INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

One of the most important characteristics of nociceptive afferents is their ability to display sensitization in response to an inflammatory stimulus. Inflammation results in the production or up-regulation of agents such as bradykinin and prostaglandins that individually can sensitize nociceptors (Beck and Handwerker 1974; Martin et al. 1987). These agents have been shown to work via intracellular signaling systems (reviewed in Levine and Taiwo 1994) in part to enhance the efficacy of nociceptor-specific TTX-resistant Na channels thereby increasing the receptor discharge (reviewed in McCleskey and Gold 1999). TTX-resistant channels are not the only target of the second-messenger activation since PGE2 has been shown to inhibit K⁺ currents (Evans et al. 1999) and to enhance capsaicin sensitivity (Lopshire and Nicol 1997; Pitchford and Levine 1991) in sensory neurons, both via PKA signaling.

Nerve growth factor (NGF) is also up-regulated in the skin as a consequence of inflammation (Weskamp and Otten 1987). Local administration of NGF has been found to sensitize nociceptive afferents to noxious heat (Rueff and Mendell 1996). In recent experiments we have demonstrated the likelihood that this occurs at least in part by its action on the VR1 receptor. This became evident when NGF was shown to acutely sensitize the response of dissociated dorsal root ganglion (DRG) cells to capsaicin (Shu and Mendell 1999a), an agent known to elicit the sensation of noxious heat (LaMotte et al. 1992). Since the VR1 receptor has also been demonstrated to respond to noxious heat (Tominaga et al. 1998), it was suggested that the sensitization of the response to capsaicin by NGF corresponded to its sensitization of the response to noxious heat (but see DISCUSSION).

In the present work we have further explored some of the properties of the NGF-induced sensitization of the capsaicin response. In our previous work with recordings from dissociated DRG cells in perforated patch mode, the NGF-induced sensitization was measured by demonstrating that the response to capsaicin was potentiated by about 100% if NGF (100 ng/ml) was included in the superfusing solution bathing the cells (Shu and Mendell 1999a). If the same concentration of NGF was introduced into the medium during the 10 min between two identical capsaicin puffs, the average test response (to the 2nd puff) was substantially elevated compared with the usual tachyphylaxis (Koplas et al. 1997; Petersen and LaMotte 1993; Shu and Mendell 1999a). This increased response was far greater than predicted from the 100% potentiation observed on the response to the initial capsaicin puff in the presence of NGF. This suggested that NGF affected tachyphylaxis in addition to potentiating the response. Here we have investigated the effect of NGF in greater detail by asking whether the magnitude of the effect of NGF on the second of two responses to capsaicin puffs in the absence of tachyphylaxis is similar to potentiation of the initial response. To eliminate tachyphylaxis, we carried out these experiments in Ca²⁺-free solution (Koplas et al. 1997). Under these conditions we were also able to determine the duration of NGF's effect on the capsaicin response. In a second group of experiments, we introduced inhibitors of certain specific signaling molecules known to be associated with actions of NGF or with sensitization of sensory neuron responses by molecules such as bradykinin or PGE2 to see whether these could affect the acute sensitizing effects of NGF.

Some of these results have been presented in abstract form (Shu and Mendell 1999b).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Cell culture

The procedures used for DRG cell culture are described in a previous paper (Shu and Mendell 1999a). Briefly, 4-5 wk old Sprague-Dawley rats, unselected as to sex, were killed in CO₂, and the thoracolumbar DRGs were quickly removed and minced under the dissecting microscope. After digestion in collagenase (1 mg/ml) at 37°C for 45 min, the cells were centrifuged for 5 min and resuspended in DMEM medium. The cells were triturated using flame-polished glass pipettes and put into the polylysine-laminin-coated dishes. They were maintained over night in DMEM medium with 10% fetal calf serum in a 5% CO₂ incubator at 37°C.

