|Home | About | Journals | Submit | Contact Us | Français|
Intraplantar injection of endothelin-1 (ET-1) (1.5-10 μM) in the rat produces mechanical allodynia. Here we identify the receptor subtypes for ET-1, glutamate and CGRP critical to such allodynia. Antagonism of ETA or ETB receptors alone, by BQ123 or BQ788, respectively, only partially suppressed allodynia; the combined antagonists prevented allodynia, showing the involvement of both receptor subtypes. Co-injection of NMDA receptor antagonists, (+)MK-801 or D-AP5, with ET-1 also prevented allodynia. In contrast, co-injection of the CGRP1 antagonist CGRP8-37 attenuated only the later phase of allodynia (>30 min). A mechanistic basis for these effects is shown by ET-1’s ability to enhance basal release from cultured sensory neurons of glutamate and CGRP (2.4-fold and 5.7-fold, respectively, for 10 nM ET-1). ETA blockade reduced ET-1’s enhancement of basal CGRP release by ~80%, but basal glutamate release by only ~30%. ET-1 also enhanced the capsaicin-stimulated release of CGRP (up to 2-fold for 0.3 nM ET-1), but did not change capsaicin-stimulated glutamate release. Release stimulated by elevated K+ was not altered by ETA blockade, nor did blockade of ETB reduce any type of release. Thus, ET-1 may induce release of glutamate and CGRP from nerve terminals innervating skin, thereby sensitizing primary afferents, accounting for ET-1-dependent tactile allodynia.
The endogenous endothelin peptides participate in a remarkable variety of pain-related processes. The present results provide evidence for the participation of ionotropic glutamatergic receptors and CGRP receptors in the hyperalgesic responses to exogenous ET-1 and suggest clinically relevant targets for further study of elevated pain caused by release of endogenous ET-1.
When endothelin-1 (ET-1) is injected subcutaneously into the human forearm or rodent hind paw, at concentrations of 30 nM - 10 μM, it induces local nociceptive sensitization or hyperalgesia. 3, 4, 12, 38 This paper addresses two important questions about this sensitization. First, which of the two G-protein-coupled receptors for ET-1, ETA and ETB, is involved? Earlier studies reported contradictory findings; ET-1-induced mechanical hyper-responsiveness in mice and rats was either unaffected by ETA antagonists12 or was inhibited by antagonists of ETA and of ETB receptors.3,4 Second, what are the signaling mechanisms underlying the enhanced nociceptive responsiveness triggered by ET-1? Several published reports inform the second question. Previous investigations demonstrated that neither prostaglandins, sympathetic amines, or glucocorticoid-sensitive/pro-inflammatory cytokines contribute to mechanical hyper-nociception from exogenous ET-1 in rats.12 Such mechanical sensitization, however, is potentiated by a phosphodiesterase inhibitor, implying a positive role for cyclic AMP (cAMP), and also is attenuated by inhibitors of protein kinase C and of MAP kinases.12, 38 In addition, as we have recently reported, this sensitization is sustained, but not initiated, by local TRPV1 receptors.4 The mechanisms and pathways involved in the rapid initial development of tactile allodynia in response to local ET-1 are, however, not known.
Glutamate and calcitonin gene related peptide (CGRP) are likely candidates to mediate an enhanced responsiveness of peripheral nociceptors, including that caused by ET-1. Glutamate plays an important role in hyperalgesia27 by participating in both central and peripheral pain transmission.5 Ionotropic glutamate receptors are present on the peripheral terminals of small diameter primary afferents 8, 10, 34 and glutamate is released in the plantar skin following high threshold A delta- and C-fiber stimulation of the sciatic nerve.13 Injection of glutamate and of agonists for ionotropic glutamate receptors into the normal rat hind paw 5, 49 results in acute mechanical hyperalgesia, while local, but not systemic, administration of the NMDA receptor antagonist MK-801 attenuates inflammatory mechanical hyperalgesia induced by Freund’s complete adjuvant (CFA). 35 Furthermore, endogenous ET-1 is involved in the inflammatory hyper-nociception evoked by carrageenan, formalin (second phase), and CFA. 3, 6, 14 Finally, it is known that ET-1 can stimulate the efflux of glutamate from neuronal 18 and non-neuronal cells (rat astrocytes 45).
