Effect of peptidases on the ability of exogenous and endogenous neurokinins to produce neurokinin 1 receptor internalization in the rat spinal cord

Abstract

1. The ability of peptidases to restrict neurokinin 1 receptor (NK1R) activation by exogenously applied or endogenously released neurokinins was investigated by measuring NK1R internalization in rat spinal cord slices.

2. Concentration–response curves for substance P and neurokinin A were obtained in the presence and absence of 10 μM thiorphan, an inhibitor of neutral endopeptidase (EC 3.4.24.11), plus 10 μM captopril, an inhibitor of dipeptidyl carboxypeptidase (EC 3.4.15.1). These inhibitors significantly decreased the EC₅₀ of substance P to produce NK1R internalization from 32 to 9 nM, and the EC₅₀ of neurokinin A from 170 to 60 nM.

3. Substance P was significantly more potent than neurokinin A, both with and without these peptidase inhibitors.

4. In the presence of peptidase inhibitors, neurokinin B was 10 times less potent than neurokinin A and 64 times less potent than substance P (EC₅₀=573 nM).

5. Several aminopeptidase inhibitors (actinonin, amastatin, bacitracin, bestatin and puromycin) failed to further increase the effect of thiorphan plus captopril on the NK1R internalization produced by 10 nM substance P.

6. Electrical stimulation of the dorsal root produced NK1R internalization by releasing endogenous neurokinins. Thiorphan plus captopril increased NK1R internalization produced by 1 Hz stimulation, but not by 30 Hz stimulation.

7. Therefore, NEN and DCP restrict NK1R activation by endogenous neurokinins when they are gradually released by low-frequency firing of primary afferents, but become saturated or inhibited when primary afferents fire at a high frequency.

Introduction

The neurokinin 1 receptor (NK1R) is one of three receptors that recognize the neurokinins (or tachykinins) substance P (SP), neurokinin A (NKA) and neurokinin B (NKB) (Regoli et al., 1994; Routh & Helke, 1995). There is ample evidence that NK1Rs in dorsal horn neurons contribute to the development of hyperalgesia and central sensitization. For example, NK1R antagonists block hyperalgesia (Traub, 1996; Henry et al., 1999), and hyperalgesic responses are absent in transgenic mice lacking the NK1R (De Felipe et al., 1998; Laird et al., 2000; 2001) or after selective elimination of dorsal horn neurons possessing NK1Rs (Mantyh et al., 1997).

NK1Rs are present throughout the spinal cord, except in lamina II neurons (Bleazard et al., 1994; Liu et al., 1994). In particular, many lamina I neurons with NK1Rs send axons to brain regions involved in pain processing, while receiving synapses from SP-containing primary afferents (Todd et al., 2002). Therefore, synapses between SP-containing primary afferents and NK1R-containing projection neurons appear to be crucial in determining the intensity of pain signals. Owing to this, mechanisms that control the release of neurokinins and their ability to bind to NK1Rs are extremely important. Indeed, neurokinin release is modulated by a staggering number of presynaptic receptors, including GABA_B, α₂-adrenergic, opioid, cannabinoid 1, vanilloid 1, NMDA and serotonin receptors (for references, see Marvizon et al., 2003).

The access of neurokinins to NK1Rs could also be controlled by their degradation by peptidases. In particular, neutral endopeptidase (NEN, EC 3.4.24.11) and dipeptidyl carboxypeptidase (DCP, EC 3.4.15.1) appear to degrade neurokinins in the dorsal horn, because inhibitors of these enzymes increased the amount of released SP detected using antibody microprobes (Duggan et al., 1992). Other studies indicated that aminopeptidases (EC 3.4.11.) are also able to degrade SP (Harbeck & Mentlein, 1991; Mauborgne et al., 1991). Inhibitors of NEN, DCP and aminopeptidases increased the ability of SP and NKA to produce scratch, bite and licking behavior (Sakurada et al., 1990) and ventral root depolarizations (Suzuki et al., 1994).

