- Email: [email protected]
Resiniferatoxin-Induced Loss of Plasma Membrane in Vanilloid Receptor Expressing Cells
NeuroToxicology 24 (2003) 895–908
Resiniferatoxin-Induced Loss of Plasma Membrane in Vanilloid Receptor Expressing Cells Robert M. Caudle1,*, Laszlo Karai2, Narasaiah Mena1, Brian Y. Cooper1, Andrew J. Mannes2, Federico M. Perez1, Michael J. Iadarola2, Zoltan Olah2 1
Department of Oral and Maxillofacial Surgery, University of Florida College of Dentistry and the University of Florida McKnight Brain Institute, 1600 Archer Road, P.O. Box 100416, Gainesville, FL 32610, USA 2 Pain and Neurosensory Mechanisms Branch, National Institute of Dental and Craniofacial Research, National Institutes of Health, Bethesda, MD, USA Received 27 August 2002; accepted 29 May 2003
Abstract Resiniferatoxin (RTX), a potent analog of capsaicin, was evaluated electrophysiologically in dorsal root ganglion (DRG) cells and cell lines ectopically expressing the vanilloid receptor type 1 (VR1) to determine if cell phenotype influenced RTXs neurotoxic properties. Furthermore, capsaicin and heat activation of VR1 were evaluated in these cells to determine if cellular damage was unique to RTX activation of the receptors. RTX application to DRG cells identified as type 1, 2 or 5, cell types known to express VR1, induced large inward currents. RTX did not induce currents in DRG cells that do not express the receptor (type 4 cells). In cell lines ectopically expressing VR1, RTX-induced similar currents. RTX produced no effect in non-transfected cells. After exposure to RTX both DRG cells and transfected cells failed to respond to subsequent applications of the agonist. In addition, whole cell capacitance was reduced up to 70%. The decrease in capacitance was associated with the loss of plasma membrane, as determined by confocal microscopy. Cell phenotype, other than VR1 expression, did not influence the response to RTX. Interestingly, capsaicin and heat activation of vanilloid receptors also decreased cell capacitance, but the loss of membrane was not as great as with RTX and responses to these stimuli were not lost after the initial exposure. The loss of cell membrane required elevated intracellular levels of Ca2þ. From these data it was concluded that the loss of cell membrane was dependent on the presence of both VR1 and intracellular Ca2þ accumulation, but not on cell phenotype. # 2003 Elsevier Inc. All rights reserved.
Keywords: Vanilloid; Nociceptors; Neuroablation; Resiniferatoxin; Capsaicin; TRPV1
INTRODUCTION Resiniferatoxin (RTX), a potent tissue irritant and an agonist at vanilloid receptor type 1 (VR1), is isolated from the latex of several members of the genus Euphorbia (Szallasi and Blumberg, 1989). This agent is 1000 to 10,000 times more potent than capsaicin, a wellknown VR1 agonist found in chili peppers. RTX was found to evoke large inward currents in dorsal root * Corresponding author. Tel.: þ352-392-5661; fax: þ352-392-7609. E-mail address: [email protected] (R.M. Caudle).
ganglion (DRG) neurons and VR1 transfected cell lines (Szallasi et al., 1999). The actions of RTX are mediated by binding, with low nanomolar affinity, directly to the capsaicin-binding site on the VR1 receptor (Biro et al., 1997; Jerman et al., 2000). VR1 is located primarily on small and medium diameter nociceptive afferents (Cortright et al., 2001; Szallasi et al., 1993; Tominaga et al., 1998) and the receptors are thought to be rapidly desensitized by vanilloid agonists (Biro et al., 1997). The apparent desensitization to RTX is long lasting; whereas capsaicin produces only acute desensitization of the VR1 mediated currents (Liu and Simon, 1998). This observation has been confirmed in Ca2þ imaging
0161-813X/$ – see front matter # 2003 Elsevier Inc. All rights reserved. doi:10.1016/S0161-813X(03)00146-3
896
R.M. Caudle et al. / NeuroToxicology 24 (2003) 895–908
experiments where multiple short treatments with capsaicin-induced smaller and smaller transients in the cytosolic-free Ca2þ (Karai et al., unpublished observation). However, capsaicin responses in these experiments completely recover within 30 min of washing out the agonist, whereas they do not recover with RTX. If sufficiently high doses of vanilloid agonists are used in vivo, selective loss of VR1 expressing neurons is possible (Goso et al., 1993). RTX was reported to induce long-term loss of response to thermal stimuli when applied systemically or epidurally in adult rats (Pan et al., 2003; Szabo et al., 1999) and systemic administration of high doses of capsaicin lesion C-fibers in neonates (Hammond and Ruda, 1989). When capsaicin is repeatedly applied to the skin of human volunteers desensitization to thermal stimuli is observed (Fuchs et al., 2000). Nolano et al. (1999) and Pini and Lynn (1991) demonstrated that long-term peripheral capsaicin treatment produced a local Cfiber denervation. Recovery of peripheral thermal responses was due to the re-innervation of the tissue. Overall, these data indicate that vanilloid agonists can be neurotoxic if the neurons are subjected to prolonged exposure to high concentrations of the agonist and suggest that the lesions are selective for a subset of neurons. Our recent work indicates that RTX is highly toxic to DRG neurons expressing VR1. The toxicity leads rapidly to damage to membranes in intracellular organelles (Olah et al., 2001a,b) and preliminary electrophysiological experiments indicated that this membrane damage also included the plasma membrane. Capsaicin and other activators of VR1 were not as damaging to the cells suggesting that RTX may have unique toxicological properties beyond its high potency. The vanilloid receptor is a member of the transient receptor potential (TRP) family, which is a group of receptors that are activated by a variety of stimuli ranging from protons to mechanical perturbations of the membrane (Gunthorpe et al., 2002; Montell et al., 2002). The vanilloid receptor is the only member of this family that responds to compounds with the vanilloid moiety in their structure such as capsaicin, the pungent agent in hot chili peppers, and resiniferatoxin. This receptor is believed to be responsible for detecting potentially damaging stimuli in the periphery in the form of heat and protons. Because the receptor’s function can be significantly enhanced by a number of inflammatory mediators it is also likely to play a role in sensitizing nociceptive afferent fibers that innervate an injured area (Caterina et al., 2000). In addition, recent work has suggested that when the pH of tissue is
low, as in inflammation, anandamide can act as an endogenous agonist at the capsaicin binding site (Olah et al., 2001a,b). These data indicate that VR1 is an important integrator of a variety of nociceptive signals in the peripheral nervous system. Recent work has demonstrated that DRG neurons that express VR1 have several distinct phenotypes (Petruska et al., 2000a,b). These cells are both immunohistochemically and electrophysiologically distinguishable and likely subserve different sensory functions. Since RTX administration in vivo results in a loss of thermal sensitivity with no detectable loss of other senses (Pan et al., 2003; Szabo et al., 1999) it is possible that some VR1 expressing neurons are not as susceptible to damage from RTX as others. If a particular DRG cell type had relatively high levels of calcium binding proteins and low expression of VR1 it might not succumb as easily to RTX. Therefore, the questions addressed in this study were whether RTXs toxicity was limited to a subset of VR1 expressing cells and whether RTX was uniquely neurotoxic when compared to other activators of VR1.
MATERIALS AND METHODS Dorsal Root Ganglion Whole Cell Patch Clamp Adult male Sprague-Dawley rats (90–180 g) were anesthetized with halothane. Following decapitation, the spinal cord was rapidly removed and the dorsal root ganglia were dissected free. All animals were housed in AAALAC approved quarters, and all procedures were reviewed and approved by the University of Florida Institutional Animal Care and Use Committee. Dissected ganglia were placed in a tube containing dispase (neutral protease, 5 mg/ml; Boehringer Mannheim) and collagenase (2 mg/ml; Sigma type 1). The tube was shaken for 60 min in a heated (358) bath. Following wash and trituration, recovered cells were plated on 10 polylysine coated Petri dishes. All recordings were made at room temperature and completed within 10 h of plating. Glass pipettes (Scientific Products B4416-1) were prepared (2–4 MO) with a Brown and Flaming type horizontal puller (Sutter model P87). Whole cell recordings were made with an Axopatch 200B (Axon Instruments). Stimuli were controlled and digital records captured with pClamp 8.0 software and Digipack 1320A ac/dc converter (Axon Instruments). Series resistance was compensated 40–60% with Axopatch 200B compensation circuitry. The ‘‘membrane test’’
R.M. Caudle et al. / NeuroToxicology 24 (2003) 895–908
897
Fig. 1. Electrophysiological characterization of dissociated dorsal root ganglion cells. DRG cells were isolated as described in the methods and voltage clamped. The cells were then classified using the described classification protocols. The traces on the left are representative traces demonstrating the response of type 1, 2, 5 and 4 neurons to classification protocol 1 (CP1) and CP2 for the type 5 neuron. CP2 is not shown for type 1, 2 and 4 cells since it is not needed for classification of these cell types (see Petruska et al., 2000a,b). Cell types 1, 2 and 5 are VR1 expressing cell types, whereas type 4 cells do not express VR1. The traces on the right are action potentials obtained in current clamp mode and evoked by a brief (1 ms, 2–3 nA) depolarizing current injection.
898
R.M. Caudle et al. / NeuroToxicology 24 (2003) 895–908
protocol was used to track cell capacitance, whole cell resistance and access resistance. Following seal and break-in, cells were permitted to equilibrate for approximately 3 min. Subsequently the cell was classified as type 1, 2, 5 or 4 as described below. DRG types 1, 2 and 5 are capsaicin sensitive nociceptive neurons, whereas type 4 cells are capsaicin insensitive neurons (Fig. 1) (Petruska et al., 2000a). Following classification, cells were brought into current clamp mode. The resting membrane potential was assessed and an action potential was evoked by current injection (2–3 nA, 1 ms). After returning to voltage clamp, cell size was determined using the membrane test. RTX (0.01–100 nM), capsaicin (5 mM) (Sigma, St. Louis, MO), or vehicle containing the identical concentration of ethanol was presented by a 1 mm diameter pipette placed near the cell. Cell size was assessed immediately afterward and again after a 2min wash with Tyrode’s solution. Recordings were made exclusively from cells with diameters between 25 and 40 mm. Cell diameter was estimated from the average of the longest and shortest axis as measured through an eyepiece micrometer scale. Cells were classified according to patterns of voltageactivated currents (current signatures) that were revealed by three classification protocols (see Petruska et al., 2000a,b for detailed descriptions). In the studies described below classification protocols 1 (CP1) and 2 (CP2) were sufficient to classify the four afferent subclasses examined. With CP1, currents were evoked by a series of hyperpolarizing pulses presented from a holding potential (VH) of 60 mV (10 mV per step to a final potential of 110 mV; 500 ms, 4 s interstimulus interval). With CP2, strongly depolarizing command steps were used to produce outward current patterns. From a VH of 60 mV, a 500 ms conditioning pulse to 100 mV was followed by 200 ms depolarizing command steps (20 mV) to a final potential of þ40 mV. If CP1 evoked a transient outward current following repolarization to 60 mV, the cell was classified as type 2. If only leak currents were present, CP2 was applied. If no A-current peaks were observed with CP2, the cell was classified as type 1. Cells were classified as type 5 cells if small amplitude H-current (<100 pA) was revealed by CP1; and CP2 revealed three A-current deflections (Fig. 1). Type 4 cells were identified by large H-currents (>400 pA) during the CP1 protocol (Fig. 1). Plated cells were superfused in rat Tyrode’s solution containing (in mM): NaCl (140), KCl (4), MgCl2 (2), CaCl2 (2), glucose (10), HEPES (10), adjusted to pH 7.4 with NaOH. Test solutions were applied via gravity fed pipette positioned approximately 1 mm from the
cell. The recording electrodes were filled with (in mM): KCl (120), Na2-ATP (5), Na2-GTP (0.4), EGTA (5), CaCl2 (2.25), MgCl2 (5), HEPES (20), and adjusted to pH 7.4 with KOH; osmolarity was approximately 315–325 mOsm. Resiniferatoxin and capsaicin were dissolved in ethanol/water (1:1) and then diluted to their final concentration in Tyrode’s solution. HEK-293 Whole Cell Patch Clamp HEK-293 cells were seeded on to 13 mm cover slips at low density and then transfected with VR1 fused to enhanced green fluorescence protein (VR1eGFP) as described (Olah et al., 2001a,b). Twenty-four hours after transfection the cells were placed in a bath and superfused with Tyrode’s solution. The cells were viewed on an inverted microscope equipped with phase contrast and fluorescence optics. Cells expressing VR1eGFP were identified by their green fluorescence. The fluorescently labeled cells were whole cell patch clamped as described for DRG neurons. The holding potential was 60 mV. Cell capacitance was measured prior to treatment and 10–15 min following exposure to RTX, capsaicin or heating of the superfusion solution. Resiniferatoxin (100–125 pM) or capsaicin (5 mM) was applied via a gravity fed pipette placed within 200 mm of the patched cell. For experiments in which heat was used to activate vanilloid receptors, the bath temperature was elevated by using an inline heater (Warner Instrument Corp., Hamden, CT). The bath temperature was monitored with a miniature thermistor placed in the recording chamber. Fluorescent Confocal Microscopy COS7, NIH3T3 and HEK-293 cells were seeded on 25 mm coverslips and transfected with the VR1eGFP plasmid construct, as described previously (Olah et al., 2001a,b). Transiently transfected cells were cultured for an additional 24 h. To observe morphological changes of neuronal processes, dissociated DRG cultures were prepared from embryonic (E16) rats as described before in detail (Olah et al., 2001a,b). Briefly, after dissection and trituration DRG cells were seeded on 25 mm glass cover slips. Surfaces were coated with poly-D-lysine and laminin. Cultures were selected in a medium containing 2 and 5 mg/ml FUDR and uridine, respectively, for 4 days to 1 week, at which point well-differentiated neurons with elongated processes and non-dividing cells dominated the population. Primary DRG cultures at this stage were used for fluorescent Ca2þ imaging.