Whole cell patch-clamp recording

All recordings were obtained from small cells (<30 µm diam) at room temperature using the Axopatch 200B amplifier with 10-kHz filtering. Patch electrodes were typically 1-2 M. Amphotericin B was added to the electrode to obtain a perforated patch clamp with series resistance <50 M. The cells displayed resting membrane potential more negative than 40 mV and overshooting action potentials. All cells were clamped at 60 mV to measure the capsaicin current.

The external solution contained (in mM) 145 NaCl, 5 KCl, 2 CaCl₂, 1 MgCl₂, 10 HEPES, and 10 glucose, adjusted to pH 7.4. The normal pipette solution contained (in mM) 130 K-gluconate, 10 HEPES, 10 EGTA, 1 MgCl₂, 1 CaCl₂, and 2 Mg-ATP, adjusted to pH 7.4 with KOH. Capsaicin (1 µM) was applied (10 psi, 400 ms) through a puffer pipette located close the cell. Following the first application of capsaicin, either NGF, NGF with another agent, or an agent alone was bath applied immediately for 10 min after which a second, identical pulse of capsaicin was administered. The percent change in response was computed as [100 * (response 2/response 1)  100]. All solutions were buffered at pH 7.4 to prevent pH effects on capsaicin responses (Petersen and LaMotte 1993). Each culture dish was used for only a single neurotrophin experiment.

Drugs

The effects of the following cell-permeable inhibitors of intracellular signal transduction pathways were tested on NGF enhancement of the capsaicin response: PKA inhibitors H₈₉ (isoquinoline sulfonamide) (Calbiochem, La Jolla, CA) and PKAI_14-22 (Calbiochem), PKC inhibitor Bisindolylmaleimide I (Calbiochem), and MAP kinase inhibitor PD 98059 (Calbiochem). In each case the initial dosage was chosen to be higher than published values of K_i or IC₅₀ for the particular agonist but substantially below the K_i or IC₅₀ for other signaling molecules to reduce the possibility of nonspecific effects (Chijiwa et al. 1990; Toullec et al. 1991). If no effects were detected, the concentration was raised 20-25 times and was delivered through the patch electrode to determine whether the inability to measure an effect was related to dosage and/or to access to the cell. We were also guided where possible by reports of experiments in which these agents were shown to have effects on DRG cells (Ganju et al. 1998; Zhang et al. 2001). Other experiments were done in the presence Ca²⁺-free solution in the presence of EGTA with or without bis-(o-aminophenoxy)-N,N,N',N'-tetraacetic acid (BAPTA)/AM (Sigma; see Figs. 1 and 2).

View larger version (7K): [in this window] [in a new window]   Fig. 1. Example of 3 responses to brief capsaicin puffs delivered 10 min apart in 0-Ca + EGTA + bis-(o-aminophenoxy)-N,N,N',N'-tetraacetic acid (BAPTA)/AM. Note that the 2nd response was similar in magnitude to the initial one (no tachyphylaxis) but that when nerve growth factor (NGF) was added between the 2nd and 3rd capsaicin puffs, the response was enhanced substantially.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

NGF-induced sensitization of capsaicin response in the absence of tachyphylaxis

The effects of NGF on capsaicin responses were examined in Ca²⁺-free solutions to determine its effect on the test response in the absence of tachyphylaxis. In an initial experiment cells were placed in Ca²⁺-free medium with 1 mM EGTA. An initial capsaicin puff was followed 10 min later by a second capsaicin puff. NGF (100 ng/ml) was placed in the medium after the second capsaicin puff, and the response to a third capsaicin puff 10 min later was determined. Sensitization was measured as the ratio of a test response to the control response converted to a percent. An example of the response with this protocol is displayed in Fig. 1. As predicted from the results of previous studies (Koplas et al. 1997), we found no tachyphylaxis when the response to the second puff was compared with the response to the first, in contrast to the severe tachyphylaxis (mean, 83%) (Shu and Mendell 1999a) observed in every case in 2 mM Ca²⁺ (Fig. 2). However, when NGF was introduced into the medium, the response to the capsaicin puff was increased substantially in many cells.

