration in rats given stannous chloride in drinking water. Toxicol Lett 1986;30(1):35–9. Silva CR, Oliveira MB, Melo SF, Dantas FJ, de Mattos JC, Bezerra RJ, et al. Biological effects of stannous chloride, a substance that can produce stimulation or depression of the central nervous system. Brain Res Bull 2002;59(3):213–6. Tomaszewski A, Bu¨sselberg D. Cisplatin modulates voltage gated channel currents of dorsal root ganglion neurons of rats. Neurotoxicology 2007;28:49–58. Yamaguchi M, Sugii K, Okada S. Tin decreases femoral calcium independently of calcium homeostasis in rats. Toxicol Lett 1982;10(1):7–10. Zareba G, Chmielnicka J. Effects of tin and lead on organ levels of essential minerals in rabbits. Biol Trace Elm Res 1989;20(3):233–42.
Contents lists available at ScienceDirect
NeuroToxicology
SnCl2 reduces voltage-activated calcium channel currents of dorsal root ganglion neurons of rats Anke Tomaszewski a, Dietrich Bu¨sselberg b,* a b
Institut fu¨r Physiologie, Universita¨tsklinikum Essen, Universita¨t Duisburg-Essen, Hufelandstrasse 55, 45122 Essen, Germany Texas Tech University Health Sciences Center, Paul L. Foster School of Medicine, Department of Medical Education, 4800 Alberta Avenue, El Paso, TX 79905, USA
A R T I C L E I N F O
A B S T R A C T
Article history: Received 15 January 2008 Accepted 12 February 2008 Available online 17 February 2008
Stannous dichloride (SnCl2) occurs in the environment where it has been especially enriched in aquatic ecosystems. Furthermore, it is used in food manufacturing (e.g. for stabilizing soft drinks or as an anticorrosive substance) and in nuclear medicine where it is employed as a reducing agent for technecium99m (99mTc) and therefore is applied intravenously to human beings. SnCl2 is known to have toxic effects on the nervous system which can be related to alterations of intracellular calcium homeostasis ([Ca2+]i). In this study the whole cell patch clamp technique is used on dorsal root ganglion neurons of 3-week-old ‘‘Wistar’’ rats to evaluate the effects of SnCl2 on voltageactivated calcium channel currents (ICa(V)). ICa(V) were reduced concentration-dependently by SnCl2 (1–50 mM). 1 mM SnCl2 reduced ICa(V) by 8.1 4.5% (peak current) and 19.2 8.9% (sustained current), whereas 50 mM inhibited ICa(V) by 50.6 4.3% (peak current) and 55.6 11.3% (sustained current). Sustained currents were slightly but not significantly more reduced than peak currents. The effect appeared not to be reversible. The threshold concentration was below 1 mM. The current–voltage relation did not shift which is an indication that different calcium channel subtypes were equally affected. There was a slight but not significant shift of the activation/inactivation curves towards the depolarizing direction. We conclude that voltage-gated calcium channels are affected by Sn2+ similarly to other divalent metal cations (e.g. Pb2+ or Zn2+). The reduction of ICa(V) could be related to the neurotoxic effects of SnCl2. ß 2008 Elsevier Inc. All rights reserved.
Keywords: Voltage-activated calcium channel currents DRG Heavy metal SnCl2
1. Introduction Stannous chloride (SnCl2) is widely used in food manufacturing processes (e.g. as a stabilizer in soft drinks or as a protector against corrosive processes in cans) (Assis et al., 1998a; Dantas et al., 1999) and it has been enriched especially in aquatic ecosystems (Sala´nki et al., 2000; Gyo¨ri et al., 2000). Furthermore, it is employed as a reducing agent in technecium-99m (99mTc), a substance used in nuclear medicine for brain and renal scintigraphies and is therefore applied intravenously to humans (Agha et al., 1983; Assis et al.,
* Corresponding author. Tel.: +1 915 783 1700x251. E-mail address: [email protected] (D. Bu¨sselberg). Abbreviations: Ca2+, calcium ions; [Ca2+]i, intracellular calcium concentration; DRG, dorsal root ganglion neurons; Hg2+, mercury; ICa(V), voltage-activated calcium channel currents; INa(V), voltage-activated sodium channel currents; IV, current– voltage relation; 99mTc, technecium-99m; Pb2+, lead; SnCl2, stannous chloride, tin chloride; VACCs, voltage-activated calcium channels; Zn2+, zinc. 0161-813X/$ – see front matter ß 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.neuro.2008.02.003
1998a,b; Dantas et al., 1999; Silva et al., 2002). Without any special exposure blood tin concentrations in humans are around 2–9 mg/l (17–76 mmol/l; Hamilton et al., 1972/1973). When administered to the body it accumulates and its highest concentrations are found in the kidney (15.0 1.64 mg/g wet weight; Chiba et al., 1984). After oral application of SnCl2 (a cumulative dose of 17.7 mmol/kg body weight was applied) to rats, blood serum levels of 16–60 pmol/g (16–60 nmol/l) were found (Savolainen and Valkonen, 1986). As a heavy metal SnCl2 has numerous toxic effects. In addition to its known genotoxic potential (effects occurred at concentrations of 12.5–200 mg/ml SnCl2 (Dantas et al., 1999) or 1.1 10 4 to 4.4 10 4 M SnCl2 (Assis et al., 1998a,b; Dantas et al., 1999; Silva et al., 2002)) it induces cellular inactivation (Assis et al., 1998a), inhibits the immune response and is speculated to play a role in tumor generation in the thyroid gland (Silva et al., 2002). Moreover, after intoxication with SnCl2, hematocrit and haemoglobin as well as serum iron are diminished (Fritsch et al., 1978) and heme biosynthesis is compromised (effects occur after oral application of 10–200 mg Sn/kg) (Chmielnicka et al., 1992, 1994).