CGRP, a pro-inflammatory and vasodilatory peptide, is synthesized by and released in skin from one class of peripheral nociceptors. 2 CGRP also binds to receptors on pre-synaptic neurons in the dorsal horn of the spinal cord 37, and to receptors in peripheral sensory neurons.11, 40 Local injection of an antagonist of the CGRP1 receptor decreases CFA-induced mechanical allodynia in the masseter muscle in rats. 1 In many cases, pain and nociceptor sensitization occur under pathological conditions associated with endogenously released ET-1. 3, 21, 41 (see 30 for review) We therefore hypothesize that ET-1 injected locally into rat plantar hind paw evokes local release of glutamate and CGRP, which agents in turn sensitize nociceptive primary afferent fibers and result in mechanical hyperalgesia.
Experiments were performed on adult male Sprague-Dawley rats (220-250 g, Charles River, USA), housed 2 per cage under a 12:12 h dark:light cycle and were provided with food and water ad libitum. Animals were experimentally treated and cared for using policies and procedures approved by the Harvard Committee on Animals.
Behavioral assessments were made with individual animals freely moving on a flat surface enclosed by an inverted (24 cm × 46 cm) Plexiglas® cage. Animals were allowed to recover for 5 min from the sevoflurane anesthesia, used during the injection of ET-1 (see below), before the onset of measurements. Spontaneous hind paw flinches were recorded every 5 min for 70 min after 100 μM ET-1 injection.
Unrestrained rats were placed on an elevated plastic mesh floor (28 × 17.5 cm; 9.5 × 9.5mm openings) and allowed to habituate for 30 min before initial testing. Withdrawal threshold to mechanical stimulation was determined using calibrated von Frey hairs (VFH) applied perpendicular to the plantar surface of a hind paw through spacing in the mesh. The minimal value to achieve 100% response was considered the paw withdrawal threshold (PWT).
The animals were tested over 5 days before each experiment to obtain a stable baseline PWT to mechanical stimulation. Each VFH was applied once starting with a force of 0.07 g (1 g, in naïve rats) and continuing until a withdrawal response occurred (or 60 g, the cutoff value in naïve rats, was reached). This PWT was verified by testing with the next thicker VFH, which always caused paw withdrawal, and then with the next thinner VFH, that did not cause paw withdrawal. 4, 7 If a response occurred to the VFH that was initially just below threshold, then the procedure was repeated until the PWT was consistent. On the treatment day, baseline was assessed 30, 20 and 10 min before the first injection The lowest value of these 3 measurements, before drug injection, was taken as the baseline level against which changes were compared, because we often found that the baseline PWT was lower at the 2nd and 3rd tests compared to the first and we did not want to bias the results towards showing the desired effect, i.e., tactile allodynia caused by ET-1. PWT were measured at 20, 30, 40, 50, 60, 75 and 90 min after injection of 10 μM ET-1.
The following agents were initially made as stock solutions: ET-1 (AXXORA, LLC, San Diego, CA, USA) was dissolved in phosphate buffered saline, PBS (pH=7.4; Invitrogen Corp., Carlsbad, CA, USA) at a concentration 0.1 mg/0.2 ml and stored in aliquots for up to 1 week at -80°C. (+)MK-801 (Sigma Chemical Co., St. Louis, MO, USA) was dissolved (5 mM) in high purity de-ionized H2O; D-AP5 (Tocris Bioscience, St. Louis, MO, USA) - (10 mM) and CGRP8-37 (Tocris Bioscience) - (1.6 mM) were dissolved in PBS. The ETA-selective antagonist BQ-123 (American Peptide Co; Ki=3.3-22 nM for ETA vs. 1.5 μM for ETB 25), and the ETB-selective antagonist BQ-788 (American Peptide; Ki=1-100 nM for BQ-788, 26), were dissolved in PBS. The ETA-selective antagonist FR139317 (Tocris Bioscience; Ki =1 nM for ETA vs. 7.3 μM for ETB 46) was initially dissolved to 50 mM in DMSO. Stock solutions were stored at -80°C for up to 2 weeks. Prior to the injection, stock aliquots were diluted in PBS (pH 7.1-7.4) to the noted final concentration. The final, injected FR139317 solution contained 2% DMSO. Vehicle for the experiments using FR139317 also contained 2% DMSO.