NEN, DCP and aminopeptidases also efficiently degrade opioid peptides (Guyon et al., 1979; Chou et al., 1984; Yaksh & Chipkin, 1989; Hiranuma et al., 1998; Roques, 2000). Indeed, we found that inhibitors of NEN, DEC and aminopeptidase N increased 10 and 100 times, respectively, the potencies of dynorphin A and Leu-enkephalin to produce μ-opioid receptor internalization (Song & Marvizon, 2003). It has been suggested that peptidase inhibitors could be used to treat pain, as they have analgesic effects (Noble et al., 1992b) and do not produce tolerance and addiction (Noble et al., 1992a, 1992c; Roques, 2000). However, if peptidase inhibitors also protect neurokinins against degradation, this would produce pronociceptive effects and even long-term hyperalgesia that would compromise any analgesic effects. Therefore, it is necessary to compare the relative efficacies of peptidase inhibitors in protecting the degradation of neurokinins and opioids.

The main goal of this study was to evaluate the ability of peptidases to prevent the activation of NK1 receptors by neurokinins, either exogenously applied or endogenously released. NK1R activation was measured by their internalization, a technique validated in numerous studies (Mantyh et al., 1995; Allen et al., 1997; Liu et al., 1997; Marvizon et al., 1997; 1999a; 2003; Cao et al., 1998; Honore et al., 1999; Riley et al., 2001). An additional goal of this study was to compare the potencies of SP, NKA and NKB to produce NK1R internalization. We used spinal cord slices, a high-quality preparation developed for electrophysiology (Yoshimura & Jessell, 1990), that contains functional synapses between dorsal horn neurons and primary afferent terminals (Randic et al., 1993; Yoshimura & Nishi, 1996; Marvizon et al., 1997; Sandkuhler et al., 1997).

Methods

Chemicals

Isoflurane was from Halocarbon Laboratories, River Edge, NJ, U.S.A. Normal goat serum (NGS) was from Jackson ImmunoResearch Laboratories Inc., West Grove, PA, U.S.A. SP was from Tocris (Ellisville, MO, U.S.A.). Other chemicals were purchased from Sigma/RBI (St Louis, MO, U.S.A.).

Media for slices

Artificial cerebrospinal fluid (aCSF) contained (in mM) 124 NaCl, 1.9 KCl, 26 NaHCO₃, 1.2 KH₂PO₄, 1.3 MgSO₄, 2.4 CaCl₂ and 10 glucose. K⁺-aCSF contained 5 mM KCl. Sucrose-aCSF was aCSF in which NaCl was iso-osmotically replaced by sucrose (215 mM) and containing 5 mM KCl. These media were constantly bubbled with 95% O₂/5% CO₂.

Preparation of spinal cord slices

All animal procedures were approved by the Chancellor's Animal Research Committee at UCLA, and conform to NIH guidelines. Efforts were made to minimize the number of animals used and their suffering. Slices were prepared as previously described (Randic et al., 1993; Marvizon et al., 1997; 1999a, 1999b; Sandkuhler et al., 1997). Briefly, Sprague–Dawley rats 3–4-weeks old (Harlan, Indianapolis, IND) were anesthetized with isoflurane. A lumbar segment of the spinal cord was rapidly extracted, placed in ice-cold sucrose-aCSF, and cleaned of dura mater and ventral roots. The spinal cord segment was glued vertically to a block of agar in the stage of a Vibratome (Technical Products International, St Louis, MO, U.S.A.), and coronal spinal cord slices (400 μm) were cut in ice-cold sucrose-aCSF, using minimum forward speed and maximum vibration amplitude. Up to six slices were obtained from each animal in the L1–L4 region. To prepare slices with dorsal roots, the selected roots were lightly pulled to separate them from the spinal cord. The blade of the Vibratome was aimed at the entrance of the root with the help of a stereo microscope. Fiber continuity between the root and the dorsal funiculus was assessed, and only slices with more than 80% of the dorsal funiculus continuous with the root were used. After cutting, slices were kept for 1 h in K⁺-aCSF at 35°C, and then transferred to aCSF.

Slice treatment

For drug application, slices were placed in a nylon net suspended halfway inside a small beaker covered with aluminum foil, and incubated for 10 min at 35°C in aCSF (5–20 ml) constantly bubbled with 95% O₂/5% CO₂. In some experiments, 10 μM thiorphan and 10 μM captopril were added to the aCSF. The incubation was ended by placing the slices in cold fixative. Slices were used and fixed within 3 h of preparing them.