R.M. Caudle et al. / NeuroToxicology 24 (2003) 895–908
Calcium Imaging DRG cells were washed three times with imaging medium (IM) containing Hanks’ balanced salt solution, supplemented with 0.01% bovine serum albumin (BSA), 1.25 mM CaCl, 0.8 mM MgCl, 1 mM ascorbic acid, 1 mM pyruvate and buffered to pH 7.4 with HEPES. Cultures were loaded with 5 mM Fluo-4 AM dye. After incubation for 30 min at room temperature cells were washed three times in IM to remove excess dye and kept in the dark for at least 15 min before starting the experiments. Recordings were carried out in an open imaging chamber (Promega). Cells were visualized by an inverted Nikon microscope equipped with the MRC-1024 Bio-Rad Confocal System. Statistics Paired and un-paired t-tests, Pearson correlation and repeated measures ANOVA followed by Dunnett’s post-hoc test were used as indicated in the figure legends and text. The alpha level was set at 0.05.
899
RESULTS Effect of RTX in DRG Neurons To examine the influence of RTX on peptidergic and non-peptidergic, capsaicin-sensitive DRG neurons, patch clamp experiments were performed on type 1, 2 and 5 cells (Petruska et al., 2000a). Cell types 1 and 5 are known to express calcitonin gene related peptide (CGRP) and somatostatin, and CGRP and substance P, respectively. Type 2 cells do not express these peptides (Petruska et al., 2000a,b). Each of these cell types was previously demonstrated to respond to capsaicin (Petruska et al., 2000a,b). Once neurons were voltage clamped and typed (Fig. 1) whole cell capacitance was measured before and after the application of various concentrations of RTX. Analysis of the concentration response data for type 1 (N ¼ 5), type 2 (N ¼ 13) and type 5 (N ¼ 4) cells demonstrated that the cells did not differ in their responses to RTX. Thus, the data for these three cell types was combined. Fig. 2a demonstrates that RTX-induced large currents in these
Fig. 2. Effect of RTX on type 1, 2 and 5 DRG neurons. (a) Representative current traces from a type 2 DRG neuron while superfusing with 10 nM RTX. The horizontal line represents the 1-min RTX application. (b) RTX evoked current concentration response relationship. (c) RTX-induced change in capacitance concentration response relationship. (d) Effect of vehicle (1.7 mM ethanol) on RTX-induced changes in membrane capacitance. Asterisk indicates P < 0:05, paired t-test.
900
R.M. Caudle et al. / NeuroToxicology 24 (2003) 895–908
neurons. Washing out the RTX did not typically reverse the current, even when the cells were held for 15 min after the RTX administration. Even currents evoked by 10 pM RTX did not completely reverse. The lack of a full reversal of the current suggests that the neurons were damaged by the RTX treatment as previously described (Olah et al., 2001a,b). The current evoked by RTX was concentration dependent with an EC50 of 0:1 1:2 nM (Fig. 2b). In addition to the inward current, RTX reduced the capacitance of these DRG neurons in a concentration dependent manner with an EC50 of 0:2 3:5 nM (Fig. 2c). The EC50s for RTX-induced current and changes in capacitance were not statistically different. The change in capacitance indicates that the neurons had a decrease in surface area following RTX treatment. The vehicle used for delivering RTX (ethanol, maximum concentration ¼ 1:7 mM) did not induce currents nor did it change the capacitance of the neurons (Fig. 2d). To determine the time course of the change in capacitance the current and capacitance were monitored simultaneously in three cells. As demonstrated in Fig. 3a, the decrease in cell capacitance paralleled the
onset of the inward current. Plotting the current versus the capacitance for a representative type 2 neuron illustrates that the relationship between these measures is linear until the capacitance reaches a plateau (Fig. 3b). The current continues to increase following the plateau of the capacitance. The role of calcium in mediating the loss of membrane by the neurons during RTX administration was examined by removing the extracellular calcium from the superfusion buffer. In normal buffer the capacitance of DRG neurons was reduced by 100 nM RTX from 26 2:7 to 9 2:0 pF (N ¼ 3 neurons, P ¼ 0:026, paired t-test). In nominally Ca2þ-free buffer 100 nM RTX did not significantly influence cell capacitance. The capacitance before treatment was 27 1:4 and 37 8:5 pF after treatment (N ¼ 4 neurons, P ¼ 0:26, paired t-test). Interestingly, the maximum current RTX induced in nominally Ca2þ-free buffer was larger than in normal buffer (2:7 0:35 nA versus 1:4 0:33 nA, P ¼ 0:024, Student’s t-test). The response of type 4 cells, which lack vanilloid receptors (Petruska et al., 2000a), to RTX was also examined. Unlike type 1, 2 and 5 cells, RTX (100 nM) did not induce currents in type 4 cells (N ¼ 4) nor
Fig. 3. Relationship between RTX-induced current onset and capacitance changes in DRG neurons. (a) Current (top) and capacitance (bottom) record for a representative type 2 neuron demonstrating the rapid loss of membrane induced by RTX. (b) Plot of current versus capacitance for the above cell. The solid line is the regression of the linear portion of the data.
R.M. Caudle et al. / NeuroToxicology 24 (2003) 895–908
901
Fig. 4. Ca2þ fluorescence images of RTX-induced membrane blebbing in cultured DRG neurons. (a) Representative DRG neurons at the start of RTX (1 nM) superfusion. (b) Same neurons 2 min after RTX addition. The arrows indicate regions of membrane blebbing. Blebbing was observed in both the cell body and the processes.
did it have any effect on the cells’ capacitance (data not shown). To visualize the changes in the membrane induced by RTX, cultured DRG neurons were examined using Ca2þ imaging techniques. Cell types cannot be evaluated in this preparation; however, RTX-induced blebbing of DRG cell bodies and neuronal processes (Fig. 4) in approximately fifty percent of the cells. The membrane forming the blebs appears to be lost in vanilloid receptor containing neurons during the application of RTX. The blebbing in the cell bodies is consistent with the reduction in capacitance observed in patch clamp experiments.
Fig. 5. Effect of capsaicin (5 mM) on DRG neurons. (a) Capsaicininduced currents from a representative type 2 neuron. The horizontal lines indicate the duration of the capsaicin application. (b) The change in capacitance in type 2 DRG neurons after capsaicin exposure. Asterisks indicate P < 0:05, repeated measures ANOVA followed by Dunnett’s test, N ¼ 9 DRG neurons.
Effect of Capsaicin in DRG Neurons
Effect of RTX on VR1eGFP Transfected HEK-293 Cells
The application of capsaicin (5 mM) to type 1, 2 and 5 DRG neurons produced currents that were similar in size to 100 nM RTX (1:4 0:4 nA, N ¼ 5 cells), but the effects on the cells’ membranes were less dramatic. Repeated capsaicin applications produced currents with reduced amplitudes when compared to the initial capsaicin response (Fig. 5a), but complete loss of the current was not observed. In addition, the capacitance was reduced from 41:9 1:8 pF to a minimum of 37:5 1:7 pF (P < 0:05, repeated measures ANOVA followed by Dunnett’s test, N ¼ 9 DRG neurons, Fig. 5b). Interestingly, unlike RTX removing Ca2þ did not significantly influence the amplitude of the capsaicin-evoked current (1:7 0:4 nA, N ¼ 3 DRG cells, P ¼ 0:33, Student’s t-test 0Ca2þ versus normal buffer). Similar to the RTX experiments type 4 neurons demonstrated no response to capsaicin (data not shown).
To further evaluate the role of vanilloid receptors in decreasing cell size during RTX treatment HEK-293 cells were transfected with the vanilloid receptor C-terminally fused to enhanced green fluorescent protein (VR1eGFP). To quantify the changes in cell membrane properties VR1eGFP transfected HEK293 cells were whole cell voltage clamped as previously described. Exposure of these cells to 125 pM RTX resulted in a large inward current (Fig. 6a, N ¼ 9 cells). This current displayed the typical VR1 kinetics and had a reversal potential of approximately 0 mV as previously described (Caterina et al., 1997). Subsequent applications of RTX evoked little or no current indicating rapid and complete loss of RTX responsiveness. The RTX evoked current did not return even when the RTX was washed out of the bath for 30 min. Unlike the DRG cells, the RTX-induced current returned to baseline levels upon washout of the drug. HEK-293
902
R.M. Caudle et al. / NeuroToxicology 24 (2003) 895–908
Fig. 6. Effect of resiniferatoxin on HEK-293 cells transfected with VR1eGFP. (a) Representative current traces demonstrating that bath applied RTX (125 pM, 1 min) induces a large inward current on the first application, but subsequent applications are ineffective. The horizontal lines represent the application of RTX. The traces were collected 10 min apart. (b) A representative current trace from an eGFP transfected HEK-293 cell during RTX (125 pM, horizontal line) treatment. (c) Summary of RTX-induced change in capacitance. Asterisk indicates P ¼ 0:03, paired t-test, N ¼ 9 HEK-293 cells.
cells that were not transfected or were transfected with eGFP did not respond to RTX (Fig. 6b, N ¼ 5 cells). In addition to the rapid loss of the RTX mediated current, the cells’ membrane capacitance was reduced 10 min following the RTX exposure (Fig. 6c). The capacitance before RTX treatment was 20:9 3:7 and 14:6 2:6 pF after treatment (N ¼ 9 cells, P ¼ 0:03, paired t-test). The time constant for the decay of the capacitive charge was not significantly altered by the RTX treatment (t ¼ 0:26 0:05 ms before resiniferatoxin and t ¼ 0:86 0:4 ms following treatment, P ¼ 0:15, paired t-test, N ¼ 5 cells). Cells not transfected with VR1eGFP did not demonstrate this decrease in cell capacitance. To assess whether the amount of membrane lost is related to the number of VR1eGFP receptors in the membrane, the rate of onset of the RTX-induced current was plotted versus the maximum change in capacitance. Analysis of this relationship demonstrated that the two measures were correlated (r 2 ¼ 0:68, P < 0:01, Pearson correlation).