A. Tomaszewski, D. Bu¨sselberg / NeuroToxicology 29 (2008) 958–963
Stannous chloride also affects the nervous system. By generating reactive oxygen species it causes oxidative stress, involved in both aging and neurological diseases (Silva et al., 2002). Furthermore behavioral abnormalities which have a possible neuronal basis have been observed in living organisms intoxicated with SnCl2 (Gyo¨ri et al., 2000), and it has been shown that SnCl2 produces stimulation or depression of the central nervous system (Silva et al., 2002). In addition, the release of neurotransmitters in nerve terminals is accelerated by SnCl2 and thus neuromuscular transmission is facilitated (Silva et al., 2002). Further investigation revealed that SnCl2 causes a decrease of the acetylcholine-induced inward current of neurons of mollusc Lymnaea stagnalis L. (Gyo¨ri et al., 2000; Sala´nki et al., 2000) which is a direct action upon the neuronal membrane, is important in modulating synaptic transmission and facilitates the release of neurotransmitters from nerve terminals (Hattori and Maehashi, 1993; Salanki et al., 1998). This last effect of SnCl2 has been related to a modulation of the calcium homeostasis by Hattori and Maehashi (1989). Calcium ions (Ca2+) are present in the intra- and extra-cellular space and their concentration is strictly regulated (Berridge, 1997; Garcı´a et al., 2006). Changes of the intracellular calcium concentration ([Ca2+]i) are necessary in processes like muscle contraction, cell proliferation, activation of oncogenes and apoptosis (Berridge, 1997; Garcı´a et al., 2006). Neuronal cells are highly sensitive to slight changes of the [Ca2+]i and thus are able to use these modifications of [Ca2+]i as a second messenger and a means of signal transmission (Garcı´a et al., 2006). For example calcium entry into the cell triggers neurotransmitter release at the synaptic membrane. While slight modifications of the [Ca2+]i are physiological, greater changes of the [Ca2+]i may result in malfunction of the cell and lead to acute or chronic cell injuries or cell death (Berridge, 1997). There are several ways for a cell to regulate the [Ca2+]i. Either Ca2+ is removed from the cell via active transport mechanisms or it is moved to internal Ca2+ stores. Both mechanisms result in a decrease of the [Ca2+]i. In order to increase the [Ca2+]i, Ca2+ can be released from internal stores or enter the cell via Ca2+ selective pores located in the cell membrane. These can be receptormediated channels like the NMDA-receptor channel complex or voltage-activated channels. Therefore, it is possible that a malfunction of voltage-activated calcium channels results in changes of [Ca2+]i which may lead to a disruption of intracellular signalling and thus to cellular dysfunction (Bu¨sselberg, 2004). SnCl2 and its effect on [Ca2+]i have been investigated by Hattori and colleagues who found an increase of the [Ca2+]i which was related to SnCl2. They conclude that voltage-activated calcium channels (VACCs) (N-type (Hattori and Maehashi, 1992) or L-type (Hattori et al., 2001)) are involved, yet as far as we know modulations of currents through voltage-activated calcium channel currents (ICa(V)) by SnCl2 have never been studied directly. Divalent metal ions like lead (Pb2+), zinc (Zn2+), mercury (Hg2+), methyl mercury and trimethyl lead are known to reduce the current of voltage-activated ion channels of mammalian neurons. This is considered to be a possible explanation of the cognitive dysfunctions caused by heavy metals (Bu¨sselberg, 1995, 2004; Bu¨sselberg et al., 1994; Florea and Bu¨sselberg, 2005, 2006). Furthermore, Needleman and colleagues showed that metal ions like Pb2+ alter cognitive functions like learning and memory directly (Needleman, 1979; Needleman et al., 1979; Needleman and Bellinger, 1991). As SnCl2, another divalent metal ion is neurotoxic; too, it is likely to cause similar effects on neuronal cells which might e.g. alter the current of voltage-activated channels, too. Dorsal root ganglion neurons (DRG) are ‘‘strategically well placed’’ to cause a wide range of neuronal deficits when harmed, as
959
a malfunction can affect cellular metabolism as well as axonal transport (Gregg et al., 1992). In addition, they are not protected by the blood–brain-barrier. This study was designed to evaluate the effect of SnCl2 on VACCs of neuronal cells. Therefore, we have examined its effect on (ICa(V)) of dorsal root ganglion neurons of rats using the whole cell patch clamp method. 2. Methods 2.1. Cell culture preparation Short-term cell cultures (cells were kept in culture for a period less than 24 h) of dorsal root ganglion neurons from 3-week-old ‘Wistar’ rats were prepared as described previously (Tomaszewski and Bu¨sselberg, 2007). Shortly after anesthetizing the animals with approximately 5 ml isoflurane (CuraMed Pharma GmbH, Karlsruhe, Germany) and in the absence of pain reflexes, the rats were decapitated; the spinal column was removed and opened from the dorsal side. Thereafter the spinal cord was taken out; the DRGs were collected and placed in ice cooled F12 media (Sigma, Taufkirchen, Germany). After cutting off the spinal nerves and opening the ganglion capsules under optical control, the ganglia were transferred into a medium containing 0.9 ml F12 and 0.1 ml collagenase (2612.5 U ml 1 Sigma Type II) to be incubated for a period of 45 min in a humidified atmosphere containing 5% CO2 at a temperature of 37 8C. The ganglia were washed three times in 1.5 ml F12 medium to remove the collagenase. Afterwards they were trypsinized (2525 U Trypsin per ml F12 medium, Sigma Type IX) for 2 min under the same conditions. Trypsin was removed by washing the cells three times with F12 medium leaving a final volume of 0.7 ml. By triturating, using a fire-polished Pasteur pipette (tip diameter 150 mM) the neurons were released from the capsules. 