View larger version (13K): [in this window] [in a new window]   Fig. 2. Percent change in the response of individual dorsal root ganglion (DRG) cells to capsaicin (ordinate) in different experimental conditions (abcissa). Each point represents the % change in the response to capsaicin pressure ejected from a nearby pipette 10 min after a prior identical puff of capsaicin. A value of 100% indicates that the 2nd (test) response was 0; a value of 0% indicates that the test response was equal in magnitude to the initial (conditioning) response. A value of +100% indicates that the test response was twice as large as the conditioning response. "Control (2 mM Ca)" (column 1) displays data previously published in Shu and Mendell (1999a). The NGF column displays data previously published in Shu and Mendell (1999a) as well as more recent data collected under identical circumstances. These were measured in a solution with 2 mM Ca2+ (column 2). In "Control" the solution was not changed during the 10-min interval. In "NGF" 100 ng/ml NGF was added to the solution. 0-Ca (column 3) denotes an experiment carried out in Ca2+-free solution with EGTA. In "0-Ca + NGF" (column 4) 100 ng/ml NGF was added during the 10-min interval. "0-Ca + BAPTA/AM" (column 5) denotes experiments in which intracellular Ca2+ was chelated, and "0-Ca + BAPTA/AM + NGF" (column 6) denotes experiments in which 100 ng/ml NGF was added to the Ca2+-free + BAPTA/AM solution during the 10-min interval. Note that the tachyphylaxis observed in Control is lacking in both Ca2+-free solutions (columns 3 and 5). Note also that NGF enhanced the test response even in Ca2+-free solution (columns 4 and 6) but that the % increase in the test response (measured in terms of the corresponding control values) was less under these conditions than in controls with 2 mM Ca2+. Dashed line at 0% (2nd response equal to 1st response) drawn for reference. Further discussion in text.

The responses over all cells studied in Ca²⁺-free solution are shown in Fig. 2. In Ca²⁺-free solution with EGTA, there was little tachyphylaxis (Fig. 2, column 3), but when NGF was added, the response to capsaicin was potentiated in several cells tested (Fig. 2, column 4; see statistical analysis in this section). To explore whether NGF could still potentiate the response to capsaicin even after internal Ca²⁺ was chelated (but see DISCUSSION), we added the permeable Ca²⁺ chelator BAPTA-AM (20 µM) to the Ca²⁺-free/EGTA solution. In a group of 10 cells studied under these conditions without NGF, the percent change in response to the second capsaicin puff averaged over all cells was +4 ± 11% (mean ± SD), indicating again that there was no tachyphylaxis under these conditions. The response of each cell is demonstrated in Fig. 2 (column 5), where it can be seen that the change in the capsaicin over all tested cells clustered around 0% (no change) in this solution. Most but not all the cells exhibited some increase in the response after NGF was added (column 6). The mean percent increase when NGF was added under conditions where tachyphylaxis was absent was +69 ± 24%, n = 36.

To examine these data statistically we log transformed the distributions of facilitation [log (facilitation + 101)] to eliminate skewness. A two-way ANOVA indicated a highly significant Ca²⁺ × NGF interaction (F = 6.0; df = 1, 60; P = 0.02), specifically that the ability of NGF to alter values of tachyphylaxis from those observed in their respective control conditions depended on Ca²⁺ levels (0 or 2 mM). Further analysis revealed that the mean observed change in tachyphylaxis after NGF was significant both in the presence of Ca²⁺ [F(1,35) = 15.6; P < 0.0004] and in its absence [F(1,25) = 5.8; P = 0.02]. Thus the significant Ca²⁺ × NGF interaction reflects the fact that NGF produced a significantly greater change in tachyphylaxis in 2 mM Ca²⁺ than in Ca²⁺-free solutions.

A subsequent two-way ANOVA for the data in Ca²⁺-free solution (columns 3-6 in Fig. 2) revealed no significant interaction with NGF (F = 1.20; df = 1, 49; P = 0.28), i.e., the effect of NGF was the same with or without BAPTA/AM. This effect of NGF on the change in the test capsaicin current was highly significant [F = 7.1; df = (1, 49); P = 0.01]. BAPTA/AM itself had no significant effect on the change in the test capsaicin current [F = 0.86; df = (1, 49); P = 0.36].