This analysis assumes that the rate of onset of the current, at a fixed concentration of agonist, is proportional to the number of VR1eGFP receptors in the membrane. Thus, this finding suggests that the number of VR1 receptors in the membrane determine the amount of membrane that will be lost following RTX treatment. Effect of Capsaicin on VR1eGFP Transfected HEK-293 Cells Application of capsaicin (5 mM) to the transfected HEK-293 cells resulted in large inward currents (Fig. 7a). In contrast to RTX, repeated application of capsaicin-induced currents, albeit they were smaller in amplitude than during the first exposure. These data were similar to those obtained from DRG neurons treated with capsaicin. Capsaicin treatment also reduced the capacitance of the VR1eGFP containing HEK-293 cells from 28:9 5:3 to 26:2 4:9 pF (mean S:E:,
R.M. Caudle et al. / NeuroToxicology 24 (2003) 895–908
903
Fig. 7. Effect of capsaicin and heat activation on VRleGFP transfected HEK-293 cells. (a) Currents from a representative HEK-293 cell exposed multiple times to capsaicin (5 mM). The horizontal lines indicate 1 min bath applications of capsaicin. (b) Effect of capsaicin on cell capacitance. Asterisks indicate P < 0:05, repeated measures ANOVA followed by Dunnett’s test on raw data, N ¼ 4 cells. (c) Top traces are heat-activated currents and the bottom traces are the corresponding records of the bath temperature. The traces were collected 5 min apart. (d) Summary of the heat-induced change in capacitance. Asterisk indicates P ¼ 0:02, paired t-test, N ¼ 7 cells.
N ¼ 4 cells, P < 0:05 repeated measures ANOVA followed by Dunnett’s test) (Fig. 7b). Effect of Heat on VR1eGFP Transfected HEK-293 Cells The vanilloid receptor was previously demonstrated to be activated by temperatures above 42 8C (Caterina et al., 1997; Davis et al., 2000; Kirschstein et al., 1999).
These studies did not report any significant cellular changes following heat. Thus, it was possible that the loss of membrane produced by RTX and capsaicin in DRG and VR1eGFP transfected cells was due specifically to the activation of the vanilloid receptor by vanilloid agonists and not simply due to the Ca2þ entry through the receptor when it was activated. To test this idea VR1eGFP transfected HEK-293 cells were patch clamped and the capacitance of the cells was measured
904
R.M. Caudle et al. / NeuroToxicology 24 (2003) 895–908
before and after heating the bathing solution to 47 8C with an inline heater. Fig. 7c illustrates that heating the bath-induced currents in the VR1eGFP containing cells. Although the currents decreased in amplitude following the first heat application a complete loss of response was not observed as was seen with the RTX treatment. This may have been due to the smaller currents induced by the heat (500 pA versus >1000 pA for RTX). However, increasing the maximum temperature to 508 resulted in the rapid loss of the cells, as did increasing the duration of the heat stimulus. Thus, larger currents could not be obtained. Even though complete loss of response to heat was not achieved, the cells’ membrane capacitance was significantly reduced from 43:8 6:6 to 36:6 6:6 pF (N ¼ 7 cells, P ¼ 0:02, paired t-test) (Fig. 7d). Role of Ca2þ in VR1-Induced Membrane Loss To confirm that the change in membrane capacitance produced by RTX, capsaicin and heat was due to Ca2þ entry into the cells, VR1eGFP transfected and nontransfected HEK-293 cells were exposed to the Ca2þ ionophore 4-bromo, 5-(methylamino)-2-[3,9,11-trimethyl-8-[1-methyl-2-oxo[6S-6-alpha-(2S*,3S*)], 8beta-(R*)], 9-beta, 11-alpha]-4-benzoxazole carboxylic acid (4-bromo A23187, Calbiochem, San Diego, CA) (1 mM) via the superfusion buffer while the cells were voltage clamped at 60 mV. In the transfected cells membrane capacitance was reduced from 23:6 5:0 to 18:8 3:8 pF (N ¼ 10 cells, P ¼ 0:03, paired t-test) 10 min after the removal of 4-bromo A23187. Non-transfected cells, on the other hand, did not demonstrate a change in capacitance. The capacitance was 73:0 40 pF before treatment and 88:7 40 pF after treatment (N ¼ 5 cells, P @ 0:05, paired t-test). These findings indicate that VR1 and Ca2þ are necessary for the observed membrane remodeling. This finding is consistent with our previously published data (Olah et al., 2001a,b). Note that although the baseline capacitance for VR1eGFP transfected cells was smaller than for non-transfected cells this difference was not statistically significant (P > 0:05, Students’ t-test). Confocal Microscopy of VR1eGFP Transfected Cell Lines Confocal fluorescence microscopy was employed to analyze the cellular distribution of VR1eGFP. Optical sections taken at the plane of cell attachment to the glass surface show VR1eGFP fluorescence in plasma
Fig. 8. Confocal images of representative COS7 cells transfected with VR1eGFP. Similar results were obtained with HEK-293 and NIH3T3 cells. (a) Control conditions prior to RTX application: N indicates the nucleus. (b) The same cells 1 min after 1 nM RTX: arrows indicate blebs. (c) The same cells 10 min after 1 nM RTX: arrows indicate blebs. Note some of the blebs have separated from the cells.
membrane processes (Fig. 8a). VR1 accumulation is also present in intracellular structures consistent with the ER, as previously described (Olah et al., 2001a,b). Fig. 8 illustrates the localization of VR1 in COS7 cells, but similar localization was seen in HEK-293 and NIH3T3 cells, indicating that it was not a cell type specific anomaly (Olah et al., 2001a,b). COS7 cells are
R.M. Caudle et al. / NeuroToxicology 24 (2003) 895–908
presented in Fig. 8 because their flat profile allowed visualization of membrane remodeling in a single plane. The more rounded HEK-293 cells require three-dimensional confocal imaging to follow the dynamics of the membrane. Exposure to 1 nM RTXinduced fragmentation of the VR1eGFP decorated ER and vesiculation of the plasma membrane processes in seconds (Fig. 8b versus Fig. 8a). Detached vesicles were seen by 10 min. In cells expressing only eGFP or the KDEL (lys–asp–glu–leu) signal linked eGFP plasmid the structure of the mitochondria and ER did not change upon exposure to vanilloids, such as RTX, olvanil or capsaicin, indicating the dependence of these effects on the presence of fully functional VR1 receptor. Thus, within the first few seconds after vanilloid exposure, coincident, structurally similar intracellular organelle remodeling occurs for both the ER and the plasma membrane processes.
DISCUSSION Resiniferatoxin is a potent analog of capsaicin, the active ingredient in hot chili peppers, that is isolated from the latex of several species of Euphorbia (Szallasi and Blumberg, 1989). RTX was previously demonstrated to produce loss of thermal nociceptive responses in rodents (Pan et al., 2003; Szabo et al., 1999). This effect appears to be different from the one that follows repeated application of low concentrations of capsaicin (Liu and Simon, 1998). The difference between RTX and capsaicin is due to a targeted destruction of DRG neurons that contain vanilloid receptors by RTX, whereas capsaicin appears to desensitize the neurons (Olah et al., 2001a,b). In the current study we demonstrate that RTX produces a rapid dismantling of the cell membrane of both DRG neurons and transfected cells that express the vanilloid receptor. If we assume that membrane capacitance is 1 mF/cm2 of membrane (Hille, 1991) then the decrease in DRG neuron capacitance indicates that approximately 2800 mm2 of membrane was removed from the surface of the neurons when they were treated with a maximally effective concentration of RTX. This represents a >50% decrease in cell surface indicating significant damage to the cells. Sensory neurons in the DRG are a diverse population of cells that have a variety of physical and biochemical properties. The different DRG cell types are believed to have distinct functions (Petruska et al., 2000a,b). In the present study we found no difference in sensitivity between DRG neurons classified as types 1, 2 and 5.
905
These three classes of nociceptive neurons express VR1. Type 4 nociceptive DRG neurons, which do not express VR1, were insensitive to vanilloid treatment. Thus, the presence of VR1 is required for RTX to damage the cells, but DRG cell type does not influence the response to RTX. In addition to the RTX-induced membrane loss we found that capsaicin and heat activation of vanilloid receptors also demonstrated decreases in capacitance. However, the degree of membrane loss was not as large as with RTX and the response to the stimulus was never fully lost as was observed with RTX. The total loss of membrane in DRG cells following multiple exposures to 5 mM capsaicin, a maximally effective concentration (Gunthorpe et al., 2000), was approximately 440 mm2. These data suggest that the presence of the vanilloid receptor and free intracellular Ca2þ are the main determinants for inducing cell membrane loss. The finding that 4-bromo A23187, a Ca2þ ionophore, induced a decrease in membrane capacitance only in cells containing VR1, in the absence of vanilloid agonists, supports this conclusion. The difference in the degree of membrane loss between RTX and the other stimuli is somewhat puzzling. The maximum current produced by RTX was no greater than the maximum current produced by capsaicin, yet there was a six-fold difference in the maximum amount of membrane lost. It is possible that RTX evoked opening of the VR1 ion channel leads to greater Ca2þ permeability when compared to the other stimuli or to the greater release of internal stores of Ca2þ. To the best of our knowledge, differences in ion permeability in vanilloid receptors have not been demonstrated for the various activators of the receptor. However, we have observed longer lasting elevations in intracellular-free Ca2þ following RTX when compared with capsaicin (Olah et al., 2001a,b). The prolonged elevation in Ca2þ may potentiate RTXs toxicity. An interesting observation in this study was that the maximum amplitude of the RTX current in DRG neurons was significantly reduced by the presence of extracellular Ca2þ, whereas the current evoked by capsaicin was not influenced by the ion. One possible explanation for this phenomenon is that because of the rapid loss of membrane with RTX in buffer containing Ca2þ vanilloid receptors were lost prior to their activation, thus reducing the maximum current that could be evoked. In the capsaicin experiments the loss of membrane was significantly less, which decreased the difference between Ca2þ-free and normal buffer. Confocal micrographs demonstrated that the membrane losses were the result of vesiculation and the
906
R.M. Caudle et al. / NeuroToxicology 24 (2003) 895–908
subsequent internalization/expulsion of the vesicles from the cell (Fig. 8). The decrease in capacitance occurs on a time scale (t50 20 s) that is intermediate between our previously reported time scales for endoplasmic reticulum vesiculation (t50 1 s) and nuclear blebbing (t50 5 min) (Olah et al., 2001a,b). Thus, the loss of plasma membrane appears to be the second phase of RTX-induced cell degradation. Possible Mechanisms for RTX-Induced Membrane Loss
were unable to protect DRG neurons from membrane loss with the free radical scavenger N-acetyl-L-cysteine (data not shown). These data suggest that oxidative/ nitrative stress is not involved in the acute membrane loss observed with RTX. Thus, the data from our two studies suggests that neither calpains, caspases nor oxidative/nitrative stress are involved in the mechanism for the VR1-induced Ca2þ mediated membrane loss. Thus, the mechanism for RTX-induced cell damage remains to be elucidated. Therapeutic Potential of RTX