Electrical stimulation of the dorsal root

One dorsal root contiguous with the slice was stimulated using previously established procedures (Malcangio et al., 1997; Marvizon et al., 1997; 1999a; Lever & Malcangio, 2002) with some modifications. Slices were placed in a custom-made slice chamber (a gift from Dr Marzia Malcangio, King's College of London, London, England) and superfused with aCSF at 35°C bubbled with 95% O₂/5% CO₂ at a rate of 4 ml min⁻¹, starting 5 min before the stimulation. In some experiments, aCSF contained 10 μM thiorphan and 10 μM captopril. The electrode consisted of two platinum wires (diameter 0.5 mm, separation 1 mm) located in a side compartment separated from the superfusion chamber by a movable partition. The dorsal root was drawn into the electrode compartment through a hole in the partition (sealed with vacuum grease) and placed on top of the platinum wires. The electrode compartment was then emptied of aCSF and filled with mineral oil. Contact between the root and the electrode wires was monitored with a stereo microscope. Electrical stimulation was generated by a Master-8 stimulator and an Iso-Flex stimulus isolating unit (A.M.P. Instruments, Jerusalem, Israel), and consisted of square pulses of 20 V and 0.4 ms. Slices were kept in the chamber for 10 min after stimulation and then fixed. The side of the slice ipsilateral to the stimulated root was marked by punching a round hole in the ventral horn.

Immunohistochemistry in spinal cord slices

Histological sections from spinal cord slices were prepared and labeled as described (Marvizon et al., 1997; 1999a; 2003). Briefly, the slices were fixed, cryoprotected, frozen on dry ice, and sectioned with a cryostat at 25 μm in the coronal plane. Free-floating sections were incubated overnight at room temperature with a 1 : 3000 dilution of a rabbit antiserum recognizing the NK1 receptor (#94168, a gift from Dr Nigel Bunnett, UCSF), previously characterized (Grady et al., 1996). The secondary antibody was Alexa-488 goat anti-rabbit IgG (Molecular Probes, Eugene, OR, U.S.A.), diluted 1 : 2000. Incubation with the secondary antibody was 2 h at room temperature.

Confocal microscopy and image processing

Confocal images were acquired at UCLA's Carol Moss Spivak Cell Imaging Facility, using a Leica TCS-SP scanning confocal microscope and an argon (488 nm) laser. The pinhole was 1.0 Airy units, and the objective × 100, with a numerical aperture of 1.4, resulting in an optical section thickness (full-width at half-maximum) of 0.62 μm. For each image, stacks of optical sections were obtained at intervals of 0.57 μm, averaging each optical section up to six times to reduce the noise. One image was obtained with a × 40 objective (0.75 numerical aperture) with an optical section thickness of 2.16 μm and optical section separation of 1.79 μm. Images were processed in Adobe Photoshop 5.5., using the ‘curves' feature of the program to adjust the contrast. Images were acquired at a digital size of 1024 × 1024 pixels, and were later cropped to the relevant part of the field without altering the original image resolution.

Quantification of NK1R internalization

The amount of NK1R internalization was quantified using a standard method (Mantyh et al., 1995; Abbadie et al., 1997; Honore et al., 1999; Schwei et al., 1999; Trafton et al., 1999; 2001) with minor modifications (Marvizon et al., 1997; 1999a). Briefly, we determined the percentage of NK1R immunoreactive neurons in lamina I that show internalization in relation to the total number of NK1R neurons sampled. The person counting the neurons was unaware of the treatment. A Zeiss Axiovert 135 (Carl Zeiss, Inc., Thornwood, NY, U.S.A.) fluorescence microscope fitted with a × 100 objective was used to count neurons. Neuronal somas with 10 or more endosomes were considered to have internalized receptors. At least three sections per slice were used, and all NK1R cells in lamina I in each section were counted.

Data analysis

Treatments were randomized between slices, and no more than two slices from the same animal received the same treatment. Data were analyzed using Prism 4.0 for Windows (GraphPad Software, San Diego, CA, U.S.A., www.graphpad.com). In concentration–response experiments, a sigmoidal dose–response function,

was fitted to the data points by nonlinear regression. The ‘top' and ‘bottom' parameters were constrained to values <100% and >0%, respectively. The statistical error associated with the calculated EC₅₀ (effective concentration for 50% of the maximum response) was expressed as 95% confidence interval (95% CI). To determine whether there were significant differences between EC₅₀ values, the data of two concentration–responses (for example, presence vs absence of peptidase inhibitors, or SP vs NKA) were fitted simultaneously to this function (‘global fitting'). Akaike's Information Criterion (AIC) (Motulsky & Christopoulos, 2003) was used to calculate the probability (‘Akaike's weight') that the log EC₅₀ was the same or different for the two concentration–responses. The ‘evidence ratio' is defined as the ratio between the probabilities of the two models (i.e., ‘EC₅₀ is the same' and ‘EC₅₀ is different'), and roughly indicates how many times one model is more likely to be correct than the other. For other statistical analyses, we used one-way or two-way ANOVA with Bonferroni's post-test; statistical significance was set at 0.05.