50 ml of this suspension were placed in the middle of each Petri dish (3 cm, ‘Easy Grip’, Falcon, Gra¨tling-Lochau, Germany), and the cells were incubated for at least 2 h in which time they adhered to the substrate. One ml F12 medium with 10% horse serum and nerve growth factor was then added to each dish. Afterwards the cells were put into the incubator until they were used the next day. 2.2. Recording of voltage-activated calcium channel currents (ICa(V)) The whole cell patch clamp technique was used to record voltage-activated channel currents of DRG neurons. For this a HEKA EPC9 amplifier with the ‘‘EPC screen’’ software (HEKA Instruments, Port Washington, NY, USA) was used. Microelectrodes, consisting of borosilicate glass with filament (o.d.: 1.5 mm and i.d.: 0.7 mm; Biomedical Instruments (BMI), Zo¨llnitz, Germany), were pulled with an electrode puller (Sutter, model P-87, Sutter Instruments, Navato, USA or a DMZ-Universal Puller, Zeitzlnstrumente GmbH Mu¨nchen). Electrodes were fire polished to a final resistance of 3–5 MV using a Narashige microforge (MF-830, Narashige Instrument Laboratory, Tokyo, Japan). An external solution (containing 1 mM MgCl2, 10 mM HEPES, 0.001 mM TTX, 10 mM Glucose, 10 mM BaCl2 and 130 mM TEA-chloride with a pH of 7.3) to separate voltage-activated calcium channel currents was prepared and glass electrodes were filled with the adequate internal solution (consisting of 140 mM CsCl, 4mM MgCl2, 10 mM HEPES, 10 mM EGTA and 2 mM Na-ATP with an pH of 7.2). Both (the external as well as the internal) solutions are nominally calcium free and barium ions were used as charge carriers. After the formation of a giga-V seal and establishment of the whole cell configuration the cells were clamped at 80 mV. By depolarizing the cells to 0 mV for 75 ms, voltage-activated calcium channel
960
A. Tomaszewski, D. Bu¨sselberg / NeuroToxicology 29 (2008) 958–963
currents (ICa(V)) were elicited. Depolarization steps were started at 60 mV and the depolarizations were increased in intervals of 10 mV until a maximum depolarization of 40 mV was reached in order to record a current–voltage relation (IV) of ICa(V) (inset in Fig. 2). Activation and inactivation curves were generated by firstly hyperpolarizing the cell to 90 mV for a period of 75 ms, secondly depolarizing the cell to 80 mV for 75 ms and thirdly depolarizing to 0 mV for 75 ms before returning to the resting membrane potential of 80 mV. The second depolarization step was increased by +10 mV until 0 mV was reached (insets Fig. 3). All data were sampled and stored on hard disc (sample rate: 25 data points/ms). Tin chloride (SnCl2) (Tin (II) chloride dihydrate; Sigma–Aldrich, Steinem, Germany) was freshly dissolved (from a stock solution of 1 mM) in the external solution and was applied by using a bath application system with a continuous flow (flow rate: 5 ml/min). 2.3. Data analysis All currents were leak corrected using a p/3 protocol, where the neuron was hyperpolarized three times one third of the proposed voltage depolarization. The resulting currents were added and the sum was subtracted from the current which was elicited by the
depolarization. This guarantees that only the voltage-activated currents (but not passive ‘‘leak’’-currents) are taken into account. Peak currents were taken during the first 10 ms of each depolarization whereas sustained currents were measured between 60 and 70 ms of each depolarization. A minimum of five single successive peak/sustained currents ( 80 to 0 mV) which were taken under control conditions were averaged. All successive currents were expressed as a percentage of these currents to allow a comparison of single experiments. A ‘‘steady state’’ was defined as a situation in which the peak current amplitude did not decline for more than 5% within 40 s. Before calculating time courses, current–voltage relationships or concentration response curves, the currents were corrected for linear rundown. Therefore, the average reduction of the current within a defined time interval was estimated under control conditions and the calculated time-dependent rundown was added to each current. To calculate IV relationships, currents were expressed as a percentage of the maximum current of the control IV. The inactivation curves were calculated by defining the degree of depolarization at which the smallest current (either under control conditions or after SnCl2 was applied) was registered and the cell was completely inactivated as 100%. All following currents are expressed as reduction in percent of this current.