The RTX-induced cell damage requires an increase in intracellular Ca2þ (present study; Olah et al., 2001a,b), but the pathways mediating the vesiculation of the plasma membrane are not known. Because the membrane is expelled from the cells, as well as internalized, it is not likely that the electrophysiologically observed loss of plasma membrane is due to typical receptor internalization as is observed with G-protein coupled receptors (Ferguson, 2001). In studies by other groups, similarities were found between excitotoxicity and both apoptosis and necrosis suggesting that there is a continuum between these two mechanisms of cell death (Chan and Mattson, 1999; Portera-Cailliau et al., 1997a,b). Similarly, we found a disruption of ER, nuclear and mitochondrial membranes (Olah et al., 2001a,b). However, chromatin fragmentation, apoptotic body formation, and caspase activity were absent indicating that the process was not apoptotic. Our observations on RTX treated DRG neurons are similar to those reported previously for excitotoxicity mediated by excitatory amino acids such as glutamate or aspartate. Faddis et al. (1997) suggested that channel openings by these amino acids induce varicosities by accumulation of excess intracellular Ca2þ similar to that demonstrated here. However, their data suggests that vesicularization is mediated by the local activation of calpain, a Ca2þ-dependent protease, in the vicinity of activated receptors. We tested the possibility that calpain was involved in the membrane loss by including calpain inhibitor peptide in the recording pipettes during some experiments. However, the peptide did not attenuate the loss of membrane in the VR1 containing HEK-293 cells (data not shown). This result suggests that calpain was not involved in the membrane loss, with the caveat that without a positive control it is also possible that the peptide did not inhibit calpain as expected. Jambrina et al. (2003) found that oxidative/nitrative stress-induced damage to Jurcat cells transfected with VR1. They described this as a paraptotic process. Similar to the calpain experiments we
Chronic neuropathic pain of peripheral origin is thought to be maintained by neuromas, or peripheral generators, that send a constant afferent barrage of nociceptive signals to the central nervous system. This excessive input to the spinal cord results in an enhancement of excitability in the spinal cord, an expansion of receptive fields, hyperalgesia, and allodynia that is often referred to as altered central processing (Eliav and Gracely, 1998; Gracely et al., 1993; Morris et al., 1997). The allodynia, derived from the altered central processing, is thought to be mediated to a large extent by the interpretation of Ab mechanoreceptor activity as pain by the central nervous system, rather than touch. The peripheral generator is likely to be injured primary afferent nociceptors. In support of this hypothesis, experimental altered central processing with the associated allodynia and hyperalgesia can be produced by an intradermal injection of capsaicin (Iadarola et al., 1998; Magerl et al., 1998; Morris et al., 1997). The activation of capsaicin sensitive primary afferents results in altered central processing of sensory information leading to the interpretation of Ab fiber activity as pain (Andersen et al., 1995). These data indicate that the peripheral generators in neuropathic pain might be targeted using the vanilloid receptors. The study by Szabo et al. (1999) demonstrates that application of RTX can eliminate [3 H]-RTX binding sites in the spinal cord and that the loss of these binding sites is associated with long lasting thermal analgesia. Thus, RTX may be useful in treating some forms of chronic pain. In summary, the present study demonstrates that RTX produces a rapid loss of cell membrane that is mediated by the heat/vanilloid-induced opening of the VR1 cation channel. The membrane loss is an early response in a Ca2þ-mediated toxicity that leads to cell death. The increased cytosolic Ca2þ vesiculates the ER, nuclear and plasma membranes with the expulsion of the membrane extra and intracellularly. The mechanism mediating the blebbing and vacuolization is not entirely known, but
R.M. Caudle et al. / NeuroToxicology 24 (2003) 895–908
does not appear to involve calpain, a Ca2þ activated protease, caspases, the principal executive enzymes of apoptosis (Olah et al., 2001a,b), or oxidative/nitrative stress. These findings do indicate that RTX rapidly and selectively disables VR1 containing cells. Because vanilloid receptors reside mainly on nociceptive primary afferents RTX may be a useful neuroablative agent for the treatment of chronic pain. ACKNOWLEDGEMENTS The National Institute of Dental and Craniofacial Research, National Institutes of Health supported this work. REFERENCES Andersen OK, Gracely RH, Arendt-Nielsen L. Facilitation of the human nociceptive reflex by stimulation of A beta-fibers in a secondary hyperalgesic area sustained by nociceptive input from the primary hyperalgesic area. Acta Physiol Scand 1995;155(1):87–97. Biro T, Acs G, Acs P, Modarres S, Blumberg PM. Recent advances in understanding of vanilloid receptors: a therapeutic target for treatment of pain and inflammation in skin. J Invest Dermatol Symp Proc 1997;2(1):56–60. Caterina MJ, Schumacher MA, Tommaga M, Rosen TA, Levine JD, Julius D. The capsaicin receptor: a heat-activated ion channel in the pain pathway. Nature 1997;389(6653):816–24. Caterina MJ, Leffler A, Malmberg AB, Martin WJ, Trafton J, Petersen-Zeitz KR, Koltzenburg M, Basbaum AI, Julius D. Impaired nociception and pain sensation in mice lacking the capsaicin receptor. Science 2000;288(5464):306–13. Chan SL, Mattson W. Caspase and calpain substrates: roles in synaptic plasticity and cell death. J Neurosci Res 1999;58(1): 167–90. Cortright DN, Crandall M, Sanchez JF, Zou T, Krause JE, White G. The tissue distribution and functional characterization of human vrl. Biochem Biophys Res Commun 2001;281(5):1183–9. Davis JB, Gray J, Gunthorpe MJ, Hatcher JP, Davey PT, Overend P, Harries MH, Latcham J, Clapham C, Atkinson K, Hughes SA, Rance K, Grau E, Harper AJ, Pugh PL, Rogers DC, Bingham S, Randall A, Sheardown SA. Vanilloid receptor-1 is essential for inflammatory thermal hyperalgesia. Nature 2000;405(6783):183–7. Eliav E, Gracely RH. Sensory changes in territory of the lingual and inferior alveolar nerves following lower third molar extraction. Pain 1998;77:191–9. Faddis BT, Hasbani MJ, Goldberg MP. Calpain activation contributes to dendritic remodeling after brief excitotoxic injury in vitro. J Neurosci 1997;17(3):951–9. Ferguson SS. Evolving concepts in G protein-coupled receptor endocytosis: the role in receptor desensitization and signaling. Pharmacol Rev 2001;53(1):1–24. Fuchs PN, Campbell JN, Meyer RA. Secondary hyperalgesia persists in capsaicin desensitized skin. Pain 2000;84(2–3):141–9.
907
Goso C, Piovacari G, Szallasi A. Resiniferatoxin-induced loss of vanilloid receptors is reversible in the urinary bladder but not in the spinal cord of the rat. Neurosci Lett 1993;162(1–2):197–200. Gracely RH, Lynch SA, Bennett GJ. Painful neuropathy: altered central processing maintained dynamically by peripheral input. Pain 1993;54(3):365–6. Gunthorpe MJ, Harries MH, Prinjha RK, Davis JB, Randall A. Voltage- and time-dependent properties of the recombinant rat vanilloid receptor (rVR1). J Physiol 2000;525(3):747–59. Gunthorpe MJ, Benham CD, Randall A, Davis JB. The diversity in the vanilloid (TRPV) receptor family of ion channels. Trends Pharmacol Sci 2002;23(4):183–91. Hammond DL, Ruda MA. Developmental alterations in thermal nociceptive threshold and the distribution of immunoreactive calcitonin gene-related peptide and substance P after neonatal administration of capsaicin in the rat. Neurosci Lett 1989; 97(1–2):57–62. Hille B. Ionic channels of excitable membranes. Sunderland, MA: Sinauer Associates Inc.; 1991. p. 9–10. Iadarola MJ, Berman KF, Zeffiro TA, Byas-Smith MG, Gracely RH, Max MB, Bennett GJ. Neural activation during acute capsaicin-evoked pain and allodynia assessed with PET. Brain 1998;121(5):931–47. Jambrina E, Alonso R, Alcalde M, Del Carmen Rodriguez M, Serrano A, Martinez AC, Garcia-Sancho J, Izquierdo M. Calcium influx through receptor-operated channel induces mitochondria-triggered paraptotic cell death. J Biol Chem 2003;278(16):14134–45. Kirschstein T, Greffrath W, Busselberg D, Treede RD. Inhibition of rapid heat responses in nociceptive primary sensory neurons of rats by vanilloid receptor antagonists. J Neurophysiol 1999;82(6):2853–60. Liu L, Simon SA. The influence of removing extracellular Ca2þ in the desensitization responses to capsaicin, zingerone and olvanil in rat trigeminal ganglion neurons. Brain Res 1998;809(2): 246–52. Magerl W, Wilk SH, Treede RD. Secondary hyperalgesia and perceptual wind-up following intradermal injection of capsaicin in humans. Pain 1998;74(2–3):257–68. Montell C, Birnbaumer L, Flockerzi V, Bindels RJ, Bruford EA, Caterina MJ, Clapham DE, Harteneck C, Heller S, Julius D, Kojima I, Mori Y, Penner R, Prawitt D, Scharenberg AM, Schultz G, Shimizu N, Zhu MX. A unified nomenclature for the superfamily of TRP cation channels. Mol Cell 2002;9(2): 229–31. Morris VH, Cruwys SC, Kidd BL. Characterization of capsaicininduced mechanical hyperalgesia as a marker for altered nociceptive processing in patients with rheumatoid arthritis. Pain 1997;71(2):179–86. Nolano M, Simone DA, Wendelschafer-Crabb G, Johnson T, Hazen E, Kennedy WR. Topical capsaicin in humans: parallel loss of epidermal nerve fibers and pain sensation. Pain 1999;81(1–2):135–45. Olah Z, Szabo T, Karai L, Hough C, Fields RD, Caudle RM, Blumberg PM, Iadarola MJ. Ligand-induced dynamic membrane changes and cell deletion conferred by vanilloid receptor 1. J Biol Chem 2001a;276(14):11021–30. Olah Z, Karai L, Iadarola MJ. Anandamide activates vanilloid receptor 1 (VR1) at acidic pH in dorsal root ganglia neurons and cells ectopically expressing VR1. J Biol Chem 2001b; 276(33):31163–70.
908
R.M. Caudle et al. / NeuroToxicology 24 (2003) 895–908
Pan HL, Khan GM, Alloway KD, Chen SR. Resiniferatoxin induces paradoxical changes in thermal and mechanical sensitivities in rats: mechanism of action. J Neurosci 2003;23(7):2911–9. Petruska JC, Napaporn J, Johnson RD, Gu JG, Cooper BY. Subclassified acutely dissociated cells of rat DRG: histochemistry and patterns of capsaicin-, proton-, and ATPactivated currents. J Neurophysiol 2000a;84(5):2365–79. Petruska JC, Cooper BY, Gu JG, Rau KK, Johnson RD. Distribution of P2X1, P2X2, and P2X3 receptor subunits in rat primary afferents: relation to population markers and specific cell types. J Chem Neuroanat 2000b;20(2):141–62. Pini A, Lynn B. C-fibre function during the 6 weeks following brief application of capsaicin to a cutaneous nerve in the rat. Eur J Neurosci 1991;3(3):274–84. Portera-Cailliau C, Price DL, Martin LJ. Excitotoxic neuronal death in the immature brain is an apoptosis–necrosis morphological continuum. J Comp Neurol 1997a;378(1):70–87. Portera-Cailliau C, Price DL, Martin LJ. Non-NMDA and NMDA receptor-mediated excitotoxic neuronal deaths in adult brain are
morphologically distinct: further evidence for an apoptosis– necrosis continuum. J Comp Neurol 1997b;378(1):88–104. Szabo T, Olah Z, ladarola MJ, Blumberg PM. Epidural resiniferatoxin induced prolong regional analgesia to pain. Brain Res 1999;840:92–8. Szallasi A, Blumberg PM. Resiniferatoxina, phorbol-related diterpene, acts as an ultrapotent analog of capsaicin, the irritant constituent in red pepper. Neuroscience 1989;30(2): 515–20. Szallasi A, Goso C, Blumberg PM, Manzini S. Competitive inhibition by capsazepine of [3 H]resiniferatoxin binding to central (spinal cord and dorsal root ganglia) and peripheral (urinary bladder and airways) vanilloid (capsaicin) receptors in the rat. J Pharmacol Exp Ther 1993;267(2):728–33. Szallasi A, Szabo T, Biro T, Modarres S, Plumberg PM, Krause JE, Cortright DN, Appendino G. Resiniferatoxin-type phorboid vanilloids display capsaicin-like selectivity at native vanilloid receptors on rat DRG neurons and at the cloned vanilloid receptor VR1. Br J Pharmacol 1999;128:428–34.