In some cases in Ca²⁺-free solution it was possible to hold a cell long enough to administer multiple capsaicin puffs. In several cases we managed to deliver three capsaicin puffs 10 min apart before the cell was lost. If NGF was introduced between the first two capsaicin puffs, the second response was elevated in four of six cases, and in these cases the third response was elevated even further, despite wash out of NGF after the second puff (Fig. 3A).

View larger version (19K): [in this window] [in a new window]   Fig. 3. Amplitude of response (ordinates in pA) to 3 capsaicin puffs given 10 min apart in Ca2+-free or 2 mM Ca2+ solutions. NGF (100 ng/ml) was placed in the bath between the 1st 2 puffs and was removed after the 2nd puff. In Ca2+-free solution, the response to the 3rd puff was generally enhanced more than the response to the 2nd one. In 2 mM Ca2+, the response to the 3rd puff was generally equal to or smaller than the response to the 2nd. The latter occurred especially in cases where the response to the 2nd puff was strongly potentiated.

In a few cases we were able to investigate the responses to even more successive capsaicin puffs. Examples of these are shown in Fig. 4. It can be seen that in the absence of NGF there was no consistent trend in amplitude changes, whereas if NGF was administered between the first and second puff, the response increased substantially and remained elevated despite removal of the NGF via wash out immediately after the second puff. Based on flow rate of the solution and volume of the chamber, we calculate that NGF concentration fell almost to 0 within 5 min, and so it was virtually eliminated from the solution within the 10-min wash out period. Thus the elevated response to capsaicin could persist for up to 1 h after a brief pulse of NGF.

View larger version (10K): [in this window] [in a new window]   Fig. 4. Response of 2 cells to repeated brief capsaicin applications. These records were obtained in Ca2+-free solution with EGTA. In the top trace the capsaicin puffs were delivered in the presence of saline at the intervals indicated (10 or 20 min). Note the lack of any consistent change in response during the 80 min in which the cell was tested. In the bottom trace NGF was applied for 10 min between the 1st and 2nd capsaicin puffs, and then capsaicin was washed out. Note the persistent elevation in the response to capsaicin that lasted about 70-90 min.

We also examined the response to a third puff of capsaicin in solutions containing 2 mM Ca²⁺ (Fig. 3B) to determine whether the effects of NGF administered only for the 10 min between the first and second puffs were as long-lasting as in the absence of Ca²⁺. Only experiments in which NGF had a measurable effect on the second response (i.e., no tachyphylaxis) were subjected to a third puff of capsaicin after NGF had been washed out for 10 min. Under these conditions the response to the third capsaicin puff was always smaller than the second response unlike the finding in Ca²⁺-free solution where the third and subsequent responses generally remained elevated despite the wash out of NGF (Fig. 3A). Furthermore, we noted a tendency for the third response to be depressed most severely with respect to the second response in cells where the second response was most elevated.

From these experiments we conclude that NGF can sensitize the capsaicin response in the absence of tachyphylaxis. The magnitude of the change is similar on the average to the potentiation of the initial response to capsaicin by NGF (Shu and Mendell 1999a) (see INTRODUCTION). However, the ability of NGF to overcome tachyphylaxis in normal Ca²⁺ decays much more rapidly than the enhancement of the response to capsaicin in Ca²⁺-free solution.