Fig. 1. Raw traces of ICa(V) generated by depolarizing the cell from the resting potential of 80 mV to 0 mV. Currents are shown before (black line) and after (grey line) the application of SnCl2 in the concentrations of (A) 1 mM, (C) 5 mM and (E) 10 mM. (B, D and F) show the reduction of ICa(V) peak currents (black diamonds) and sustained currents (white diamonds) caused by SnCl2 in the course of time for SnCl2 1 mM (B), 5 mM (D) and 10 mM (F). The grey line indicates the time during which SnCl2 has been applied. Data are shown with S.D.
A. Tomaszewski, D. Bu¨sselberg / NeuroToxicology 29 (2008) 958–963
961
Activation curves were created following a similar procedure. The degree of depolarization resulting in the largest current (under control conditions or after the application of SnCl2) was elicited and defined as 100%. All other currents are expressed as a percentage of this current. Data are given as means S.D. ‘‘P-values’’ were calculated with student’s t-test (double sided, unpaired or paired as required, type 1) (***P < 0.001; **P < 0.005; *P < 0.01) using Excel software. 3. Results Voltage-activated calcium channel currents were generated by depolarizing the neuron from the resting potential of 80 mV to 0 mV. A total of 33 neurons were included in this study. SnCl2 (10 mM) was applied to 22 cells, 5 cells (50 mM), 3 cells (5 mM) and 3 cells (1 mM). ICa(V) were sensitive to the application of SnCl2. Parts A, C and E of Fig. 1 represent raw traces of ICa(V) before and after the application of SnCl2. The peak currents of 1 nA (A), 800 pA (C) or 800–900 pA (E) were elicited 3 1 ms (A), 8 3 ms (C) or 5 2 ms (E) after starting the depolarization of the cell. The currents taken under control conditions (black line) declined by up to one third of the peak current during the 75 ms of depolarization. The application of SnCl2 in the concentrations of A, 1 mM; C, 5 mM; E, 10 mM resulted in a reduction of the ICa(V) (grey line). After the application of 1 mM SnCl2 the ICa(V) was comparable to control (8.1 4.5% reduction (peak current) and 19.2 8.9% (sustained current), respectively) as illustrated in Fig. 1B. A steady state was reached after 200 s. SnCl2 (5 mM) reduced the peak (sustained) currents significantly (P = 0.009, * (peak); P = 0.004, ** (sustained)) by
Fig. 2. (A) Current–voltage relationship of ICa(V) peak currents before (black diamonds) and after (white diamonds) the application of SnCl2 10 mM. Currents are expressed as a percentage of the maximum current of the control IV. Data are shown with S.D. N = 9. Inset: current–voltage relationship. Depolarization steps were started at 60 mV and increased by 10 mV (indicated by the broken lines) until a maximal depolarization of +40 mV was reached. Each depolarization lasted 75 ms. (B) Reduction of ICa(V) peak currents taken after the application of 10 mM SnCl2 in percent of the control IV.
Fig. 3. Activation (black) and inactivation (grey) curves taken before (solid line) and after (broken line) the application of SnCl2 10 mM. Currents are expressed as a percentage of the maximum current of the control IV. Data are shown with S.D. N = 8 (activation curves); N = 6 (inactivation curves). Insets: Waveforms of the recording protocol used to measure activation/inactivation curves (a) and IV relationships (b). A: voltage-activated channel currents were elicited by hyperpolarizing the cell to 90 mV. The second depolarization step was increased (indicated by the broken lines) by 10 mV until a final depolarization to 0 mV was reached. Each depolarization/hyperpolarization was carried out for the duration of 75 ms.