Resiniferatoxin-Induced Loss of Plasma Membrane in Vanilloid Receptor Expressing Cells Robert M. Caudle1,*, Laszlo Karai2, Narasaiah Mena1, Brian Y. Cooper1, Andrew J. Mannes2, Federico M. Perez1, Michael J. Iadarola2, Zoltan Olah2 1
Department of Oral and Maxillofacial Surgery, University of Florida College of Dentistry and the University of Florida McKnight Brain Institute, 1600 Archer Road, P.O. Box 100416, Gainesville, FL 32610, USA 2 Pain and Neurosensory Mechanisms Branch, National Institute of Dental and Craniofacial Research, National Institutes of Health, Bethesda, MD, USA Received 27 August 2002; accepted 29 May 2003
Abstract Resiniferatoxin (RTX), a potent analog of capsaicin, was evaluated electrophysiologically in dorsal root ganglion (DRG) cells and cell lines ectopically expressing the vanilloid receptor type 1 (VR1) to determine if cell phenotype influenced RTXs neurotoxic properties. Furthermore, capsaicin and heat activation of VR1 were evaluated in these cells to determine if cellular damage was unique to RTX activation of the receptors. RTX application to DRG cells identified as type 1, 2 or 5, cell types known to express VR1, induced large inward currents. RTX did not induce currents in DRG cells that do not express the receptor (type 4 cells). In cell lines ectopically expressing VR1, RTX-induced similar currents. RTX produced no effect in non-transfected cells. After exposure to RTX both DRG cells and transfected cells failed to respond to subsequent applications of the agonist. In addition, whole cell capacitance was reduced up to 70%. The decrease in capacitance was associated with the loss of plasma membrane, as determined by confocal microscopy. Cell phenotype, other than VR1 expression, did not influence the response to RTX. Interestingly, capsaicin and heat activation of vanilloid receptors also decreased cell capacitance, but the loss of membrane was not as great as with RTX and responses to these stimuli were not lost after the initial exposure. The loss of cell membrane required elevated intracellular levels of Ca2þ. From these data it was concluded that the loss of cell membrane was dependent on the presence of both VR1 and intracellular Ca2þ accumulation, but not on cell phenotype. # 2003 Elsevier Inc. All rights reserved.
Keywords: Vanilloid; Nociceptors; Neuroablation; Resiniferatoxin; Capsaicin; TRPV1
INTRODUCTION Resiniferatoxin (RTX), a potent tissue irritant and an agonist at vanilloid receptor type 1 (VR1), is isolated from the latex of several members of the genus Euphorbia (Szallasi and Blumberg, 1989). This agent is 1000 to 10,000 times more potent than capsaicin, a wellknown VR1 agonist found in chili peppers. RTX was found to evoke large inward currents in dorsal root * Corresponding author. Tel.: þ352-392-5661; fax: þ352-392-7609. E-mail address: [email protected] (R.M. Caudle).
ganglion (DRG) neurons and VR1 transfected cell lines (Szallasi et al., 1999). The actions of RTX are mediated by binding, with low nanomolar affinity, directly to the capsaicin-binding site on the VR1 receptor (Biro et al., 1997; Jerman et al., 2000). VR1 is located primarily on small and medium diameter nociceptive afferents (Cortright et al., 2001; Szallasi et al., 1993; Tominaga et al., 1998) and the receptors are thought to be rapidly desensitized by vanilloid agonists (Biro et al., 1997). The apparent desensitization to RTX is long lasting; whereas capsaicin produces only acute desensitization of the VR1 mediated currents (Liu and Simon, 1998). This observation has been confirmed in Ca2þ imaging
0161-813X/$ – see front matter # 2003 Elsevier Inc. All rights reserved. doi:10.1016/S0161-813X(03)00146-3
896
R.M. Caudle et al. / NeuroToxicology 24 (2003) 895–908
experiments where multiple short treatments with capsaicin-induced smaller and smaller transients in the cytosolic-free Ca2þ (Karai et al., unpublished observation). However, capsaicin responses in these experiments completely recover within 30 min of washing out the agonist, whereas they do not recover with RTX. If sufficiently high doses of vanilloid agonists are used in vivo, selective loss of VR1 expressing neurons is possible (Goso et al., 1993). RTX was reported to induce long-term loss of response to thermal stimuli when applied systemically or epidurally in adult rats (Pan et al., 2003; Szabo et al., 1999) and systemic administration of high doses of capsaicin lesion C-fibers in neonates (Hammond and Ruda, 1989). When capsaicin is repeatedly applied to the skin of human volunteers desensitization to thermal stimuli is observed (Fuchs et al., 2000). Nolano et al. (1999) and Pini and Lynn (1991) demonstrated that long-term peripheral capsaicin treatment produced a local Cfiber denervation. Recovery of peripheral thermal responses was due to the re-innervation of the tissue. Overall, these data indicate that vanilloid agonists can be neurotoxic if the neurons are subjected to prolonged exposure to high concentrations of the agonist and suggest that the lesions are selective for a subset of neurons. Our recent work indicates that RTX is highly toxic to DRG neurons expressing VR1. The toxicity leads rapidly to damage to membranes in intracellular organelles (Olah et al., 2001a,b) and preliminary electrophysiological experiments indicated that this membrane damage also included the plasma membrane. Capsaicin and other activators of VR1 were not as damaging to the cells suggesting that RTX may have unique toxicological properties beyond its high potency. The vanilloid receptor is a member of the transient receptor potential (TRP) family, which is a group of receptors that are activated by a variety of stimuli ranging from protons to mechanical perturbations of the membrane (Gunthorpe et al., 2002; Montell et al., 2002). The vanilloid receptor is the only member of this family that responds to compounds with the vanilloid moiety in their structure such as capsaicin, the pungent agent in hot chili peppers, and resiniferatoxin. This receptor is believed to be responsible for detecting potentially damaging stimuli in the periphery in the form of heat and protons. Because the receptor’s function can be significantly enhanced by a number of inflammatory mediators it is also likely to play a role in sensitizing nociceptive afferent fibers that innervate an injured area (Caterina et al., 2000). In addition, recent work has suggested that when the pH of tissue is
low, as in inflammation, anandamide can act as an endogenous agonist at the capsaicin binding site (Olah et al., 2001a,b). These data indicate that VR1 is an important integrator of a variety of nociceptive signals in the peripheral nervous system. Recent work has demonstrated that DRG neurons that express VR1 have several distinct phenotypes (Petruska et al., 2000a,b). These cells are both immunohistochemically and electrophysiologically distinguishable and likely subserve different sensory functions. Since RTX administration in vivo results in a loss of thermal sensitivity with no detectable loss of other senses (Pan et al., 2003; Szabo et al., 1999) it is possible that some VR1 expressing neurons are not as susceptible to damage from RTX as others. If a particular DRG cell type had relatively high levels of calcium binding proteins and low expression of VR1 it might not succumb as easily to RTX. Therefore, the questions addressed in this study were whether RTXs toxicity was limited to a subset of VR1 expressing cells and whether RTX was uniquely neurotoxic when compared to other activators of VR1.
MATERIALS AND METHODS Dorsal Root Ganglion Whole Cell Patch Clamp Adult male Sprague-Dawley rats (90–180 g) were anesthetized with halothane. Following decapitation, the spinal cord was rapidly removed and the dorsal root ganglia were dissected free. All animals were housed in AAALAC approved quarters, and all procedures were reviewed and approved by the University of Florida Institutional Animal Care and Use Committee. Dissected ganglia were placed in a tube containing dispase (neutral protease, 5 mg/ml; Boehringer Mannheim) and collagenase (2 mg/ml; Sigma type 1). The tube was shaken for 60 min in a heated (358) bath. Following wash and trituration, recovered cells were plated on 10 polylysine coated Petri dishes. All recordings were made at room temperature and completed within 10 h of plating. Glass pipettes (Scientific Products B4416-1) were prepared (2–4 MO) with a Brown and Flaming type horizontal puller (Sutter model P87). Whole cell recordings were made with an Axopatch 200B (Axon Instruments). Stimuli were controlled and digital records captured with pClamp 8.0 software and Digipack 1320A ac/dc converter (Axon Instruments). Series resistance was compensated 40–60% with Axopatch 200B compensation circuitry. The ‘‘membrane test’’
R.M. Caudle et al. / NeuroToxicology 24 (2003) 895–908
897
Fig. 1. Electrophysiological characterization of dissociated dorsal root ganglion cells. DRG cells were isolated as described in the methods and voltage clamped. The cells were then classified using the described classification protocols. The traces on the left are representative traces demonstrating the response of type 1, 2, 5 and 4 neurons to classification protocol 1 (CP1) and CP2 for the type 5 neuron. CP2 is not shown for type 1, 2 and 4 cells since it is not needed for classification of these cell types (see Petruska et al., 2000a,b). Cell types 1, 2 and 5 are VR1 expressing cell types, whereas type 4 cells do not express VR1. The traces on the right are action potentials obtained in current clamp mode and evoked by a brief (1 ms, 2–3 nA) depolarizing current injection.
898
R.M. Caudle et al. / NeuroToxicology 24 (2003) 895–908
protocol was used to track cell capacitance, whole cell resistance and access resistance. Following seal and break-in, cells were permitted to equilibrate for approximately 3 min. Subsequently the cell was classified as type 1, 2, 5 or 4 as described below. DRG types 1, 2 and 5 are capsaicin sensitive nociceptive neurons, whereas type 4 cells are capsaicin insensitive neurons (Fig. 1) (Petruska et al., 2000a). Following classification, cells were brought into current clamp mode. The resting membrane potential was assessed and an action potential was evoked by current injection (2–3 nA, 1 ms). After returning to voltage clamp, cell size was determined using the membrane test. RTX (0.01–100 nM), capsaicin (5 mM) (Sigma, St. Louis, MO), or vehicle containing the identical concentration of ethanol was presented by a 1 mm diameter pipette placed near the cell. Cell size was assessed immediately afterward and again after a 2min wash with Tyrode’s solution. Recordings were made exclusively from cells with diameters between 25 and 40 mm. Cell diameter was estimated from the average of the longest and shortest axis as measured through an eyepiece micrometer scale. Cells were classified according to patterns of voltageactivated currents (current signatures) that were revealed by three classification protocols (see Petruska et al., 2000a,b for detailed descriptions). In the studies described below classification protocols 1 (CP1) and 2 (CP2) were sufficient to classify the four afferent subclasses examined. With CP1, currents were evoked by a series of hyperpolarizing pulses presented from a holding potential (VH) of 60 mV (10 mV per step to a final potential of 110 mV; 500 ms, 4 s interstimulus interval). With CP2, strongly depolarizing command steps were used to produce outward current patterns. From a VH of 60 mV, a 500 ms conditioning pulse to 100 mV was followed by 200 ms depolarizing command steps (20 mV) to a final potential of þ40 mV. If CP1 evoked a transient outward current following repolarization to 60 mV, the cell was classified as type 2. If only leak currents were present, CP2 was applied. If no A-current peaks were observed with CP2, the cell was classified as type 1. Cells were classified as type 5 cells if small amplitude H-current (<100 pA) was revealed by CP1; and CP2 revealed three A-current deflections (Fig. 1). Type 4 cells were identified by large H-currents (>400 pA) during the CP1 protocol (Fig. 1). Plated cells were superfused in rat Tyrode’s solution containing (in mM): NaCl (140), KCl (4), MgCl2 (2), CaCl2 (2), glucose (10), HEPES (10), adjusted to pH 7.4 with NaOH. Test solutions were applied via gravity fed pipette positioned approximately 1 mm from the
cell. The recording electrodes were filled with (in mM): KCl (120), Na2-ATP (5), Na2-GTP (0.4), EGTA (5), CaCl2 (2.25), MgCl2 (5), HEPES (20), and adjusted to pH 7.4 with KOH; osmolarity was approximately 315–325 mOsm. Resiniferatoxin and capsaicin were dissolved in ethanol/water (1:1) and then diluted to their final concentration in Tyrode’s solution. HEK-293 Whole Cell Patch Clamp HEK-293 cells were seeded on to 13 mm cover slips at low density and then transfected with VR1 fused to enhanced green fluorescence protein (VR1eGFP) as described (Olah et al., 2001a,b). Twenty-four hours after transfection the cells were placed in a bath and superfused with Tyrode’s solution. The cells were viewed on an inverted microscope equipped with phase contrast and fluorescence optics. Cells expressing VR1eGFP were identified by their green fluorescence. The fluorescently labeled cells were whole cell patch clamped as described for DRG neurons. The holding potential was 60 mV. Cell capacitance was measured prior to treatment and 10–15 min following exposure to RTX, capsaicin or heating of the superfusion solution. Resiniferatoxin (100–125 pM) or capsaicin (5 mM) was applied via a gravity fed pipette placed within 200 mm of the patched cell. For experiments in which heat was used to activate vanilloid receptors, the bath temperature was elevated by using an inline heater (Warner Instrument Corp., Hamden, CT). The bath temperature was monitored with a miniature thermistor placed in the recording chamber. Fluorescent Confocal Microscopy COS7, NIH3T3 and HEK-293 cells were seeded on 25 mm coverslips and transfected with the VR1eGFP plasmid construct, as described previously (Olah et al., 2001a,b). Transiently transfected cells were cultured for an additional 24 h. To observe morphological changes of neuronal processes, dissociated DRG cultures were prepared from embryonic (E16) rats as described before in detail (Olah et al., 2001a,b). Briefly, after dissection and trituration DRG cells were seeded on 25 mm glass cover slips. Surfaces were coated with poly-D-lysine and laminin. Cultures were selected in a medium containing 2 and 5 mg/ml FUDR and uridine, respectively, for 4 days to 1 week, at which point well-differentiated neurons with elongated processes and non-dividing cells dominated the population. Primary DRG cultures at this stage were used for fluorescent Ca2þ imaging.