Results

In untreated spinal cord slices, the NK1R was located at the cell surface in the majority of dorsal horn neurons (Marvizon et al., 1997; 1999a; 2003). For example, Figure 1a shows several lamina I neurons with NK1Rs in their surface. The localization of NK1R at the cell surface can be appreciated in more detail in Figure 1b, showing a high-magnification confocal image of a large lamina I neuron. In contrast, incubating spinal cord slices with SP produced the internalization of NK1Rs in neurons throughout the dorsal horn (Marvizon et al., 1997). For example, Figure 1d shows a lamina I neuron with NK1R internalization in a slice incubated in 100 nM SP. The number of lamina I neurons with NK1R internalization increased with the concentration of SP, following a sigmoidal concentration–response curve (Figure 2a).

Figure 1.

Confocal images of NK1R neurons after incubation with neurokinins. The peptidase inhibitors (PI) were 10 μM thiorphan and captopril. The dorsal side is up for all panels. (a) Slice incubated in aCSF (control); this low-magnification image (× 40, scale bar is 50 μm, two optical sections separated by 1.79 μm) shows six NK1R neurons in central lamina I without NK1R internalization. Images in the rest of the panels were taken at × 100 (scale bar is 20 μm), and consist of two optical sections (three for b, c, h, four for g) separated 0.57 μm. (b) Slice incubated in aCSF (control); large lamina I neuron showing no NK1R internalization. (c) Slice incubated with peptidase inhibitors; these four neurons in lamina I show no NK1R internalization. (d) Slice incubated with 100 nM SP and no peptidase inhibitors; lamina I neuron showing NK1R internalization. (e) Slice incubated with 100 nM SP and peptidase inhibitors; two lamina I neurons show NK1R internalization. (f) Slice incubated with 100 nM SP and peptidase inhibitors; lamina IV neuron showing NK1R internalization. (g) Slice incubated with 100 nM NKA and no peptidase inhibitors; lamina I neuron showing NK1R internalization. (h) Slice incubated with 100 nM NKA and peptidase inhibitors; of these five neurons in lamina I, four show NK1R internalization. (i) Slice incubated with 100 nM NKA and peptidase inhibitors; lamina IV neuron showing NK1R internalization. (j) Slice incubated with 1 μM NKB and peptidase inhibitors; four lamina I neurons show NK1R internalization.

Figure 2.

Effect of peptidase inhibitors on the potencies of SP and NKA to produce NK1R internalization. Slices were incubated for 10 min at 35°C with different concentrations of SP (a) or NKA (b) in the absence (control) or presence of the peptidase inhibitors thiorphan and captopril (10 μM). Points are the mean±s.e.m. of 3–4 slices. Curves were obtained by fitting a sigmoidal dose–response function to the data, and the calculated EC₅₀ values are given next to the corresponding curve.

It is possible, however, that the ability of SP and NKA to bind to the NK1R to produce its internalization is restricted by their degradation by peptidases present in the spinal cord. To investigate to what extent this occurs, we incubated spinal cord slices for 10 min with several concentrations of SP or NKA in the presence and absence of peptidase inhibitors, and measured the amount of NK1R internalization in lamina I neurons (Figure 2). The peptidase inhibitors used were thiorphan, an NEN inhibitor, and captopril, a DCP inhibitor. Inhibitors of NEN and DCP have been found to increase the release of SP (Duggan et al., 1992), and are routinely used to protect released SP against degradation (Malcangio & Bowery, 1993; 1994; Malcangio et al., 1997; Lever & Malcangio, 2002). Thiorphan and captopril were both used at 10 μM, a concentration at which they completely protect opioid peptides against degradation (Suzuki et al., 1997; Hiranuma et al., 1998; Song & Marvizon, 2003) and increase responses to SP and NKA in the spinal cord (Suzuki et al., 1994). Although some of the SP or the NKA may have reached the peptidases before they were blocked by the inhibitors, this would not have affected our results, given the presence of a large supply of neurokinins outside the slices and more than enough time (10 min) for NK1R internalization to occur: it is completed in about 2 min (Wang & Marvizon, 2002). The peptidase inhibitors by themselves did not produce NK1R internalization, as can be observed in the cluster of neurons in lamina I shown in Figure 1c, and in measurements in the absence of SP (‘0 M') in Figure 2a.