15.5 5.4% (23.8 5.5%) (Fig. 1D). The effect of SnCl2 (10 mM) was even more pronounced. Peak (sustained) currents were reduced significantly (P = 1.8 10 8, *** (peak); P = 1.2 10 7; *** (sustained)) by 27.9 6.7% (40.3 8.5%). A steady state was reached after 200 s. Since after eliminating SnCl2 from the bath solution there was hardly any recovery, the effect of SnCl2 on ICa(V) appeared not to be reversible (Fig. 1F). Gyo¨ri et al. (2000) tested the effect of SnCl2 on acetylcholineinduced and voltage-dependent inward currents of the mollusc, L. stagnalis L. In addition, they found a slight shift to the left of the current–voltage relation of voltage-activated sodium channels caused by SnCl2 (10 mM). To estimate whether ICa(V) are effected similarly by SnCl2 we measured the current–voltage dependency of ICa(V) before (Fig. 2A, black diamonds) and after (white diamonds) the application of SnCl2 (10 mM). The reduction of the current after the application of SnCl2 expressed as a percentage of the current measured under control conditions is shown in part B of Fig. 2. ICa(V) were reduced equally by 30–40% over the whole voltage range tested. There was no indication of a shift of the current–voltage relation. To test whether the activation or the inactivation of voltagegated calcium channels are influenced by SnCl2 (10 mM), activation as well as inactivation protocols were applied. This is illustrated in Fig. 3 where the activation (black) and the inactivation (grey) curves of VACCs before (solid line) and after (broken line) the
Fig. 4. Concentration dependency of ICa(V) peak currents (black diamonds) and sustained currents (white diamonds) after the application of SnCl2 in the concentrations of 1, 5, 10 and 50 mM. P-values are indicated by asterisks (***P < 0.0001; **P < 0.005; *P < 0.01).
962
A. Tomaszewski, D. Bu¨sselberg / NeuroToxicology 29 (2008) 958–963
application of SnCl2 are shown. The degree of inactivation decreased with an increasing pre-depolarization of the neurons. With a pre-depolarization of 30 mV 25% of the channels were inactivated and with a pre-depolarization of 20 mV the VACCs were not inactivated. Voltage-activated calcium channels were 50% activated at a depolarization to 30 mV under control conditions while at 20 mV a maximal depolarization was observed. After SnCl2 (10 mM) was applied, the activation as well as the inactivation curves were shifted slightly, but not significantly to more positive depolarization steps. The concentration dependency of the effect of SnCl2 on ICa(V) is shown in Fig. 4. As 1 mM SnCl2 reduced the ICa(V) peak current by 8.1 4.5% the threshold concentration is about 1 mM. The highest concentration tested in this study (50 mM) reduced the ICa(V) significantly (P = 0.0008; *** (peak), P = 0.0009; *** (sustained)) by 50.6 4.3% (peak) and by 55.6 11.3% (sustained), respectively. As after either oral or intraperitoneal admissions of SnCl2 or lead to rats the blood levels of SnCl2 have been less than or 1/10 of the blood lead levels (Zareba and Chmielnicka, 1989) and as lead causes a total blockage of ICa(V) at concentrations about 1 mM (Bu¨sselberg et al., 1994) we abstained from testing higher concentrations of SnCl2. Although the sustained current of all concentrations tested was more reduced than the peak current this difference was not significant for any of the concentrations tested. 4. Discussion We found that voltage-activated calcium channel currents were reduced by SnCl2 depending on the concentration applied. The threshold concentration was about 1 mM and 50 mM SnCl2 reduced the ICa(V) peak currents by 50.6 4.3%. This effect appeared to be irreversible in the course of our experiments. Seemingly contradictory to our findings, earlier research has revealed that SnCl2 results in an increased [Ca2+]i (Hattori and Maehashi, 1989, 1990; Hattori et al., 2001). As mentioned above there are two possibilities for a cell to raise its [Ca2+]i. Calcium could enter from the extracellular space or be released from internal stores. Hattori and Maehashi (1989, 1990) studied extracellularly recorded potentials of motor nerve terminals of the bullfrog. They found that apart from raising the [Ca2+]i, SnCl2 increased the prolonged positive deflection of this potential, which they suspect is related to calcium entry. By studying the interactions of SnCl2 with calcium channel blockers they found that the SnCl2-induced increase of the [Ca2+]i can be reduced by specific calcium channel blockers. As CdCl2 or v-conotoxin proved to be more effective than NiCl2 or nifedipine