R.M. Caudle et al. / NeuroToxicology 24 (2003) 895–908
Calcium Imaging DRG cells were washed three times with imaging medium (IM) containing Hanks’ balanced salt solution, supplemented with 0.01% bovine serum albumin (BSA), 1.25 mM CaCl, 0.8 mM MgCl, 1 mM ascorbic acid, 1 mM pyruvate and buffered to pH 7.4 with HEPES. Cultures were loaded with 5 mM Fluo-4 AM dye. After incubation for 30 min at room temperature cells were washed three times in IM to remove excess dye and kept in the dark for at least 15 min before starting the experiments. Recordings were carried out in an open imaging chamber (Promega). Cells were visualized by an inverted Nikon microscope equipped with the MRC-1024 Bio-Rad Confocal System. Statistics Paired and un-paired t-tests, Pearson correlation and repeated measures ANOVA followed by Dunnett’s post-hoc test were used as indicated in the figure legends and text. The alpha level was set at 0.05.
899
RESULTS Effect of RTX in DRG Neurons To examine the influence of RTX on peptidergic and non-peptidergic, capsaicin-sensitive DRG neurons, patch clamp experiments were performed on type 1, 2 and 5 cells (Petruska et al., 2000a). Cell types 1 and 5 are known to express calcitonin gene related peptide (CGRP) and somatostatin, and CGRP and substance P, respectively. Type 2 cells do not express these peptides (Petruska et al., 2000a,b). Each of these cell types was previously demonstrated to respond to capsaicin (Petruska et al., 2000a,b). Once neurons were voltage clamped and typed (Fig. 1) whole cell capacitance was measured before and after the application of various concentrations of RTX. Analysis of the concentration response data for type 1 (N ¼ 5), type 2 (N ¼ 13) and type 5 (N ¼ 4) cells demonstrated that the cells did not differ in their responses to RTX. Thus, the data for these three cell types was combined. Fig. 2a demonstrates that RTX-induced large currents in these
Fig. 2. Effect of RTX on type 1, 2 and 5 DRG neurons. (a) Representative current traces from a type 2 DRG neuron while superfusing with 10 nM RTX. The horizontal line represents the 1-min RTX application. (b) RTX evoked current concentration response relationship. (c) RTX-induced change in capacitance concentration response relationship. (d) Effect of vehicle (1.7 mM ethanol) on RTX-induced changes in membrane capacitance. Asterisk indicates P < 0:05, paired t-test.
900
R.M. Caudle et al. / NeuroToxicology 24 (2003) 895–908
neurons. Washing out the RTX did not typically reverse the current, even when the cells were held for 15 min after the RTX administration. Even currents evoked by 10 pM RTX did not completely reverse. The lack of a full reversal of the current suggests that the neurons were damaged by the RTX treatment as previously described (Olah et al., 2001a,b). The current evoked by RTX was concentration dependent with an EC50 of 0:1 1:2 nM (Fig. 2b). In addition to the inward current, RTX reduced the capacitance of these DRG neurons in a concentration dependent manner with an EC50 of 0:2 3:5 nM (Fig. 2c). The EC50s for RTX-induced current and changes in capacitance were not statistically different. The change in capacitance indicates that the neurons had a decrease in surface area following RTX treatment. The vehicle used for delivering RTX (ethanol, maximum concentration ¼ 1:7 mM) did not induce currents nor did it change the capacitance of the neurons (Fig. 2d). To determine the time course of the change in capacitance the current and capacitance were monitored simultaneously in three cells. As demonstrated in Fig. 3a, the decrease in cell capacitance paralleled the
onset of the inward current. Plotting the current versus the capacitance for a representative type 2 neuron illustrates that the relationship between these measures is linear until the capacitance reaches a plateau (Fig. 3b). The current continues to increase following the plateau of the capacitance. The role of calcium in mediating the loss of membrane by the neurons during RTX administration was examined by removing the extracellular calcium from the superfusion buffer. In normal buffer the capacitance of DRG neurons was reduced by 100 nM RTX from 26 2:7 to 9 2:0 pF (N ¼ 3 neurons, P ¼ 0:026, paired t-test). In nominally Ca2þ-free buffer 100 nM RTX did not significantly influence cell capacitance. The capacitance before treatment was 27 1:4 and 37 8:5 pF after treatment (N ¼ 4 neurons, P ¼ 0:26, paired t-test). Interestingly, the maximum current RTX induced in nominally Ca2þ-free buffer was larger than in normal buffer (2:7 0:35 nA versus 1:4 0:33 nA, P ¼ 0:024, Student’s t-test). The response of type 4 cells, which lack vanilloid receptors (Petruska et al., 2000a), to RTX was also examined. Unlike type 1, 2 and 5 cells, RTX (100 nM) did not induce currents in type 4 cells (N ¼ 4) nor
Fig. 3. Relationship between RTX-induced current onset and capacitance changes in DRG neurons. (a) Current (top) and capacitance (bottom) record for a representative type 2 neuron demonstrating the rapid loss of membrane induced by RTX. (b) Plot of current versus capacitance for the above cell. The solid line is the regression of the linear portion of the data.
R.M. Caudle et al. / NeuroToxicology 24 (2003) 895–908
901
Fig. 4. Ca2þ fluorescence images of RTX-induced membrane blebbing in cultured DRG neurons. (a) Representative DRG neurons at the start of RTX (1 nM) superfusion. (b) Same neurons 2 min after RTX addition. The arrows indicate regions of membrane blebbing. Blebbing was observed in both the cell body and the processes.
did it have any effect on the cells’ capacitance (data not shown). To visualize the changes in the membrane induced by RTX, cultured DRG neurons were examined using Ca2þ imaging techniques. Cell types cannot be evaluated in this preparation; however, RTX-induced blebbing of DRG cell bodies and neuronal processes (Fig. 4) in approximately fifty percent of the cells. The membrane forming the blebs appears to be lost in vanilloid receptor containing neurons during the application of RTX. The blebbing in the cell bodies is consistent with the reduction in capacitance observed in patch clamp experiments.
Fig. 5. Effect of capsaicin (5 mM) on DRG neurons. (a) Capsaicininduced currents from a representative type 2 neuron. The horizontal lines indicate the duration of the capsaicin application. (b) The change in capacitance in type 2 DRG neurons after capsaicin exposure. Asterisks indicate P < 0:05, repeated measures ANOVA followed by Dunnett’s test, N ¼ 9 DRG neurons.
Effect of Capsaicin in DRG Neurons
Effect of RTX on VR1eGFP Transfected HEK-293 Cells
The application of capsaicin (5 mM) to type 1, 2 and 5 DRG neurons produced currents that were similar in size to 100 nM RTX (1:4 0:4 nA, N ¼ 5 cells), but the effects on the cells’ membranes were less dramatic. Repeated capsaicin applications produced currents with reduced amplitudes when compared to the initial capsaicin response (Fig. 5a), but complete loss of the current was not observed. In addition, the capacitance was reduced from 41:9 1:8 pF to a minimum of 37:5 1:7 pF (P < 0:05, repeated measures ANOVA followed by Dunnett’s test, N ¼ 9 DRG neurons, Fig. 5b). Interestingly, unlike RTX removing Ca2þ did not significantly influence the amplitude of the capsaicin-evoked current (1:7 0:4 nA, N ¼ 3 DRG cells, P ¼ 0:33, Student’s t-test 0Ca2þ versus normal buffer). Similar to the RTX experiments type 4 neurons demonstrated no response to capsaicin (data not shown).
To further evaluate the role of vanilloid receptors in decreasing cell size during RTX treatment HEK-293 cells were transfected with the vanilloid receptor C-terminally fused to enhanced green fluorescent protein (VR1eGFP). To quantify the changes in cell membrane properties VR1eGFP transfected HEK293 cells were whole cell voltage clamped as previously described. Exposure of these cells to 125 pM RTX resulted in a large inward current (Fig. 6a, N ¼ 9 cells). This current displayed the typical VR1 kinetics and had a reversal potential of approximately 0 mV as previously described (Caterina et al., 1997). Subsequent applications of RTX evoked little or no current indicating rapid and complete loss of RTX responsiveness. The RTX evoked current did not return even when the RTX was washed out of the bath for 30 min. Unlike the DRG cells, the RTX-induced current returned to baseline levels upon washout of the drug. HEK-293
902
R.M. Caudle et al. / NeuroToxicology 24 (2003) 895–908
Fig. 6. Effect of resiniferatoxin on HEK-293 cells transfected with VR1eGFP. (a) Representative current traces demonstrating that bath applied RTX (125 pM, 1 min) induces a large inward current on the first application, but subsequent applications are ineffective. The horizontal lines represent the application of RTX. The traces were collected 10 min apart. (b) A representative current trace from an eGFP transfected HEK-293 cell during RTX (125 pM, horizontal line) treatment. (c) Summary of RTX-induced change in capacitance. Asterisk indicates P ¼ 0:03, paired t-test, N ¼ 9 HEK-293 cells.
cells that were not transfected or were transfected with eGFP did not respond to RTX (Fig. 6b, N ¼ 5 cells). In addition to the rapid loss of the RTX mediated current, the cells’ membrane capacitance was reduced 10 min following the RTX exposure (Fig. 6c). The capacitance before RTX treatment was 20:9 3:7 and 14:6 2:6 pF after treatment (N ¼ 9 cells, P ¼ 0:03, paired t-test). The time constant for the decay of the capacitive charge was not significantly altered by the RTX treatment (t ¼ 0:26 0:05 ms before resiniferatoxin and t ¼ 0:86 0:4 ms following treatment, P ¼ 0:15, paired t-test, N ¼ 5 cells). Cells not transfected with VR1eGFP did not demonstrate this decrease in cell capacitance. To assess whether the amount of membrane lost is related to the number of VR1eGFP receptors in the membrane, the rate of onset of the RTX-induced current was plotted versus the maximum change in capacitance. Analysis of this relationship demonstrated that the two measures were correlated (r 2 ¼ 0:68, P < 0:01, Pearson correlation).