The peptidase inhibitors produced a left shift of the concentration–response curves of both SP and NKA (Figure 2). Thus, the EC₅₀ for SP decreased from a value of 32 nM (95% CI 19–55 nM) in the absence of peptidase inhibitors to 9 nM (95% CI 5–18 nM) in their presence. Similarly, the EC₅₀ for NKA decreased from 170 nM (95% CI 118–246 nM) in the absence of peptidase inhibitors to 60 nM (95% CI 29–125 nM) in their presence. To determine whether the differences between these EC₅₀ values were statistically significant, we performed a ‘global fitting' of the data in each panel to a dose–response function (see ‘Data analysis') and used the AIC to calculate the probability that the log EC₅₀ were different in the absence and presence of peptidase inhibitors. The probabilities (Akaike's weights) that the log EC₅₀ were different were 99% in the case of SP (evidence ratio of 97.3), and 80% in the case of NKA (evidence ratio of 4.1), showing that the peptidase inhibitors did increase the potencies of SP and NKA to produce NK1R internalization. This, in turn, indicates that a certain amount of the SP and NKA entering the slices was degraded by peptidases when peptidase inhibitors were not present.

Since a previous study (Trafton et al., 2001) reported that the potencies of SP and NKA to produce NK1R internalization were the same, we also used the AIC to compare the log EC₅₀ of SP and NKA. The probabilities that their log EC₅₀ were different were >99.99% in the absence of peptidase inhibitors, and 99.4% (evidence ratio of 154) in their presence. Therefore, SP was clearly more potent (five to seven times) than NKA to produce NK1R internalization in our experiments.

Images of NK1R neurons in the slices used for this study are shown in Figure 1. These images were taken from slices treated with 100 nM concentrations of SP or NKA, which produced submaximal amounts NK1R internalization (Figure 2). They show representative lamina I neurons with NK1R internalization after incubating the slices with SP alone (Figure 1d), SP and peptidase inhibitors (Figure 1e), NKA alone (Figure 1g) and NKA and peptidase inhibitors (Figure 1h). Although the quantitative analysis in Figure 2 was limited to lamina I neurons, SP and NKA also produced NK1R internalization in neurons in laminae III and IV. For example, lamina IV neurons with internalization after incubation with SP and peptidase inhibitors, or NKA and peptidase inhibitors, are shown in Figure 1f and i, respectively.

To determine whether neurokinin B (NKB) was able to produce NK1R internalization in dorsal horn neurons, and to compare its potency to those of SP and NKA, we incubated spinal cord slices with different concentrations of NKB (Figure 3). The experiment was done in the presence of 10 μM thiorphan and captopril to avoid any possible degradation of NKB by peptidases. Relatively high concentrations (10 μM) of NKB produced NK1R internalization in practically all NK1R neurons in lamina I. Thus, the ‘top' parameter of the concentration–response curve for NKB was 100±5% of NK1R cells. The calculated EC₅₀ for NKB was 573 nM (95% CI 329–997 nM), indicating that its potency to produce NK1R internalization was 64 times and 10 times lower than the potencies of SP and NKA, respectively. The probability (Akaike's weights) that the log EC₅₀ of NKB was different from that of SP and NKA was >99.99%. Representative neurons showing NK1R internalization in slices incubated with 1 μM NKB are shown in Figure 1j. Since, in the absence of peptidase inhibitors, the potency of NKB may be still lower, NKB does not appear to contribute substantially to the activation of NK1R in lamina I neurons.

Figure 3.

Concentration–response of NKB to produce NK1R internalization. Slices were incubated for 10 min at 35°C with different concentrations of NKB in the presence of thiorphan and captopril (10 μM). Points are the mean±s.e.m. of three slices. The curve was obtained by fitting a dose–response function to the data, and the calculated EC₅₀ value is given next to the curve.