they concluded that this calcium entry into motor nerve terminals is linked to ICa(V) through N-type calcium channels. Furthermore, after proving that SnCl2 has similar effects on amphibians as on mammals (Hattori and Maehashi, 1993), Hattori et al. (2001) studied the effect of SnCl2 on the [Ca2+]i of osteoblastic MC3T3-E1 cells. They found, that in combination with a high extracellular concentration of potassium (10–100 mM) the [Ca2+]i was increased by SnCl2 (100 mM). This increase was inhibited by blockers of the L-type VACC (nicardipine, verapamil, and diltiazem) but not by N-type blockers (v-conotoxin GVIA). Further investigation revealed that SnCl2 applied without a high extracellular concentration of potassium (10–100 mM) did not change the [Ca2+]i while the rise of [Ca2+]i was inhibited by using a Ca2+-free medium or nifedipine. Hattori et al. (2001) conclude that calcium enters the cells through VACCs of the L-type although they also offer another possible explanation of the increase of the [Ca2+]i in suggesting that this increase is due to a release of Ca2+ from internal Ca2+ stores. A metal-triggered Ca2+ release from the calcium stores is also described for other metal compounds, e.g. Levesque et al. (1992),
who studied the effect of methyl mercury on nerve terminal mitochondria. They found that Ca2+ is released from internal stores under the influence of the heavy metal. In the light of our findings this seems to be the more plausible explanation and could explain the contradictory results. Calcium channels can be inhibited by either calcium or modulations of the membrane potential. In this study short depolarization pulses (75 ms) were used with Ba2+ as a charge carrier which minimizes the calcium-induced calcium channel inactivation as well as the voltage-dependent calcium channel inactivation. If, as is shown in this study, the VACCs are inhibited by SnCl2 but the calcium-induced calcium channel inactivation is prevented similarly, the [Ca2+]i could still be increased despite the inhibition of VACCs. In this case Ca2+ would not enter the cell through ICa(V) but Ca2+ influx would follow long-duration depolarizing pulses. As SnCl2 increased the prolonged positive deflection of extracellularly recorded potentials of motor nerve terminals of the bullfrog (Hattori and Maehashi, 1989, 1990) this could explain the contradictory results. On the contrary, the findings that the increase of [Ca2+]i can be inhibited by either a Ca2+-free medium or by blockers of calcium channels indicate that calcium enters the cell from the extracellular space. Yet in our study we found a clear reduction of ICa(V) after the application of SnCl2. One possible explanation of these differences is that Ca2+ enters the cell from the extracellular space not through voltage-activated channels but different calcium selective channels. Unfortunately this is not in agreement with the finding that the co-application of highly specific L- and N-type channel blockers inhibits the increase of the [Ca2+]i. Another possibility to explain the contradictory effects caused by SnCl2 is that Ca2+ enters the cell via voltage-activated calcium channels but not through the L or T-type. In our study the sustained currents of ICa(V) were slightly (but not significantly) more reduced by SnCl2 than peak currents of ICa(V). In addition to that the largest current elicited in the current–voltage relations both under control conditions and after the application of SnCl2 was generated by a depolarization from 80 to 20 or 10 mV. There was no indication of a shift of the current–voltage relation. Furthermore, there was no significant difference in activation/inactivation curves taken before and after the application of SnCl2. Therefore, we have no proof that sub-types (e.g. L-, N-or T-type) of VACCs are influenced differently as was proposed by Hattori and Maehashi (1990, 1991) and Hattori et al. (2001). Nevertheless, not all studies on SnCl2 are contradictory to our findings. Gyo¨ri et al. (2000), who compared the effect of SnCl2 on voltage-activated sodium channel currents (INa(V)) of the mollusc L. stagnalis L. to acetylcholine-induced currents, found that these currents were reduced by SnCl2 with an effective threshold concentration of 0.1 mM. Furthermore, there is a possibility that SnCl2 causes distinct effects in different concentrations. Hattori et al. (2001) applied SnCl2 in relatively high concentrations (about 0.1 mM) and found an increase of the [Ca2+]i, while we used SnCl2 in concentrations between 1 and 50 mM which reduced