This analysis assumes that the rate of onset of the current, at a fixed concentration of agonist, is proportional to the number of VR1eGFP receptors in the membrane. Thus, this finding suggests that the number of VR1 receptors in the membrane determine the amount of membrane that will be lost following RTX treatment. Effect of Capsaicin on VR1eGFP Transfected HEK-293 Cells Application of capsaicin (5 mM) to the transfected HEK-293 cells resulted in large inward currents (Fig. 7a). In contrast to RTX, repeated application of capsaicin-induced currents, albeit they were smaller in amplitude than during the first exposure. These data were similar to those obtained from DRG neurons treated with capsaicin. Capsaicin treatment also reduced the capacitance of the VR1eGFP containing HEK-293 cells from 28:9 5:3 to 26:2 4:9 pF (mean S:E:,
R.M. Caudle et al. / NeuroToxicology 24 (2003) 895–908
903
Fig. 7. Effect of capsaicin and heat activation on VRleGFP transfected HEK-293 cells. (a) Currents from a representative HEK-293 cell exposed multiple times to capsaicin (5 mM). The horizontal lines indicate 1 min bath applications of capsaicin. (b) Effect of capsaicin on cell capacitance. Asterisks indicate P < 0:05, repeated measures ANOVA followed by Dunnett’s test on raw data, N ¼ 4 cells. (c) Top traces are heat-activated currents and the bottom traces are the corresponding records of the bath temperature. The traces were collected 5 min apart. (d) Summary of the heat-induced change in capacitance. Asterisk indicates P ¼ 0:02, paired t-test, N ¼ 7 cells.
N ¼ 4 cells, P < 0:05 repeated measures ANOVA followed by Dunnett’s test) (Fig. 7b). Effect of Heat on VR1eGFP Transfected HEK-293 Cells The vanilloid receptor was previously demonstrated to be activated by temperatures above 42 8C (Caterina et al., 1997; Davis et al., 2000; Kirschstein et al., 1999).
These studies did not report any significant cellular changes following heat. Thus, it was possible that the loss of membrane produced by RTX and capsaicin in DRG and VR1eGFP transfected cells was due specifically to the activation of the vanilloid receptor by vanilloid agonists and not simply due to the Ca2þ entry through the receptor when it was activated. To test this idea VR1eGFP transfected HEK-293 cells were patch clamped and the capacitance of the cells was measured
904
R.M. Caudle et al. / NeuroToxicology 24 (2003) 895–908
before and after heating the bathing solution to 47 8C with an inline heater. Fig. 7c illustrates that heating the bath-induced currents in the VR1eGFP containing cells. Although the currents decreased in amplitude following the first heat application a complete loss of response was not observed as was seen with the RTX treatment. This may have been due to the smaller currents induced by the heat (500 pA versus >1000 pA for RTX). However, increasing the maximum temperature to 508 resulted in the rapid loss of the cells, as did increasing the duration of the heat stimulus. Thus, larger currents could not be obtained. Even though complete loss of response to heat was not achieved, the cells’ membrane capacitance was significantly reduced from 43:8 6:6 to 36:6 6:6 pF (N ¼ 7 cells, P ¼ 0:02, paired t-test) (Fig. 7d). Role of Ca2þ in VR1-Induced Membrane Loss To confirm that the change in membrane capacitance produced by RTX, capsaicin and heat was due to Ca2þ entry into the cells, VR1eGFP transfected and nontransfected HEK-293 cells were exposed to the Ca2þ ionophore 4-bromo, 5-(methylamino)-2-[3,9,11-trimethyl-8-[1-methyl-2-oxo[6S-6-alpha-(2S*,3S*)], 8beta-(R*)], 9-beta, 11-alpha]-4-benzoxazole carboxylic acid (4-bromo A23187, Calbiochem, San Diego, CA) (1 mM) via the superfusion buffer while the cells were voltage clamped at 60 mV. In the transfected cells membrane capacitance was reduced from 23:6 5:0 to 18:8 3:8 pF (N ¼ 10 cells, P ¼ 0:03, paired t-test) 10 min after the removal of 4-bromo A23187. Non-transfected cells, on the other hand, did not demonstrate a change in capacitance. The capacitance was 73:0 40 pF before treatment and 88:7 40 pF after treatment (N ¼ 5 cells, P @ 0:05, paired t-test). These findings indicate that VR1 and Ca2þ are necessary for the observed membrane remodeling. This finding is consistent with our previously published data (Olah et al., 2001a,b). Note that although the baseline capacitance for VR1eGFP transfected cells was smaller than for non-transfected cells this difference was not statistically significant (P > 0:05, Students’ t-test). Confocal Microscopy of VR1eGFP Transfected Cell Lines Confocal fluorescence microscopy was employed to analyze the cellular distribution of VR1eGFP. Optical sections taken at the plane of cell attachment to the glass surface show VR1eGFP fluorescence in plasma
Fig. 8. Confocal images of representative COS7 cells transfected with VR1eGFP. Similar results were obtained with HEK-293 and NIH3T3 cells. (a) Control conditions prior to RTX application: N indicates the nucleus. (b) The same cells 1 min after 1 nM RTX: arrows indicate blebs. (c) The same cells 10 min after 1 nM RTX: arrows indicate blebs. Note some of the blebs have separated from the cells.
membrane processes (Fig. 8a). VR1 accumulation is also present in intracellular structures consistent with the ER, as previously described (Olah et al., 2001a,b). Fig. 8 illustrates the localization of VR1 in COS7 cells, but similar localization was seen in HEK-293 and NIH3T3 cells, indicating that it was not a cell type specific anomaly (Olah et al., 2001a,b). COS7 cells are
R.M. Caudle et al. / NeuroToxicology 24 (2003) 895–908
presented in Fig. 8 because their flat profile allowed visualization of membrane remodeling in a single plane. The more rounded HEK-293 cells require three-dimensional confocal imaging to follow the dynamics of the membrane. Exposure to 1 nM RTXinduced fragmentation of the VR1eGFP decorated ER and vesiculation of the plasma membrane processes in seconds (Fig. 8b versus Fig. 8a). Detached vesicles were seen by 10 min. In cells expressing only eGFP or the KDEL (lys–asp–glu–leu) signal linked eGFP plasmid the structure of the mitochondria and ER did not change upon exposure to vanilloids, such as RTX, olvanil or capsaicin, indicating the dependence of these effects on the presence of fully functional VR1 receptor. Thus, within the first few seconds after vanilloid exposure, coincident, structurally similar intracellular organelle remodeling occurs for both the ER and the plasma membrane processes.
DISCUSSION Resiniferatoxin is a potent analog of capsaicin, the active ingredient in hot chili peppers, that is isolated from the latex of several species of Euphorbia (Szallasi and Blumberg, 1989). RTX was previously demonstrated to produce loss of thermal nociceptive responses in rodents (Pan et al., 2003; Szabo et al., 1999). This effect appears to be different from the one that follows repeated application of low concentrations of capsaicin (Liu and Simon, 1998). The difference between RTX and capsaicin is due to a targeted destruction of DRG neurons that contain vanilloid receptors by RTX, whereas capsaicin appears to desensitize the neurons (Olah et al., 2001a,b). In the current study we demonstrate that RTX produces a rapid dismantling of the cell membrane of both DRG neurons and transfected cells that express the vanilloid receptor. If we assume that membrane capacitance is 1 mF/cm2 of membrane (Hille, 1991) then the decrease in DRG neuron capacitance indicates that approximately 2800 mm2 of membrane was removed from the surface of the neurons when they were treated with a maximally effective concentration of RTX. This represents a >50% decrease in cell surface indicating significant damage to the cells. Sensory neurons in the DRG are a diverse population of cells that have a variety of physical and biochemical properties. The different DRG cell types are believed to have distinct functions (Petruska et al., 2000a,b). In the present study we found no difference in sensitivity between DRG neurons classified as types 1, 2 and 5.
905
These three classes of nociceptive neurons express VR1. Type 4 nociceptive DRG neurons, which do not express VR1, were insensitive to vanilloid treatment. Thus, the presence of VR1 is required for RTX to damage the cells, but DRG cell type does not influence the response to RTX. In addition to the RTX-induced membrane loss we found that capsaicin and heat activation of vanilloid receptors also demonstrated decreases in capacitance. However, the degree of membrane loss was not as large as with RTX and the response to the stimulus was never fully lost as was observed with RTX. The total loss of membrane in DRG cells following multiple exposures to 5 mM capsaicin, a maximally effective concentration (Gunthorpe et al., 2000), was approximately 440 mm2. These data suggest that the presence of the vanilloid receptor and free intracellular Ca2þ are the main determinants for inducing cell membrane loss. The finding that 4-bromo A23187, a Ca2þ ionophore, induced a decrease in membrane capacitance only in cells containing VR1, in the absence of vanilloid agonists, supports this conclusion. The difference in the degree of membrane loss between RTX and the other stimuli is somewhat puzzling. The maximum current produced by RTX was no greater than the maximum current produced by capsaicin, yet there was a six-fold difference in the maximum amount of membrane lost. It is possible that RTX evoked opening of the VR1 ion channel leads to greater Ca2þ permeability when compared to the other stimuli or to the greater release of internal stores of Ca2þ. To the best of our knowledge, differences in ion permeability in vanilloid receptors have not been demonstrated for the various activators of the receptor. However, we have observed longer lasting elevations in intracellular-free Ca2þ following RTX when compared with capsaicin (Olah et al., 2001a,b). The prolonged elevation in Ca2þ may potentiate RTXs toxicity. An interesting observation in this study was that the maximum amplitude of the RTX current in DRG neurons was significantly reduced by the presence of extracellular Ca2þ, whereas the current evoked by capsaicin was not influenced by the ion. One possible explanation for this phenomenon is that because of the rapid loss of membrane with RTX in buffer containing Ca2þ vanilloid receptors were lost prior to their activation, thus reducing the maximum current that could be evoked. In the capsaicin experiments the loss of membrane was significantly less, which decreased the difference between Ca2þ-free and normal buffer. Confocal micrographs demonstrated that the membrane losses were the result of vesiculation and the
906
R.M. Caudle et al. / NeuroToxicology 24 (2003) 895–908
subsequent internalization/expulsion of the vesicles from the cell (Fig. 8). The decrease in capacitance occurs on a time scale (t50 20 s) that is intermediate between our previously reported time scales for endoplasmic reticulum vesiculation (t50 1 s) and nuclear blebbing (t50 5 min) (Olah et al., 2001a,b). Thus, the loss of plasma membrane appears to be the second phase of RTX-induced cell degradation. Possible Mechanisms for RTX-Induced Membrane Loss
were unable to protect DRG neurons from membrane loss with the free radical scavenger N-acetyl-L-cysteine (data not shown). These data suggest that oxidative/ nitrative stress is not involved in the acute membrane loss observed with RTX. Thus, the data from our two studies suggests that neither calpains, caspases nor oxidative/nitrative stress are involved in the mechanism for the VR1-induced Ca2þ mediated membrane loss. Thus, the mechanism for RTX-induced cell damage remains to be elucidated. Therapeutic Potential of RTX