In addition to the DCP and NEN, aminopeptidases could also degrade neurokinins, restricting their ability to activate NK1R. There are several aminopeptidases in the nervous tissue that can potentially degrade neurokinins, including aminopeptidases A (EC 3.4.11.2.), N (EC 3.4.11.7.) and W (EC 3.4.11.16.). Therefore, we screened several aminopeptidase inhibitors for their ability to increase NK1R internalization produced by 10 nM SP (Figure 4). Aminopeptidase inhibitors were tested in the presence of 10 μM thiorphan and captopril in order to assess any possible effect of aminopeptidases in the absence of NEN and DCP activities, which may compete with aminopeptidases for the peptides. If the aminopeptidase inhibitors produce a decrease in the EC₅₀ for SP, this would be reflected in an increase in NK1R internalization elicited by 10 nM SP. Nevertheless, none of the aminopeptidase inhibitors tested increased NK1R internalization produced by SP (Figure 3). Actinonin inhibits aminopeptidase N; bestatin inhibits aminopeptidase W and leucine-aminopeptidase (EC 3.4.11.1) (Burley et al., 1991); amastatin inhibits aminopeptidases A, N and W (Tieku & Hooper, 1992; Suzuki et al., 1997) and puromycin and bacitracin inhibit other aminopeptidases. Our results indicate that none of these aminopeptidases were able to appreciably decrease the concentration of SP inside the slices.

Figure 4.

Lack of the effect of aminopeptidase inhibitors on SP-induced NK1R internalization. Slices were incubated for 10 min at 35°C with 10 nM SP, 10 μM thiorphan, 10 μM captopril and the indicated aminopeptidase inhibitors (10 μM). Bars are the mean± s.e.m. of three slices. ANOVA yielded no statistically significant differences (P>0.05).

To determine whether peptidases restrict the access of endogenously released neurokinins to NK1R in dorsal horn neurons, neurokinin release from primary afferent terminals was evoked by electrical stimulation of the dorsal root. Slices were superfused with 10 μM thiorphan and captopril starting 5 min before the stimulation, to allow enough time for the inhibitors to penetrate the slice and saturate the peptidases. One dorsal root entering the slice was placed on a bipolar electrode in pool of mineral oil, which served to direct the electrical current through the root. We used pulses of 20 V and 0.4 ms, an intensity and duration sufficient to recruit most C fibers in the root (Koslow et al., 1973; Li & Bak, 1976). Since high-frequency stimulation is more effective than low frequency stimulation to release SP (Go & Yaksh, 1987) and to produce NK1R internalization (Marvizon et al., 1997), we used two stimulation patterns: (1) low frequency, consisting of 1000 pulses at 1 Hz (Figure 5a); (2) high frequency, consisting of 300 pulses at 30 Hz (Figure 5b). Both stimulation patterns produced a significant increase in NK1R internalization when comparing the side of the slice ipsilateral to the stimulated root with the contralateral side (P<0.0001, two-way ANOVA). However, 30 Hz stimulation produced significantly more internalization than 1 Hz stimulation, despite it being delivered with a smaller number of pulses, both in the absence (P<0.001, two-way ANOVA and Bonferroni's post-test) and in the presence (P<0.05) of peptidase inhibitors. Internalization in the contralateral side was similar to control (untreated) slices. Superfusing the slices with thiorphan and captopril significantly increased (P<0.01) the amount of NK1R internalization produced by 1 Hz stimulation (Figure 5a), but did not modify the internalization produced by 30 Hz stimulation.

Figure 5.

Effect of peptidase inhibitors on NK1R internalization produced by endogenously released peptides. Slices with one dorsal root attached were superfused with aCSF alone (−PI) or aCSF containing 10 μM thiorphan and captopril (+PI), starting 5 min before electrical stimulation and until they were fixed 10 min after stimulation. The dorsal root was stimulated with pulses of 20 V, 0.4 ms, delivered in a single train of 1000 pulses at 1 Hz (a) or 300 pulses at 30 Hz (b). Bars are the mean±s.e.m. of three slices. The percentage of lamina I NK1R neurons showing internalization was measured in the dorsal horn ipsilateral (‘stimulated') or contralateral to the stimulated root. Statistical analyses consisted of two-way ANOVA with ‘side' (contralateral versus stimulated) and ‘peptidase inhibitors' (‘−PI' versus ‘+PI') as the variables. For 1 Hz stimulation, there was a significant effect of ‘side' (P<0.0001) and ‘peptidase inhibitors' (P=0.01); Bonferroni's post-test revealed significant differences (**P<0.01) between ‘−PI' and ‘+PI'. For 30 Hz stimulation, there was a significant effect of ‘side' (P<0.0001), but not of ‘peptidase inhibitors'.