Intracellular signaling for NGF sensitization of the capsaicin current

Previous studies have implicated the MAP kinase-signaling cascade in the response of DRG cells to NGF, specifically its long-term effect on the NGF-induced expression of capsaicin receptors (Ganju et al. 1998). We have extended the investigation of the role of this cascade to the acute actions of NGF. In initial experiments we exposed the cells to 5 µM or 10 µM PD98059 in the bath and found no significant effect; i.e., the acute response to NGF was not statistically different from values observed in NGF alone. Administering 50 µM PD98059 through the patch electrode, a concentration that has been found to significantly inhibit the chronic effect of NGF on the sensitivity of DRG cells to capsaicin (Ganju et al. 1998), also had no effect on NGF's ability to acutely facilitate the response to the test capsaicin puff. With all doses of PD98059, there was a population of cells, comprising about one-third of the total, that did not respond to NGF in that the test response exhibited tachyphylaxis similar to that observed in controls in the absence of NGF.

To obtain further information about possible mediators of the NGF-induced sensitization of the capsaicin response, we also examined the effects of NGF when blockers of PKA were introduced into the medium. We did this because enhancing PKA activity with forskolin can sensitize the response to capsaicin (Lopshire and Nicol 1998) and because inhibition of the PKA signaling pathway can block sensitization of the capsaicin response by PGE2 or by forskolin (Hingtgen et al. 1995; Lopshire and Nicol 1998). We initially confirmed the findings of Lopshire and Nicol (1998) that forskolin (10 µM) in normal Ca²⁺ abolished the tachyphylaxis observed in normal saline (not illustrated). The effect of forskolin was similar to that of NGF (Shu and Mendell 1999a) except that it affected all cells unlike NGF, which affected only about two-thirds of the cells (see DISCUSSION). Dibutyryl cyclic-AMP (D-cAMP), another activator of PKA, exerted effects similar to forskolin on the five cells tested (not illustrated).

In further experiments we administered blockers of PKA [H₈₉ (500 nM) or PKAI_14-22 (20 nM)] to the cells for at least 15 min before we tested the effect of NGF on the response to capsaicin. These experiments were carried out in the presence of 2 mM Ca²⁺ (Shu and Mendell 1999a), and so it was expected that in the absence of NGF, the response to the second of two capsaicin puffs delivered 10 min apart would be substantially reduced (tachyphylaxis). In control experiments no significant effect of these PKA blockers was detected on either the initial capsaicin current or on the degree of tachyphylaxis (not illustrated). Thus this degree of inhibition of PKA did not influence the initial response to capsaicin.

We then examined the effects of these PKA inhibitors on the response to NGF. As noted previously, in the absence of PKA inhibitors (Shu and Mendell 1999a), NGF eliminated tachyphylaxis in approximately two-thirds of the cells tested. In the presence of the PKA inhibitors, a significant minority of cells exhibited no apparent response to NGF, and cells exhibiting a response to NGF did so in an attenuated manner, i.e., no very positive values of facilitation were observed (Fig. 5; see statistics below in this section). There was no difference in the effects of the two PKA antagonists used. Thus, although the effects of NGF are dependent on the availability of PKA, some response remained indicating either that PKA activity was not completely abolished or that parallel signaling pathways are involved in the response to NGF.

View larger version (16K): [in this window] [in a new window]   Fig. 5. Percent change in the response of individual DRG cells to capsaicin (ordinate) under different experimental conditions (abcissa). Column 1 is the same as column 2 in Fig. 2. Horizontal dotted line at 38% is placed at 3 SD above the mean of the response under control conditions (column 1 of Fig. 2, see text). Any cell with a response above this line is arbitrarily considered to be one that responded to NGF (see text and also Shu and Mendell 1999a). The PKA inhibitors (combined because there were no differences between results with them) reduced the magnitude of the response to NGF (column 2). The MAP kinase inhibitor (PD98059; columns 3 and 4) and PKC inhibitor (Bisindolylmaleimide I; columns 5 and 6) did not (see text for more quantitative analysis of these results). Note cells below dashed line exhibit control levels of tachyphylaxis (i.e., not apparently responsive to NGF). Further discussion in text.

In a third group of experiments, we applied Bisindolylmaleimide I (20 nM), a blocker of PKC to determine whether it would affect the degree of tachyphylaxis induced by NGF. This was prompted by previous experiments indicating that the effects of bradykinin on the capsaicin current are mediated by the PKC pathway (Cesare et al. 1999). As with PD98059, we found no effects (see statistics below in this section), even when the dosage was raised to 500 nM and administered through the patch electrode.