ET-1 (10 μl) or its vehicle, PBS, or PBS + 2% DMSO (used for the vehicle control for FR139317), was injected subcutaneously (s.c.) into the mid-plantar hind paw, 1 cm distal from the heel using a 30-G needle attached to a 10 μl Hamilton microsyringe (Hamilton Co., Reno, NV, USA). Injection occurred under brief general anesthesia (0.5-1 min) which was accomplished with the rapidly reversible agent sevoflurane (Abbott Labs, N. Chicago, IL, USA). In this procedure ca. 0.3 ml of sevoflurane liquid was placed on cotton gauze in the bottom of a 50 ml conical centrifuge tube, and the rat was gently restrained as the open end of the tube was placed over its nose. Evident by the animal’s relaxation, usually within 10-15 sec, and recovery of the righting reflex occurs in <30 sec after the anesthetic-containing tube is removed.
Antagonists were always delivered in two injections, the first at 15 min prior to the ET-1 injection, and the second mixed with ET-1 (the concentrations and total doses are indicated in the Results and figure legends). The concentration of ET-1, in these and in control experiments, was 10 μM or 100 μM. During all experiments drug solutions were kept on ice. The ET receptor antagonists were administered at concentrations/doses which were previously shown to be selective and effective in in vivo studies. 4, 19, 24, 29, 31 (+)MK-801 and D-AP5 were shown in other in vivo studies 27, 44 to be effective without having a systemic or toxic action at the concentrations/doses used here.
Cells were dissociated from the dorsal root ganglia of 150-175 g male Sprague Dawley rats using collagenase and mechanical agitation. The dissociated cells were suspended in F-12 medium supplemented with 2 mM glutamine, 10% horse serum, 50 U/ml penicillin, 50 μg/ml streptomycin, 100 μg/ml normosin O™, 50 μM 5-fluoro-deoxyuridine, 150 μM uridine, and 250 ng/ml nerve growth factor except as indicated. Approximately 22,500 cells were plated per well of a 24-well Falcon culture plates pre-coated with poly-D-lysine and laminin. The cells were maintained in culture at 37° C in a humidified atmosphere of 3.0% CO2 in air. The culture medium was changed every 2-3 days. Release experiments were performed on sensory neurons after 8-10 days in culture. The cells were washed in HEPES buffer containing 25 mM HEPES, 135 mM NaCl, 3.5 mM KCl, 2.5 mM CaCl2, 1 mM MgCl2, 3.3 mM D-glucose, 0.001 mM phosphoramidon, and 0.1% bovine serum albumin (BSA), at pH 7.4, then incubated for 10 min in the same buffer to determine the basal level of glutamate or CGRP release. Subsequently, cultures were treated with HEPES buffer containing 10 nM ET-1 (AXXORA or American Peptide Co.), or ET-1 in the presence of 30 nM capsaicin (Sigma) or 50 mM KCl. Final 10 min incubation was again in HEPES buffer to ascertain that the levels of glutamate and CGRP were returning to baseline upon removal of stimuli. In all experiments a group of culture wells was used as controls, i.e. were exposed to the release protocol in the absence of ET-1. In another series of experiments ETA or ETB receptor antagonists, BQ-123 or BQ-788 (American Peptide Co., Sunnyvale, CA, USA), respectively, dissolved in PBS to final concentrations of 1 and 3 μM, were added with ET-1 to measure the resting transmitter release in the presence of ET-1 under the conditions of the blockade of either ET- receptor. The buffer was collected after each incubation and glutamate and CGRP content determined from the same sample.