Figure 6 shows representative NK1R neurons in lamina I ipsilateral to the stimulated root in each of these conditions. After 1 Hz stimulation, little internalization was observed in the soma of the neuron shown in Figure 6a, although some endosomes are found in its large dendrite. In the presence of peptidase inhibitors, 1 Hz stimulation produced substantial internalization in the neuron shown in Figure 6b. Stimulation at 30 Hz produced NK1R internalization both in the absence (Figure 6c) and in the presence (Figure 6d) of peptidase inhibitors.

Figure 6.

Confocal images of NK1R neurons after electrical stimulation of the dorsal root. Slices with one dorsal root attached were superfused with aCSF alone (a, c) or aCSF containing 10 μM thiorphan and captopril (peptidase inhibitors, PI) (b, d) 5 min before, during, and 10 min after stimulation. The dorsal root was stimulated with pulses of 20 V, 0.4 ms, delivered in a single train of 1000 pulses at 1 Hz (a, b) or 300 pulses at 30 Hz (c, d). Images were taken at × 100 (scale bar is 10 μm), and consist of three optical sections (four for (d)) separated 0.57 μm. The dorsal side is up. There is little NK1R internalization in the neuron in panel (a), but substantial internalization is observed in the other panels.

Discussion

Thiorphan, an inhibitor of NEN, combined with captopril, an inhibitor of DCP, moderately increased the potencies of SP and NKA to produce NK1R internalization. These findings are consistent with previous studies measuring SP release from the spinal cord. For example, Geppetti et al. (1989) found that thiorphan increased SP release from the guinea-pig spinal cord 2.4 times. Duggan et al. (1992) reported that spinal injections of the NEN inhibitor kelatorphan combined with the DCP inhibitor enalaprilat increased the amount of SP release evoked by primary afferent stimulation. However, relating SP release with NK1R activation is not straightforward: when we compared capsaicin-induced SP release and NK1R internalization in the same slices (Marvizon et al., 2003), we found that NK1R internalization could be elicited by amounts of released neurokinins too low to be detected by RIA, whereas amounts of neurokinins within the range of detection of RIA saturated NK1Rs. We did not compare the relative abilities of NEN and DCP to degrade SP and NKA, because the combined effect of thiorphan and captopril was already quite small. Geppetti et al. (1989) found that captopril did not affect capsaicin-evoked SP release from guinea-pig dorsal horn slices, whereas thiorphan produced a 2.4-fold increase in SP release. Similarly, Duggan et al. (1992) found that the addition of the DCP inhibitor enalaprilat did not increase SP release from the cat spinal cord observed in the presence of the NEN inhibitor kelatorphan. Therefore, NEN appears to be the main enzyme degrading SP.

Aminopeptidases are also thought to contribute to the degradation of neurokinins. In particular, the aminopeptidase inhibitor bacitracin was reported in one study (Mauborgne et al., 1991) as the most potent agent protecting SP from degradation, and has been used together with inhibitors of NEN and DCP to improve the recovery of released SP (Malcangio & Bowery, 1993; 1994; Chen et al., 1996; Malcangio et al., 1997; Lever & Malcangio, 2002). However, none of the aminopeptidase inhibitors that we tested, including bacitracin, increased NK1R internalization produced by a low concentration of SP. It is difficult to reconcile our results using bacitracin with those of Mauborgne et al. (1991), particularly in view of the fact that they found no effect for NEN and DCP inhibitors, in contrast with our findings and those of other investigators (Geppetti et al. 1989; Duggan et al., 1992). We cannot rule out that neurokinins are degraded by aminopeptidases that are not inhibited by the compounds that we used. For example, aminopeptidase P (EC 3.4.11.9) appears to degrade SP, but is not inhibited by amastatin, bestatin, puromycin or bacitracin (Harbeck & Mentlein, 1991).