the calcium entry through ICa(V) and therefore should lower the [Ca2+]i. Moreover, it has been shown, that other divalent metal ions such as Zn2+ or Pb2+ also reduce voltage-activated channel currents (Bu¨sselberg et al., 1994, 1998; Bu¨sselberg, 1995). Apart from Pb2+, Zn2+ has been found to block the ICa(V) completely (more than 80%), but both metals differ in their effective concentration range: while 1 mM of Pb2+ was sufficient, 150–200 mM Zn2+ were needed to accomplish the same effects. The threshold concentration for Pb2+ was even below 0.1 mM (Pb2+). The hill coefficient of both metals was about 1 which suggests that the ICa(V) were directly blocked (Bu¨sselberg et al., 1994). As SnCl2 is also a divalent metal ion which affects human beings in a similar way as Pb2+ and Zn2+, it is
A. Tomaszewski, D. Bu¨sselberg / NeuroToxicology 29 (2008) 958–963
possible that it follows a similar mechanism of reaction at the channel side. Therefore, a reduction of ICa(V) by SnCl2 is reasonable since as a divalent cation it also could compete with Ca2+ at the calcium binding sites at or within any calcium conducting pore. 4.1. Toxicity of SnCl2 As SnCl2 occurs in our environment (especially in aquatic ecosystems (Sala´nki et al., 2000; Gyo¨ri et al., 2000)), is used in food manufacturing, and applied intravenously during radiological diagnostic procedures, human beings are likely to have contact with SnCl2. In various studies it has been shown that SnCl2 can be absorbed by mammal after oral as well as after intraperitoneal administration. It accumulates in internal organs such as brain, liver, bones and kidney in which the highest concentration (15.0 1.6 mg/ g wet weight after the administration of 5 mM SnCl2/(kg day) for a period of 6 days (Chiba et al., 1984)) is found (Chiba et al., 1984; Fritsch et al., 1978; Hayashi et al., 1994; Savolainen and Valkonen, 1986; Yamaguchi et al., 1982; Zareba and Chmielnicka, 1989). As described above SnCl2 causes behavioural abnormalities (Gyo¨ri et al., 2000) and produces stimulation or depression of the nervous system (Silva et al., 2002). Lead, another divalent metal ion might have similar toxic effects. Lead interferes with processes like learning and memory (Needleman, 1979; Needleman et al., 1979; Needleman and Bellinger, 1991). Furthermore, the SnCl2-induced effect on the blood system as described by Chmielnicka et al. (1993) has the same mechanism of toxicity that lead has. The toxic effect of lead has been partly related to changes of the [Ca2+]i (Bu¨sselberg et al., 1994; Bu¨sselberg, 2004; Florea and Bu¨sselberg, 2005), but this is definitely not the only explanation of lead toxicity as it has been shown to interact with other critical pathways as well (e.g. it activates PKC, calmodulin CAMKII and synaptogamin). In this study we found that SnCl2 causes a reduction of ICa(V) depending on the concentration, which could be at least partly responsible for the toxic effects caused by SnCl2. Acknowledgements We thank Frank Splettstoesser and Kirsten Go¨pelt for supporting our work with great enthusiasm and excellent technical assistance and Dr. W. Michael King for critical reading of the manuscript. References Agha NH, Al-Hilli M, Karim HM, Al-Hissoni MH, Jassim MN. Instant 99mTc-labelled glucoheptonate kit for kidney imaging. Nuklearmedizin 1983;22(5):246–50. Assis ML, Caceres MR, De Mattos JC, Caldeira-de-Arau´jo A, Bernardo-Filho M. Cellular inactivation induced by a radiopharmaceutical kit: role of stannous chloride. Toxicol Lett 1998a;99(3):199–205. Assis ML, Neto JB, Souza JE, Caldeira-de-Arau´jo A, Bernardo-Filho M. Stannous chloride and the glucoheptonic acid effect: study of a kit used in nuclear medicine. Cancer Lett 1998b;130(1–2):127–31. Berridge MJ. Elementary and global aspects of calcium signalling. J Physiol 1997; 499(2):291–306. Bu¨sselberg D, Platt B, Michael D, Carpenter DO, Haas HL. Mammalian voltage-activated calcium channel currents are blocked by Pb2+, Zn2+, and Al3+. J Neurophysiol 1994;71(4):1491–7. Bu¨sselberg D. Calcium channels as target sites of heavy metals. Toxicol Lett 1995;82/ 83:255–62. Bu¨sselberg D, Schirrmacher K, Domann R, Wiemann M. Lead interferes with calcium entry through membrane pores. Fresenius J Anal Chem 1998;361(4):372–6. Bu¨sselberg D. Actions of metals on membrane channels, calcium homeostasis and synaptic plasticity. In: Hirner AV, Emons H, editors. Organometal and metalloid specism in the environment: analysis, distribution, processes and toxicological evaluation. Wien, New York: Springer; 2004. p. 259–81.