The amount of immunoreactive CGRP (iCGRP) in samples was measured by radioimmunosssay as previously described.47 CGRP antibody was a generous gift from Dr. M. Iadarola, NIDCR, NIH, Bethesda, MD. The amount of iCGRP was determined by comparing the amount of radioactivity in unknown samples to those values of a standard curve, generated for each experiment. Capsaicin and ET-receptor antagonists at the concentrations used in these experiments did not alter the radioimmunoassay.
Glutamate was measured by HPLC separation and electrochemical detection of the OPA-mercaptoethanol derivative using a modification of the methods of Lindroth and Mopper 36 and Gamache et al.22 Two hundred microliter portions of the release samples, 200 μl of the release buffer alone to serve as a blank, or glutamate standards in 200 μl of the release buffer were each combined with 20 μl of 1.0 M perchloric acid containing 12.5 μM homoserine as an internal standard. After mixing, this solution was transferred to the top of an Ultrafree Biomax filter with a 10 kDa cutoff or a Microcon unit with a cellulose filter with a 30 KDa cutoff (both Millipore Corp., Bedford, MA) and centrifuged immediately at 8,000 × g for 10 min at 4° C. A 20 μl sample of the ultra-filtrate was transferred to the autosampler vials and placed in the autosampler (4° C). Precolumn derivitization occured by the addition of 15 μl of the OPA/β-mercaptoethanol reagent to the samples and mixing. One min after addition of the OPA/β-mercaptoethanol reagent, 27 μl of the mixture of sample-reagent was injected onto a 3 reverse phase C-18 column, 80×4.6 mm (HR-80 from ESA, Inc., Chelmsford, MA, U.S.A.). Separation of the derivatives was carried out isocratically with a mobile phase of 0.1 M sodium phosphate at pH 6.75 with 24% methanol pumped at a flow rate of 1.2 ml/min. The electrochemical detection was performed by an ESA Coulochem II system with a dual electrode (#5011) set at 300 mV on channel 1 to oxidize impurities and 600 mV on channel 2 to oxidize the amino acid derivatives. The ESA software, running on a HP Pentium IV computer, calculated and plotted standard curves based on the blank and 4 standards for each experiment and calculated the amount of glutamate in the sample based on the recovery of the internal standard and the standard curve. The maximum concentrations of substances added to the release buffer in the various experiments did not alter the glutamate assay.
All primary data are reported as means ± S.E.M. Raw data for PWT values in grams are shown in Figure 1A, but thereafter the “normalized” values for PWT are reported. Normalization, in this case, consists of dividing each rat’s PWT value, at any specified time after treatment, by the baseline PWT value determined before treatment (see above). By such normalization it is possible to combine the data for different sets of rats that receive the same treatment and yet may have disparate baseline values; the comparison of various treatments is also much clearer when the normalized data are shown, since the relative effects of the treatments can be directly compared. Friedman’s test was applied to compare the responses to mechanical stimulation over all times after ET-1 (with vehicle) to responses over all times after ET-1 + antagonists. Two-tailed Mann-Whitney U-tests were used to compare the normalized PWTs (given as % of baseline value) that occurred with ET-1+ antagonist to that of controls, ET-1+vehicle, at each time point. For all these tests, P<0.05 was considered significant.
Basal release of neurotransmitters from sensory neurons treated with 10 nM ET-1 was compared to the basal release from non-treated cells using Student t-test. All the other comparisons for basal (resting) and stimulated (by capsaicin or potassium) release of iCGRP or glutamate and effects of ET-1 at different concentrations versus appropriate controls were made using ANOVA followed by Dunnett’s post hoc test. P<0.05 was considered significant.