A similar percentage of cells (36 ± 7%, mean ± SE) with values of tachyphylaxis similar to those in untreated cells was found in each of the six treatment groups (Fig. 5). These cells were considered to be unresponsive to NGF (Shu and Mendell 1999a). However, their presence in each treatment group complicated the statistical analysis since it diluted any possible difference in the response of the different groups to NGF. Thus a one-way ANOVA of these data revealed no significant difference in effect of treatment with NGF and a signal blocker compared with treatment with NGF alone. The analyses were then performed on the 76 cells for which a percent change in facilitation greater than 38% [3 SD above the mean value in controls (Shu and Mendell 1999a)]. Because of skewness in the distributions of the responses within treatments, the values were transformed to a log scale setting the response = log (%facilitation +101).

Analysis of covariance where log (amplitude) was included as a covariate was performed to compare the responses of the six groups. This was done because the degree of facilitation was found to be negatively correlated with amplitude of the response to capsaicin (r = 0.3; P = 0.008). Analysis of covariance of the responses resulted in a significant between-treatment difference in log of facilitation responses across treatments [F(5,69) = 2.95; P = 0.018], indicating that the response in at least one treatment was different from the others. Subsequent analysis of covariance of results for five of the treatments (all but NGF + PKA antagonists) indicated that the mean responses for these treatments were equal [F(4,53) = 1.87; P = 0.13]. We then compared the mean facilitation (log %change) of these five treatments (adjusting for log amplitude) to the mean of the NGF + PKA antagonist response and found that this mean was different [F(1,73) = 5.89; P = 0.018]. Overall, we conclude that only PKA antagonists attenuated the response to NGF but that its action was only partial.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The purpose of these experiments was to investigate mechanisms underlying the acute effects of NGF on capsaicin currents to gain further understanding of mechanisms underlying NGF-induced sensitization of the response to noxious heat. A compelling reason underlying this approach is that the receptor for capsaicin, VR-1, has been proposed to act as a transducer for at least some forms of noxious heat (Caterina and Julius 2001; Caterina et al. 2000). A correspondence between sensitivity to capsaicin and noxious heat sensitivity of individual small-diameter DRG cells has been established, although the correlation is not perfect (Greffrath et al. 2001). Also, Nagy and Rang (1999) have demonstrated in single-channel recordings in inside-out patches from sensory neurons that, although some individual channels respond to noxious heat and capsaicin, a significant proportion respond to one and not the other. Furthermore, VR-1 knockouts still display some sensitivity to noxious heat (Caterina et al. 2000), and Davis et al. (2000) have reported that VR-1 knockouts are deficient in displaying hyperalgesia rather than in responding to noxious heat stimuli per se. Therefore treating capsaicin as a "surrogate" for noxious heat (Shu and Mendell 1999a) may be somewhat of an oversimplification. Nonetheless, because administration of capsaicin intradermally is known to elicit burning pain in humans, it is important to investigate mechanisms underlying its action. In addition, NGF has been shown to sensitize the behavioral response of the organism (Lewin et al. 1993; McMahon et al. 1995; Rueff et al. 1996; Woolf et al. 1994), and the electrophysiological response of individual nociceptive afferents (Rueff and Mendell 1996) to noxious heat. Thus these results potentially have very broad application to mechanisms involved in peripheral sensitization.

The VR1 receptor has a noxious heat threshold of about 45°C (Caterina et al. 1997), which is similar to that of C-fiber mechanoheat nociceptors (LaMotte and Campbell 1978). These nociceptors exhibit a diminished response to a second heat pulse with a time course of seconds to minutes after the first (reviewed in Meyer et al. 1994), qualitatively similar to the time course of tachyphylaxis observed in response to capsaicin. However, the tachyphylaxis in response to noxious heat is less in magnitude and duration than that in response to capsaicin pulses, presumably because activation of the receptor is not as intense.