963
Chiba M, Ogihara K, Inaba Y, Nishima T, Kikuchi M. The organ distribution of tin and the effect of tin on concentrations of several essential elements in rabbit. Toxicology 1984;31(1):23–32. Chmielnicka J, Zareba G, Grabowska U. Protective effect of zinc on heme biosynthesis disturbances in rabbits after administration per os of tin. Ecotoxicol Environ Saf 1992;24(3):266–74. Chmielnicka J, Zareba G, Polkowska-Kulesza E, Najder M, Korycka A. Comparison of tin and lead toxic action on erythropoietic system in blood and bone marrow of rabbits. Biol Trace Elem Res 1993;36(1):73–87. Chmielnicka J, Zareba G, Nasiadek M. Combined effect of tin and lead on heme biosynthesis in rats. Ecotoxicol Environ Saf 1994;29(2):165–73. Dantas FJ, Moraes Mo. de Mattos JC, Bezerra RJ, Carvalho EF, Filho MB, et al. Stannous chloride mediates single strand breaks in plasmid DNA through reactive oxygen species formation. Toxicol Lett 1999;110(3):129–36. Florea A-M, Bu¨sselberg D. Toxic effects of metals: modulation of intracellular calcium homeostasis. Materialwiss Werkst 2005;36:757–60. Florea A-M, Bu¨sselberg D. Occurrence, use and potential toxic effects of metals and metal compounds. Biometals 2006;19:419–27. Fritsch P, de saint-Blanquat G, Derache R. Nutritional and toxicological impacts of administered inorganic tin for six months in rats (author’s trans.). Toxicol Eur Res 1978;1(4):253–61. Garcı´a AG, Garcı´a-De-Diego AM, Gandia L, Borges R, Garcı´a-Sancho J. Calcium signalling and exocytosis in adrenal chromaffin cells. Physiol Rev 2006;86:1093–131. Gregg RW, Molepo JM, Monpetit VA, Mikael NZ, Redmond D, Gadia M, et al. Cisplatin neurotoxicity: the relationship between dosage, time, and platinum concentration in neurologic tissues, and morphologic evidence of toxicity. J Clin Oncol 1992; 10(5):795–803. Gyo¨ri J, Platoshyn O, Carpenter DO, Sala´nki J. Effect of inorganic and organic tin compounds on ACh- and voltage-activated Na currents. Cell Mol Neurobiol 2000;20(5):591–604. Hamilton El, Minski MJ, Cleary JY. The concentration and distribution of some stable elements in healthy human tissues from the United Kingdom. An environmental study. Sci Total Environ 1972/1973;1:341–74. Hattori T, Maehashi H. Stannous chloride-induced increase in calcium entry into motor nerve terminals of the frog. Eur J Pharmacol 1989;166(3):527–30. Hattori T, Maehashi H. Evidence for enhancement of calcium inward current by stannous chloride at frog motor nerve terminals. Res Commun Chem Pathol Pharmacol 1990;69(3):369–72. Hattori T, Maehashi H. Activation of N-type calcium channels by stannous chloride at frog motor nerve terminals. Res Commun Chem Pathol Pharmacol 1991;74(1): 125–8. Hattori T, Maehashi H. Interaction between stannous chloride and calcium channel blockers in frog neuromuscular transmission. Res Commun Chem Pathol Pharmacol 1992;75(2):243–6. Hattori T, Maehashi H. Facilitation of transmitter release from mouse motor nerve terminals by stannous chloride. Res Commun Chem Pathol Pharmacol 1993;82(1): 121–4. Hattori T, Maehashi H, Miyazawa T, Naito M. Potentiation by stannous chloride of calcium entry into osteoblastic MC3T3-E1 cells through voltage-dependent L-type calcium channels. Cell Calcium 2001;30(1):67–72. Hayashi R, Shima S, Taniwaki H, Nagaoka K, Kurita H. Comparative study on placental transfer and maternal distribution of organic and inorganic tin following oral administration in pregnant rats. Nippon Eiseigaku Zasshi 1994;49(5):869–76. Levesque PC, hare MF, Atchison WD. Inhibition of mitochondrial Ca2+ release diminishes the effectiveness of methyl mercury to release acetylcholine from synaptosomes. Toxicol Appl Pharmacol 1992;115(1):11–20. Needleman HL. Lead levels and children’s psychologic performance [letter]. N Engl J Med 1979;301:163. Needleman HL, Gunnoe C, Leviton A, Reed R, Persie H, Maher C, et al. Deficits in psychologic and classroom performance of children with elevated dentine lead levels. N Engl J Med 1979;300:689–95. Needleman HL, Bellinger D. The health effects of low level exposure to lead. Annu Rev Public Health 1991;12:111–40. Salanki J, Gyeri F, Platokhin A, Rozsa KS. Neurotoxicity of the environmental pollutants: tin (Sn2+) effects on currents induced by acetylcholine in the snail neurons. Ross Fiziol Zh Im I M Sechenova 1998;84(10):1061–73. Sala´nki Y, D’eri Y, Platokhin A, Sh-Ro´zsa K. The neurotoxicity of environmental pollutants: the effects of tin (Sn2+) on acetylcholine-induced currents in greater pond snail neurons. Neurosci Behav Physiol 2000;30(1):63–73. Savolainen H, Valkonen S. Dose-dependent brain tin concentration in rats given stannous chloride in drinking water. Toxicol Lett 1986;30(1):35–9. Silva CR, Oliveira MB, Melo SF, Dantas FJ, de Mattos JC, Bezerra RJ, et al. Biological effects of stannous chloride, a substance that can produce stimulation or depression of the central nervous system. Brain Res Bull 2002;59(3):213–6. Tomaszewski A, Bu¨sselberg D. Cisplatin modulates voltage gated channel currents of dorsal root ganglion neurons of rats. Neurotoxicology 2007;28:49–58. Yamaguchi M, Sugii K, Okada S. Tin decreases femoral calcium independently of calcium homeostasis in rats. Toxicol Lett 1982;10(1):7–10. Zareba G, Chmielnicka J. Effects of tin and lead on organ levels of essential minerals in rabbits. Biol Trace Elm Res 1989;20(3):233–42.