Tactile allodynia induced by 10 μM ET-1 injected into the rat plantar hind paw has been reported to be attenuated by antagonists of ETA and ETB receptors, BQ-123 and BQ-788, respectively. 4 The initial, rapid development of allodynia, however, was only partially inhibited by the pre-emptive administration of these agents, even though BQ-123 was given at a concentration (3.2 mM) known to completely inhibit acute pain (nocifensive flinching) that occurred in response to much higher ET-1 concentrations 24 and BQ-788 was given at a concentration (3 mM), higher than the one that strongly inhibited the local analgesia resulting from activation of the ETB receptor. 50
To verify that ETA blockade only partially inhibits ET-1-induced hyperalgesia, a different, more potent and specific tri-peptide ETA antagonist, FR139317, with ~10-fold higher potency than BQ-123 (see Methods) was used. FR139317, injected 15 min prior to and then together with 10 μM ET-1, suppressed but did not abolish the tactile allodynia induced by ET-1 (Figure 1), similar to the earlier results with the less potent BQ-123 4. Figure 1A shows the raw data, with PWT in grams, and these data, normalized to the baseline values (see Methods), are redrawn in Figure 1B. (All subsequent PWT results are shown normalized to the baseline.) Paw withdrawal thresholds after ET-1 + FR139317 were higher than those after ET-1 + vehicle (Friedman’s test, p<0.0001; Mann-Whitney U-test, P<0.05 at 10, 20, 30, 40 for raw data, Fig. 1A, and, additionally, at 75 min for the normalized data, Fig. 1B). In order to establish a role for ETB receptors during ETA blockade, we pre- and co-injected ETA and ETB antagonists BQ-123 and BQ-788 together, 15 min before and then together with ET-1. Consistent with a role of both ET receptors, the combination of antagonists to both types of ET receptors resulted in no overall change from the baseline value and in smaller changes in PWT compared to those from ET-1 plus the ETA receptor antagonist alone (data from Fig 1B compared to that in Fig 2; Mann-Whitney U-test gives P<0.05 at 40, 50, 60, 75 and 90 min).
To inhibit CGRP1 receptors, CGRP8-37, a competitive, peptide antagonist, was injected into the rat plantar paw prior to and then with 10 μM ET-1. hind Allodynia developed similarly to controls (ET-1 + vehicle) for the first phase, 10-50 min, but with the antagonist it recovered faster and at 60-90 min after ET-1 injection it was significantly less than that from ET-1 + vehicle (Figure 3; P<0.05, Mann-Whitney U-test).
The CGRP1 receptor antagonist, at the same dose, also was able to strongly suppress acute pain behavior from intraplantar 100 μM ET-1. Hind paw flinching was decreased by ~70% by CGRP8-37, with total flinches over 70 min equal to 38±11 with antagonist (n=9) vs. 134±14 after ET-1 + vehicle (n=7, P<0.0005, Friedman’s test), and with a maximum flinch frequency equal to 9±2 flinches/5 min with the ET-1 + antagonist vs. 50±6 flinches/5 min with ET-1 + vehicle. Biting/licking was also attenuated by CGRP8-37 , with 9±4 biting events occurring over 70 min with the CGRP1 antagonist + ET-1 vs. 22±3 biting events after vehicle + ET-1, observed over the same time period (P<0.005, Friedman’s test) (time-course not shown). (Due to the spontaneous pain behavior of hind paw flinching and biting and licking, it is not possible to evaluate PWT for at least 90 min after the rats have been injected with 100 μM ET -1.)
To evaluate the involvement of NMDA receptors in ET-1-induced allodynia, we tested effects of two different inhibitors, the non-competitive antagonist (+)MK-801 and the competitive antagonist D-AP5. Local injection of MK-801, 15 min before and then together with ET-1, strongly suppressed tactile allodynia over the entire 90 min observation period (P<0.0001, Friedman’s test), including PWT values at 20, 30 and 60 min (Mann-Whitney U-test; Figure 4A). Injection of D-AP5 also showed a strong suppression over all times (Friedman’s test, P< 0.0001) and specific differences at 20, 30, 50 and 60 min (Mann-Whitney U-test; Fig 4B). In contrast to these results, and those with CGRP antagonist (above), the acute behavioral pain responses, e.g., spontaneous hind paw flinching (unprovoked by VFH stimulation), that were caused by a 10-fold higher ET-1 concentration, 100 μM, were not significantly affected by D-AP5 (5 mM, total dose 100 nmol/paw). The total number of flinches over 70 min were 79±13 in rats injected with antagonist + ET-1 and 90±27 flinches in rats injected with vehicle + ET-1, (P>0.05, Friedman’s test, time-course not shown).