The RTX-induced cell damage requires an increase in intracellular Ca2þ (present study; Olah et al., 2001a,b), but the pathways mediating the vesiculation of the plasma membrane are not known. Because the membrane is expelled from the cells, as well as internalized, it is not likely that the electrophysiologically observed loss of plasma membrane is due to typical receptor internalization as is observed with G-protein coupled receptors (Ferguson, 2001). In studies by other groups, similarities were found between excitotoxicity and both apoptosis and necrosis suggesting that there is a continuum between these two mechanisms of cell death (Chan and Mattson, 1999; Portera-Cailliau et al., 1997a,b). Similarly, we found a disruption of ER, nuclear and mitochondrial membranes (Olah et al., 2001a,b). However, chromatin fragmentation, apoptotic body formation, and caspase activity were absent indicating that the process was not apoptotic. Our observations on RTX treated DRG neurons are similar to those reported previously for excitotoxicity mediated by excitatory amino acids such as glutamate or aspartate. Faddis et al. (1997) suggested that channel openings by these amino acids induce varicosities by accumulation of excess intracellular Ca2þ similar to that demonstrated here. However, their data suggests that vesicularization is mediated by the local activation of calpain, a Ca2þ-dependent protease, in the vicinity of activated receptors. We tested the possibility that calpain was involved in the membrane loss by including calpain inhibitor peptide in the recording pipettes during some experiments. However, the peptide did not attenuate the loss of membrane in the VR1 containing HEK-293 cells (data not shown). This result suggests that calpain was not involved in the membrane loss, with the caveat that without a positive control it is also possible that the peptide did not inhibit calpain as expected. Jambrina et al. (2003) found that oxidative/nitrative stress-induced damage to Jurcat cells transfected with VR1. They described this as a paraptotic process. Similar to the calpain experiments we
Chronic neuropathic pain of peripheral origin is thought to be maintained by neuromas, or peripheral generators, that send a constant afferent barrage of nociceptive signals to the central nervous system. This excessive input to the spinal cord results in an enhancement of excitability in the spinal cord, an expansion of receptive fields, hyperalgesia, and allodynia that is often referred to as altered central processing (Eliav and Gracely, 1998; Gracely et al., 1993; Morris et al., 1997). The allodynia, derived from the altered central processing, is thought to be mediated to a large extent by the interpretation of Ab mechanoreceptor activity as pain by the central nervous system, rather than touch. The peripheral generator is likely to be injured primary afferent nociceptors. In support of this hypothesis, experimental altered central processing with the associated allodynia and hyperalgesia can be produced by an intradermal injection of capsaicin (Iadarola et al., 1998; Magerl et al., 1998; Morris et al., 1997). The activation of capsaicin sensitive primary afferents results in altered central processing of sensory information leading to the interpretation of Ab fiber activity as pain (Andersen et al., 1995). These data indicate that the peripheral generators in neuropathic pain might be targeted using the vanilloid receptors. The study by Szabo et al. (1999) demonstrates that application of RTX can eliminate [3 H]-RTX binding sites in the spinal cord and that the loss of these binding sites is associated with long lasting thermal analgesia. Thus, RTX may be useful in treating some forms of chronic pain. In summary, the present study demonstrates that RTX produces a rapid loss of cell membrane that is mediated by the heat/vanilloid-induced opening of the VR1 cation channel. The membrane loss is an early response in a Ca2þ-mediated toxicity that leads to cell death. The increased cytosolic Ca2þ vesiculates the ER, nuclear and plasma membranes with the expulsion of the membrane extra and intracellularly. The mechanism mediating the blebbing and vacuolization is not entirely known, but
R.M. Caudle et al. / NeuroToxicology 24 (2003) 895–908
does not appear to involve calpain, a Ca2þ activated protease, caspases, the principal executive enzymes of apoptosis (Olah et al., 2001a,b), or oxidative/nitrative stress. These findings do indicate that RTX rapidly and selectively disables VR1 containing cells. Because vanilloid receptors reside mainly on nociceptive primary afferents RTX may be a useful neuroablative agent for the treatment of chronic pain. ACKNOWLEDGEMENTS The National Institute of Dental and Craniofacial Research, National Institutes of Health supported this work. REFERENCES Andersen OK, Gracely RH, Arendt-Nielsen L. Facilitation of the human nociceptive reflex by stimulation of A beta-fibers in a secondary hyperalgesic area sustained by nociceptive input from the primary hyperalgesic area. Acta Physiol Scand 1995;155(1):87–97. Biro T, Acs G, Acs P, Modarres S, Blumberg PM. Recent advances in understanding of vanilloid receptors: a therapeutic target for treatment of pain and inflammation in skin. J Invest Dermatol Symp Proc 1997;2(1):56–60. Caterina MJ, Schumacher MA, Tommaga M, Rosen TA, Levine JD, Julius D. The capsaicin receptor: a heat-activated ion channel in the pain pathway. Nature 1997;389(6653):816–24. Caterina MJ, Leffler A, Malmberg AB, Martin WJ, Trafton J, Petersen-Zeitz KR, Koltzenburg M, Basbaum AI, Julius D. Impaired nociception and pain sensation in mice lacking the capsaicin receptor. Science 2000;288(5464):306–13. Chan SL, Mattson W. Caspase and calpain substrates: roles in synaptic plasticity and cell death. J Neurosci Res 1999;58(1): 167–90. Cortright DN, Crandall M, Sanchez JF, Zou T, Krause JE, White G. The tissue distribution and functional characterization of human vrl. Biochem Biophys Res Commun 2001;281(5):1183–9. Davis JB, Gray J, Gunthorpe MJ, Hatcher JP, Davey PT, Overend P, Harries MH, Latcham J, Clapham C, Atkinson K, Hughes SA, Rance K, Grau E, Harper AJ, Pugh PL, Rogers DC, Bingham S, Randall A, Sheardown SA. Vanilloid receptor-1 is essential for inflammatory thermal hyperalgesia. Nature 2000;405(6783):183–7. Eliav E, Gracely RH. Sensory changes in territory of the lingual and inferior alveolar nerves following lower third molar extraction. Pain 1998;77:191–9. Faddis BT, Hasbani MJ, Goldberg MP. Calpain activation contributes to dendritic remodeling after brief excitotoxic injury in vitro. J Neurosci 1997;17(3):951–9. Ferguson SS. Evolving concepts in G protein-coupled receptor endocytosis: the role in receptor desensitization and signaling. Pharmacol Rev 2001;53(1):1–24. Fuchs PN, Campbell JN, Meyer RA. Secondary hyperalgesia persists in capsaicin desensitized skin. Pain 2000;84(2–3):141–9.
907
Goso C, Piovacari G, Szallasi A. Resiniferatoxin-induced loss of vanilloid receptors is reversible in the urinary bladder but not in the spinal cord of the rat. Neurosci Lett 1993;162(1–2):197–200. Gracely RH, Lynch SA, Bennett GJ. Painful neuropathy: altered central processing maintained dynamically by peripheral input. Pain 1993;54(3):365–6. Gunthorpe MJ, Harries MH, Prinjha RK, Davis JB, Randall A. Voltage- and time-dependent properties of the recombinant rat vanilloid receptor (rVR1). J Physiol 2000;525(3):747–59. Gunthorpe MJ, Benham CD, Randall A, Davis JB. The diversity in the vanilloid (TRPV) receptor family of ion channels. Trends Pharmacol Sci 2002;23(4):183–91. Hammond DL, Ruda MA. Developmental alterations in thermal nociceptive threshold and the distribution of immunoreactive calcitonin gene-related peptide and substance P after neonatal administration of capsaicin in the rat. Neurosci Lett 1989; 97(1–2):57–62. Hille B. Ionic channels of excitable membranes. Sunderland, MA: Sinauer Associates Inc.; 1991. p. 9–10. Iadarola MJ, Berman KF, Zeffiro TA, Byas-Smith MG, Gracely RH, Max MB, Bennett GJ. Neural activation during acute capsaicin-evoked pain and allodynia assessed with PET. Brain 1998;121(5):931–47. Jambrina E, Alonso R, Alcalde M, Del Carmen Rodriguez M, Serrano A, Martinez AC, Garcia-Sancho J, Izquierdo M. Calcium influx through receptor-operated channel induces mitochondria-triggered paraptotic cell death. J Biol Chem 2003;278(16):14134–45. Kirschstein T, Greffrath W, Busselberg D, Treede RD. Inhibition of rapid heat responses in nociceptive primary sensory neurons of rats by vanilloid receptor antagonists. J Neurophysiol 1999;82(6):2853–60. Liu L, Simon SA. The influence of removing extracellular Ca2þ in the desensitization responses to capsaicin, zingerone and olvanil in rat trigeminal ganglion neurons. Brain Res 1998;809(2): 246–52. Magerl W, Wilk SH, Treede RD. Secondary hyperalgesia and perceptual wind-up following intradermal injection of capsaicin in humans. Pain 1998;74(2–3):257–68. Montell C, Birnbaumer L, Flockerzi V, Bindels RJ, Bruford EA, Caterina MJ, Clapham DE, Harteneck C, Heller S, Julius D, Kojima I, Mori Y, Penner R, Prawitt D, Scharenberg AM, Schultz G, Shimizu N, Zhu MX. A unified nomenclature for the superfamily of TRP cation channels. Mol Cell 2002;9(2): 229–31. Morris VH, Cruwys SC, Kidd BL. Characterization of capsaicininduced mechanical hyperalgesia as a marker for altered nociceptive processing in patients with rheumatoid arthritis. Pain 1997;71(2):179–86. Nolano M, Simone DA, Wendelschafer-Crabb G, Johnson T, Hazen E, Kennedy WR. Topical capsaicin in humans: parallel loss of epidermal nerve fibers and pain sensation. Pain 1999;81(1–2):135–45. Olah Z, Szabo T, Karai L, Hough C, Fields RD, Caudle RM, Blumberg PM, Iadarola MJ. Ligand-induced dynamic membrane changes and cell deletion conferred by vanilloid receptor 1. J Biol Chem 2001a;276(14):11021–30. Olah Z, Karai L, Iadarola MJ. Anandamide activates vanilloid receptor 1 (VR1) at acidic pH in dorsal root ganglia neurons and cells ectopically expressing VR1. J Biol Chem 2001b; 276(33):31163–70.
908
R.M. Caudle et al. / NeuroToxicology 24 (2003) 895–908
Pan HL, Khan GM, Alloway KD, Chen SR. Resiniferatoxin induces paradoxical changes in thermal and mechanical sensitivities in rats: mechanism of action. J Neurosci 2003;23(7):2911–9. Petruska JC, Napaporn J, Johnson RD, Gu JG, Cooper BY. Subclassified acutely dissociated cells of rat DRG: histochemistry and patterns of capsaicin-, proton-, and ATPactivated currents. J Neurophysiol 2000a;84(5):2365–79. Petruska JC, Cooper BY, Gu JG, Rau KK, Johnson RD. Distribution of P2X1, P2X2, and P2X3 receptor subunits in rat primary afferents: relation to population markers and specific cell types. J Chem Neuroanat 2000b;20(2):141–62. Pini A, Lynn B. C-fibre function during the 6 weeks following brief application of capsaicin to a cutaneous nerve in the rat. Eur J Neurosci 1991;3(3):274–84. Portera-Cailliau C, Price DL, Martin LJ. Excitotoxic neuronal death in the immature brain is an apoptosis–necrosis morphological continuum. J Comp Neurol 1997a;378(1):70–87. Portera-Cailliau C, Price DL, Martin LJ. Non-NMDA and NMDA receptor-mediated excitotoxic neuronal deaths in adult brain are
morphologically distinct: further evidence for an apoptosis– necrosis continuum. J Comp Neurol 1997b;378(1):88–104. Szabo T, Olah Z, ladarola MJ, Blumberg PM. Epidural resiniferatoxin induced prolong regional analgesia to pain. Brain Res 1999;840:92–8. Szallasi A, Blumberg PM. Resiniferatoxina, phorbol-related diterpene, acts as an ultrapotent analog of capsaicin, the irritant constituent in red pepper. Neuroscience 1989;30(2): 515–20. Szallasi A, Goso C, Blumberg PM, Manzini S. Competitive inhibition by capsazepine of [3 H]resiniferatoxin binding to central (spinal cord and dorsal root ganglia) and peripheral (urinary bladder and airways) vanilloid (capsaicin) receptors in the rat. J Pharmacol Exp Ther 1993;267(2):728–33. Szallasi A, Szabo T, Biro T, Modarres S, Plumberg PM, Krause JE, Cortright DN, Appendino G. Resiniferatoxin-type phorboid vanilloids display capsaicin-like selectivity at native vanilloid receptors on rat DRG neurons and at the cloned vanilloid receptor VR1. Br J Pharmacol 1999;128:428–34.