We also compared the potencies of SP, NKA and NKB to produce NK1R internalization. NKA was 5–7 times less potent than SP, whereas NKB was 10 times less potent than NKA and 64 times less potent than SP. This agrees with the established sequence of affinities of NK1R for its endogenous ligands: SP>NKA>NKB (Routh & Helke, 1995). Our EC₅₀ values for SP (32 nM in the absence and 9 nM in the presence of peptidase inhibitors) were similar to those reported by Trafton et al. (2001) in cultured dorsal horn neurons (14 nM). However, we found substantially higher EC₅₀ values for NKA (170 nM in the absence and 60 nM in the presence of peptidase inhibitors) than those reported by Trafton et al. (27 nM). This disparity may derive from differences in NK1Rs between adult neurons and the embryonic neurons used for cell cultures. Although Trafton et al. also reported that SP and NKA had the same potency in adult rats, they injected these peptides in the sacral spinal cord and measured NK1R internalization in the lumbar spinal cord, introducing a diffusion factor that substantially increased the EC₅₀ values obtained: 171 μM for SP and 210 μM for NKA. Hence, the apparently similar potencies for SP and NKA may have been caused by a higher susceptibility of SP to degradation by peptidases, as reported in guinea-pig airways (Martling et al., 1987; Devillier et al., 1988). Our own results support this idea: the increases in potency produced by peptidase inhibitors were greater for SP (3.5 times) than for NKA (2.8 times).

We also investigated the ability of peptidases to restrict the activation of NK1Rs by endogenous neurokinins released from primary afferent terminals. Peptidase inhibitors were able to increase NK1R internalization produced by dorsal root stimulation at 1 Hz but, surprisingly, not at 30 Hz. One possible explanation for this finding is that 1 Hz stimulation evokes a more gradual release of neurokinins, leaving enough time for peptidases to degrade them before their concentrations can substantially build up in the extracellular space. In contrast, 30 Hz stimulation produces a much rapid release of neurokinins, which may lead to the saturation of peptidase activity. Alternatively, high-frequency stimulation may release compounds that inhibit the peptidases. We have previously reported (Marvizon et al., 1997; 1999a) that high-frequency stimulation of the dorsal root is more efficacious than low-frequency stimulation to produce NK1R internalization, and this agrees with the findings of Go & Yaksh (1987) measuring SP release. However, another group successfully used 1 Hz root stimulation to evoke SP release (Malcangio & Bowery, 1993; 1994; Malcangio et al., 1997). Interestingly, Malcangio et al. used peptidase inhibitors, whereas Go & Yaksh did not, and neither did we in our previous studies on NK1R internalization. Thus, our finding resolves this apparent discrepancy. However, note that 1000 pulses at 1 Hz produced less NK1R internalization than 300 pulses at 30 Hz, even in the presence of peptidase inhibitors (Figure 5). Therefore, the action of peptidases does not completely account for the frequency dependence of neurokinin release.

To gain an adequate perspective on the ability of peptidases to degrade neurokinins, it is important to compare it with their ability to degrade opioid peptides. In a previous study (Song & Marvizon, 2003), we demonstrated that inhibitors of NEN, DEC and aminopeptidase N increased the potencies of dynorphin A and Leu-enkephalin 10 and 100 times, respectively, to produce μ-opioid receptor internalization. In contrast, we show here that inhibitors of NEN and DEC produced a relatively smaller increase (2.8–3.5 times) in the potencies of SP and NKA to produce NK1R internalization. Moreover, whereas peptidase inhibitors are required to observe μ-opioid receptor internalization in dorsal horn neurons produced by endogenously released opioids (Song & Marvizon, 2003), endogenously released neurokinins produce abundant NK1R internalization in the absence of peptidase inhibitors (Mantyh et al., 1995; Allen et al., 1997; Liu et al., 1997; Marvizon et al., 1997; 1999a; Cao et al., 1998; Honore et al., 1999; Riley et al., 2001). Therefore, it is possible to use peptidase inhibitors to enhance the analgesic effect of endogenous opioids (Noble et al., 1992b; Roques, 2000) without increasing NK1R activation by endogenous neurokinins.

In summary, our results show that peptidases limit the ability of exogenously applied or endogenously released neurokinins to activate NK1Rs. Furthermore, they confirm the idea that all three neurokinins are able to activate NK1Rs and may exert physiological actions through this receptor (Maggi & Schwartz, 1997).

Abbreviations

AIC

Akaike's Information Criterion

ANOVA

analysis of variance

CI

confidence interval

DCP

dipeptidyl carboxypeptidase, EC 3.4.15.1

IR

immunoreactive

K⁺-aCSF

aCSF with 5 mM KCl

NEN

neutral endopeptidase, EC 3.4.24.11

NGS

normal goat serum

NK1R

neurokinin 1 receptor

NKA

neurokinin A

NKB

neurokinin B

SP

substance P

sucrose-aCSF

aCSF with 5 mM KCl and 215 mM sucrose instead of NaCl