To control for systemic effects, the same dose of D-AP5 (n=6) was injected subcutaneously into the nuchal midline, 15 min prior to and then just before ET-1 (10 μM) was injected into the plantar hind paw. No differences in tactile allodynia were found for up to 2 h after ET-1’s injection, when compared to controls where PBS instead of D-AP5 was injected at the nuchal midline before the ET-1 injection (n=6) (P>0.05, Friedman’s test; data not shown).
Since the behavioral experiments showed an essential contribution of CGRP and glutamate receptors in local ET-1-induced sensitization, we sought evidence for a connection between ET-1 injection and the local activation of these receptors. Release of the endogenous ligands by cutaneous nerves seems a likely mechanism, as these neurons are known to synthesize and release both substances. 13, 32, 33 Rather than attempt to measure their release directly into skin, we used a model system of isolated sensory neurons to investigate the effects of ET-1 on release.
Primary cultures of sensory neurons isolated from adult rat DRG showed an ET-1-related enhancement of basal release of both glutamate (2.4-fold maximum increase, n=24) and CGRP (5.7-fold maximum) (Figure 5); the EC50 for ET-1 for this enhancement was ~0.5 nM (data not shown). Removal of extracellular Ca2+ or inhibition of ETA receptors with BQ-123 (3 μM) each suppressed ET-1-enhanced basal glutamate release by ~30% (Figure 6). In contrast, Ca2+ removal strongly suppressed ET-1-enhanced basal CGRP release, by ~95%, and 3 μM BQ-123 suppressed it by ~80% (Figure 6); the requirement for extracellular Ca2+ appears to parallel the involvement of ETA receptors in this effect of ET-1. The ETB receptor antagonist BQ-788 (at 1 and 3 μM) did not change ET-1-induced glutamate release, although the lower concentration significantly elevated CGRP release (Figure 6), a phenomenon for which we have no explanation.
Exposure of isolated sensory neurons to the TRPV1 agonist capsaicin (30 nM) stimulated CGRP release by ~20-fold (Figure 7A), but increased glutamate release by only ~ 1.8-fold (Fig. 7B). ET-1, which is known to enhance Ca+2 entry through TRPV1, 42, 48 enhanced capsaicin-stimulated iCGRP release by 50-100% when present at 0.3 -10 nM (n=3-9, Fig. 7A). In contrast, the capsaicin-stimulated release of glutamate was not affected by ET-1 (10 nM) (n=6 nor was there a significant difference between glutamate’s basal release in the presence of ET-1 and release with ET-1 and capsaicin (Fig 7B). Exposing neurons to 10 nM ET-1 significantly increased the basal release of glutamate, from 308 ± 5 pmol/well/10 min to 613 ± 18 pmol/well/10 min, in a manner analogous to the results observed in Figure 5.
Finally, ET-1 (10 nM) did not affect the iCGRP or glutamate release from neurons stimulated by high (50 mM) extracellular potassium (data not shown), implying that Ca+2 entry through pathways stimulated by depolarization is not modified by ET-1.
Acccumulating evidence implicates endogenous ET-1 in the pathogenesis of pain after many injuries and diseases. 30 In this report we show that tactile allodynia following subcutaneous plantar injection of ET-1 (10 μM) into the rat hind paw appears to be mediated by both sub-types of ET-receptors. Previous publications report mixed and often contradictory findings on ET receptor involvement. Inhibitors of both ETA and ETB receptors were found to be effective against mechanical hyperalgesia induced by ET-1 (2-24 μM) injected into the mouse plantar surface.3 However, mechanical hyper-nociception from 30-300 nM ET-1 was inhibited by ETB receptor and not by ETA receptor blockade, and such hypernociception was mimicked by ETB receptor agonists, indicating a pro-nociceptive role for the ETB receptor.12 Motta and colleagues 38 reported later that, despite the fact that the ETB -selective agonist sarafotoxin S6c was capable of inducing mechanical hypersensitivity similar to that from ET-1, both BQ-123 and BQ-788 were effective in preventing ET-1- induced mechano-sensitization in rats, implying that both receptor sub-types contribute to allodynia. Our data, using von Frey testing, confirms this last finding, that both sub-types of ET receptors contribute to mechanical sensitization from local ET-1 administered subcutaneously at micromolar concentrations in rats.