In agreement with previous studies (Koplas et al. 1997), we found that eliminating external Ca²⁺ abolished tachyphylaxis (Fig. 2). Under these conditions or after chelating internal Ca²⁺ with BAPTA-AM in addition, NGF was still able to increase the magnitude of the response to capsaicin. The similarity of the findings under these two Ca²⁺-free conditions suggests that BAPTA/AM contributed little additional depletion of [Ca²⁺]_i; this may be related to the finding that putting cells in Ca²⁺-free solutions can deplete virtually all the intracellular Ca²⁺ within minutes (Wagenknecht et al. 1994).

A question that arises is whether the enhanced effect of NGF on the second of two capsaicin-evoked currents in the presence of Ca²⁺ is due to the Ca²⁺ itself or to some change associated with the tachyphylaxis. The latter seems more likely because when NGF was introduced into the medium with 2 mM Ca²⁺ in advance of the initial capsaicin puff, the response to capsaicin averaged about twice as large (103%) as in identical trials without NGF (Shu and Mendell 1999a). This is similar to the increase induced by NGF on the capsaicin response in Ca²⁺-free solution with no tachyphylaxis (70%), but significantly less than was found in 2 mM Ca²⁺ (850%) when NGF was introduced between the first and second capsaicin puff; i.e., the increase in response to NGF was larger when superimposed on tachyphylaxis. At present we are not able to determine the mechanisms underlying this association between the effect of NGF and the presence of tachyphylaxis.

The duration of NGF's effect on the capsaicin current also differed according to the availability of Ca²⁺. In Ca²⁺-free solution, the enhanced response always lasted at least 10 min beyond the withdrawal of NGF, and it could last up to 1 h. In 2 mM Ca²⁺, a 10-min exposure to NGF led to an immediate sensitization in the response to capsaicin, but 10 min later the response was depressed, generally below the amplitude of the initial response. Although these findings do not illuminate the mechanisms of interaction of NGF and Ca²⁺ in determining the duration of NGF-induced sensitization of the capsaicin response, they do allow some useful conclusions concerning the duration of NGF's action in behavioral experiments. Many investigators have shown that local application of NGF subcutaneously induces an almost immediate hyperalgesia (reviewed in Shu and Mendell 1999c) and that the effect lasts for at least several hours (Shu et al. 1999). The data in the present experiments, specifically the rapid response to NGF, support the notion that the onset of the behavioral hyperalgesia could be due to a direct effect of NGF on the DRG cell. It is difficult to be certain of this correspondence since the behavioral assessments cannot be made earlier than about 30 min due to the need for anesthesia during local NGF administration (Shu et al. 1999). However, the persistence of NGF's behavioral effect at physiological levels of Ca²⁺ is likely due to other mechanisms because of the tachyphylaxis under conditions of normal Ca²⁺. Since NGF degranulates mast cells, one possibility is that other substances such as serotonin and histamine might lead to later components of the effect (Lewin et al. 1994). Other cells such as postganglionic sympathetic neurons might also be involved (Andreev et al. 1995).

The failure of an inhibitor of MAPkinase to reduce the effects of NGF was unexpected since this signaling system is known to be an important effector pathway for NGF in DRG cells, specifically its ability to enhance capsaicin sensitivity when provided over a period of days (Ganju et al. 1998). The present experiment is quite different since it measured the acute effects of NGF on capsaicin sensitivity, involving a time course of minutes rather than days. Two very recent reports suggest the involvement of phospholipase C- in NGF-mediated potentiation of capsaicin effects (Chuang et al. 2001), and the acute effect of NT-3 on transmitter release at the neuromuscular junction via activation of the trkC receptor (Yang et al. 2001).