From a comparison of the receptors that appear to be involved in different responses to ET-1, it appears that ETB receptor activation mediates the sensitizing effects of local, nanomolar, exogenous ET-1. 12 In a similar manner, a pro-algesic role of ETB receptors is evident in inflammatory hyperalgesia that occurs without exogenous ET-1, e.g., experimental arthritis, 15 where local tissue levels of ET-1 could approach nanomolar concentrations. 6, 9 This pro-algesic role of the ETB receptor in tactile allodynia differs from the ET receptor contributions to overt, spontaneous nociception caused by high concentrations (≥100 μM) of exogenous ET-1. Indeed, overt pain from high [ET-1] is triggered overwhelmingly by ETA receptor activation, independent of the site of ET-1 administration. 15, 19, 24 In contrast, the activation of ETB receptors, by such high doses of ET-1or by selective agonists, lessens the pain, suggesting that the opposing effects of ETA and ETB receptor activation determine the final nocifensive behavioral response to high concentrations of ET-1. 30 The variable contribution of the ETB receptor, which is pro-algesic at low concentrations of ET-1 and anti-hyperalgesic at high concentrations, suggests that there is more than one ETB receptor sub-type or that the receptor is distributed on different types of cells whose respective contributions have opposing effects on pain behavior.
An example of opposing effects of what appears to be one type of ET receptor is seen from the vascular actions of the ETB receptor, which in vascular endothelial cells stimulates the release of nitric oxide (NO), thereby indirectly causing the relaxation of vascular smooth muscle, while in these same smooth muscle cells directly triggers their contraction (for review see 30).
ETA receptors have been found on a large proportion of the cell bodies of small diameter sensory neurons (DRGs) which are associated with C- and Aδ-fibers that carry pain impulses 43 and presumably are located on nociceptors in skin, 16, 31 while ETB receptors appear to be mainly expressed in DRG satellite cells and ensheating Schwann cells 43 as well as keratinocytes of skin. 31 Despite their often separate and distinct tissue distributions, ETA and ETB receptors also appear to be co-expressed in a number of cells, e.g., in vascular smooth muscle cells.30
The antagonist CGRP8-37 affected only the late phase of tactile allodynia evoked by ET-1, implicating CGRP in the maintenance but not the initiating mechanisms of sensitization. The decrease in nocifensive hind paw flinching and biting/licking of the ET-1-injected paw, effected by CGRP8-37 (respectively, ~3.5- and ~2-fold), shows that CGRP receptors mediate acute pain as well as tactile sensitization from ET-1. These findings are consistent with the demonstration of mechanical hyperalgesia in the rat paw after intraplantar injections of CGRP, 39 indicating that low endogenous levels of CGRP can sensitize nociceptors. Furthermore, CGRP8-37 delays the onset of tactile allodynia or reverses this phenomenon, when injected into the rat hind paw, respectively, immediately before or after ipsilateral L5 spinal nerve ligation, suggesting that CGRP is involved in the genesis and maintenance of injury-induced neuropathic pain. 28 It thus appears that local CGRP is able to sensitize peripheral nociceptor terminals in pathological conditions that are also characterized by endothelin-dependent hypernociception, 30 which supports the plausible participation of CGRP in hyperalgesic responses to endogenous, local ET-1.