We carried out similar experiments with a PKC inhibitor (Bisindolylmaleimide I) known to block many isozymes of PKC, including the isozyme PKC, at the dosage employed in these experiments (Toullec et al. 1991). This isozyme has recently been suggested to be a major intermediate in NGF-induced behavioral hyperalgesia, both mechanical and thermal (Khasar et al. 1999). However, we found no effect on the NGF-induced sensitization of the capsaicin current (see also Chuang et al. 2001). One possibility is that the PKC-mediated effects of NGF on behavioral hyperalgesia are mediated by cells other than DRG cells. A possible candidate is mast cells that are known to be involved in NGF-elicited hyperalgesia (Lewin et al. 1994; Rueff and Mendell 1996). The release of histamine from mast cells is known to require PKC and a rise in intracellular free Ca²⁺ (Lin and Gilfillan 1992).

There are precedents for NGF effects being weakly blocked by inhibitors of PKA such as H₈₉ (Cai et al. 1999; Freeland et al. 2000; Kalman et al. 1990) and being unaffected by blockers of PKC (Freeland et al. 2000). Although these findings do not indicate how significant the effects of PKA signaling are on NGF action, they certainly suggest that further exploration of direct and/or indirect roles of PKA signaling in NGF-induced sensitization of DRG cell responses to capsaicin is warranted. The possibility that PKA signaling works in concert with other signaling mechanisms (e.g., see Gold et al. 1998), perhaps even ones whose elimination does not alter NGF-induced sensitization, also deserves further study. This might explain our inability to find a single antagonist that completely eliminated the acute effects of NGF on the capsaicin current, although it may also be that other pathways are more important than the ones tested here.

A role for PKA signaling in sensitization of the capsaicin receptor has been suggested previously by the demonstration that bath-applied forskolin and D-cAMP also sensitized the test response to capsaicin, and that PGE2-induced sensitization of the capsaicin response is blocked by inhibitors of PKA (Lopshire and Nicol 1998). Lopshire and Nicol (1998) further speculated that phosphorylation of the VR1 receptor via PKA might be responsible for the sensitization, a proposal supported by the finding of PKA-mediated enhancement of the capsaicin receptor open time (but see Lee et al. 2000). Recently, a desensitizing action of capsaicin on the response of sensory neurons mediated by PKA has been identified, specifically inhibition of voltage-gated sodium currents (VGSC) (Liu et al. 2001). However, the effect of NGF on the VGSC was not studied in the present experiments.

One of the assumptions underlying the statistical analysis of the effects of the signal molecule antagonists is that cells displaying values of tachyphylaxis similar to those observed in the absence of NGF were cells with no trkA receptors. We eliminated these cells from the statistical analysis because they would not be expected to differ in response from treatment to treatment and thus would be expected to dilute any differences among cells that were trkA positive. In a previous report (Shu and Mendell 1999a) we inferred that about 68% of all cells we sampled (<30 µm diam) were affected by NGF. Of the 96 cells tested here with NGF and a signaling molecule blocker in the presence of Ca²⁺, 61 (64%) exhibited a loss of tachyphylaxis as defined above. This agrees with the estimate from our previous work that about 65% of this cell population expresses trkA according to this physiological test. Unlike NGF, forskolin and D-cAMP eliminated tachyphylaxis in all cells tested. The more widespread effect of forskolin and D-cAMP is presumably the result of all cells having intracellular signaling molecules (PKA) activated by these molecules, i.e., they can bypass the trkA receptor.

The fact that NGF can influence capsaicin receptor function acutely and chronically via different intracellular signaling systems speaks to complexity of intracellular signaling mechanisms activated by neurotrophins (Kaplan and Cooper 2001). Another possible factor in determining these differences is the low affinity p75 receptor. The acute action of NGF studied in this paper reflects activation of the high-affinity neurotrophin receptor, trkA, because its effect on the capsaicin current is eliminated when K-252a is administered in the bath (Shu and Mendell 1999a). Although we have not directly tested the role of the p75 receptor, the fact that NGF can sensitize the response of p75 knockout mice to noxious heat (Bergmann et al. 1998) suggests that this low-affinity neurotrophin receptor is not a crucial participant in the acute NGF-induced sensitization of the capsaicin response. The role of p75 in determining the effect of chronic NGF on the response to capsaicin remains to be elucidated.