The initial increase in ET-1-induced mechanical allodynia, which is unaffected by antagonists of TRPV1 4 or of the CGRP1 receptor, requires that some mediators released by ET receptor activation contribute to the development of allodynia. We hypothesized that glutamate, localized in the axons of primary afferent neurons and released in peripheral tissues by nerve stimulation, also could be released in skin by neuronal ET receptor activation. This glutamate would, in turn, activate a population of small diameter, peptide containing neurons which would then release CGRP, and perhaps other peptides. Our data demonstrate that inhibition of local NMDA receptors virtually abolished mechanical allodynia in response to ET-1 (see Fig. 4). In the rat glabrous and hairy skin and human hairy skin, ionotropic glutamate receptors are expressed in the peripheral terminals of both unmyelinated C-fibers and thinly myelinated A delta-fibers at the dermal-epidermal junction. 8, 10, 17 Injection of glutamate into the cutaneous receptive fields of nociceptors can induce significant spontaneous discharge, as well as sensitize the nociceptors to subsequent stimulation. 16, 17 In studies of inflammatory hyperalgesia, the non-competitive NMDA-R antagonist (+)MK-801 attenuated carrageenan-induced thermal hyper-responsiveness.27 A similar result was observed for mechanical hyperalgesia evoked by CFA, wherein local injection of MK-801 rapidly returned mechanical sensitivity of the rat hind paw inflamed by CFA to baseline. 17 Collectively, these observations are evidence that peripheral NMDA receptors are activated similarly by exogenous ET-1 injection and by local inflammation, and in both cases contribute to local sensitization of peripheral nociceptors, and hence, hyperalgesic responsiveness of the skin.
In experiments on isolated sensory neurons in culture, ET-1 (10 nM) increased the basal release of glutamate (by ~2.4 fold) and of iCGRP (by ~5.7 fold). (Earlier work on the effect of ET-1 on release of CGRP and Substance P 18 had produced similar results, but using much higher concentrations of ET-1 (0.5-2 μM) than in the present study.) The elevation of release of either substance by ET-1 was inhibited by blockade of ETA receptors, weakly for glutamate release but strongly for CGRP release, and these inhibitions paralleled the dependence of these substances’ release on extracellular Ca2+. Inhibition of ETB receptors was ineffective on basal glutamate release but slightly enhanced basal release of iCGRP, an unexpected finding since ETB receptors had previously been localized exclusively on satellite and Schwann cells, and never on neurons. 43
The large, capsaicin-stimulated release of CGRP was enhanced by ET-1, while the much more modest capsaicin-stimulated release of glutamate was unaffected. Potentiation of capsaicin’s effects probably results from the enhancement of Ca2+ influx through TRPV1 channels that have been modulated by ET-1. 42, 48 The difference in response to ET-1 between capsaicin-stimulated iCGRP release and glutamate release parallels the differential sensitivity to ET-1 of basal release, in the absence of TRPV1 activation, and suggests that Ca2+ entry by any path across the plasma membrane is more effective for stimulating the release of CGRP than that of glutamate. That the stimulated release of either substance by elevated [K+]o was not altered by ET-1 treatment implies that the direct entry of Ca2+ through voltage-gated calcium channels in the plasma membrane is not modified by ET-1.
Local ET-1, acting via ETA receptors, is presumably capable of inducing release of these substances from fiber terminals innervating the skin. In this regard, several endogenous mediators and algogens released in inflamed tissue can act directly on nociceptors to induce neuropeptide release from the peripheral nerve endings.20, 23, 47 The release of CGRP increases after incubation of rat skin with inflammatory mediators 2 or superfusion with capsaicin. 32 Similarly, peripheral primary afferents contribute to the release of glutamate in rat skin, induced by capsaicin. 13
In conclusion, the inhibitory effects of NMDA and CGRP1 receptor antagonists on ET-1-evoked mechanical allodynia, and ET-1-induced enhancement of basal glutamate and iCGRP release from primary sensory neurons, provide evidence for the participation of ionotropic glutamatergic receptors and CGRP receptors in the pain sensitizing responses to ET-1 in the rat paw. We hypothesize that activation of ETA receptors on nerve fibers by subcutaneous delivery of exogenous ET-1 provokes an initial release of glutamate and, less importantly, neuropeptides from peripheral terminals innervating skin. These substances diffuse within the epidermis and in turn sensitize other primary afferents to non-noxious stimulation, accounting for tactile allodynia. We predict that similar effects underlying tactile allodynia from injury or inflammation would be induced by endogenously released ET-1.
We thank Mr. Jamie Bell for excellent technical assistance with the figures. Supported by USPHS grants CA080153, and NS048565. Some of the work was conducted in a facility constructed with support from Research Facilities Improvement Program Grant Number C06 RR015481- 01 from the National Center for Research Resources, National Institutes of Health.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.