Neurokinin-1 Receptor Antagonists Inhibit the Recruitment of Opioid-containing Leukocytes and Impair Peripheral Antinociception

Animal protocols were approved by local authorities and are in accordance with the guidelines of the International Association for the Study of Pain.29Male Wistar rats weighing 180–240 g were injected with 150 μl CFA (Calbiochem, La Jolla, CA) into the right hind paw (intraplantar) during brief isoflurane anesthesia and developed an inflammation confined to the inoculated paw. Experiments were conducted at 0–48 h after inoculation of CFA. All further injections were also performed during brief isoflurane anesthesia.

Mechanical nociceptive thresholds were assessed using the paw pressure algesiometer (modified Randall-Selitto test; Ugo Basile, Comerio, Italy).30Rats were handled for 4 days before testing. On the day of testing, rats were held under paper wadding, and incremental pressure was applied via  a wedge-shaped, blunt piston onto the dorsal surface of the hind paw by an automated gauge. The pressure required to elicit paw withdrawal, the paw pressure threshold, was recorded (cutoff at 250 g). The average of three trials, separated by 10-s intervals, was calculated. The same measurement was performed on the contralateral paw in alternated sequence to preclude order effects. To examine endogenous mechanisms of antinociception, paw pressure thresholds were determined at baseline conditions and 1 min after swimming in 2°–4°C cold water for 1 min.8An increase in paw pressure threshold was interpreted as mechanical antinociception. All behavioral experiments were performed by an examiner blinded to the treatment protocol.

Implantation was performed under continuous isoflurane anesthesia as described.31In brief, a 150-mm polyethylene tube (PE 10; Portex, Hythe, United Kingdom) was inserted intrathecally for 25 mm in cervical direction through an incision at the L3–L4 level. During the recovery period of 2 days, animals showing neurologic damage (e.g. , paralysis of the hind limbs) were excluded from the study. To ensure intrathecal localization of the catheter, 10 μl lidocaine, 2% (Braun, Melsungen, Germany), followed by a 10-μl solvent flush was applied, and the animals were monitored for development and reversal of paralysis of both hind limbs.

All hematopoietic cells were stained by mouse anti-rat CD45 Cy5 phycoerythrin monoclonal antibody (clone OX-1, identifies the leukocyte common antigen, 4 μg/ml; BD Biosciences, Heidelberg, Germany).32PMNs were identified by mouse anti-rat RP-1 phycoerythrin monoclonal antibody (12 μg/ml). Macrophages were stained by mouse anti-rat CD68 fluorescein isothiocyanate monoclonal antibody (formerly called ED1; 2 μg/ml; Serotec, Oxford, United Kingdom). Opioid-containing cells were labeled by mouse 3E7 monoclonal antibody (recognizing the pan opioid sequence, 20 μg/ml; Gramsch Laboratories, Schwabhausen, Germany) followed by rabbit anti-mouse immunoglobulin (Ig)G^2a+bphycoerythrin antibody (15 μg/ml; BD Biosciences).5,7NK¹receptor was stained by rabbit anti-rat NK¹receptor serum (1:100; Chemicon, Hampshire, United Kingdom) at 4°C for 30 min and subsequently by goat anti-rabbit IgG- fluorescein isothiocyanate antitbody (15 μg/ml; Vector Laboratories, Burlingame, CA). Specificity of staining was controlled by isotype-matched control antibodies (mouse IgG^2a, 20 μg/ml, BD Biosciences; and rabbit IgG 4 μg/ml, Santa Cruz, Santa Cruz, CA).

Cell suspensions from paw tissue were prepared and stained as described.5,7,32Briefly, subcutaneous paw tissue was enzymatically digested and was pressed through a 70-μm nylon filter (BD Biosciences). Staining with anti-rat CD45 Cy5 phycoerythrin was done without permeabilization. For intracellular stains using 3E7, anti-rat–CD68 fluorescein isothiocyanate, and anti-rat RP-1 phycoerythrin, cells were fixed with 1% paraformaldehyde and permeabilized with saponin buffer. Permeabilized cells were incubated with the aforementioned primary and secondary antibodies. Replacement of the primary antibodies with isotype-matched irrelevant antibodies was used for negative controls.11 

To calculate absolute numbers of cells per paw, fluorescence-activated cell sorting events from fluorescent TruCOUNT beads and CD45⁺stained cells were collected simultaneously, and the number of CD45⁺cells per tube was calculated accordingly. For quantification, 70,000 fluorescence-activated cell sorting events were acquired. Data were analyzed using CellQuest software (all BD Biosciences).

All experiments were performed 0–48 h after intraplantar injection of CFA. Paw tissue was retrieved, the skin was removed, and subcutaneous tissue was cut into small pieces and processed as described.6Substance P, CXC chemokine ligand 1 (CXCL1; synonym: keratinocyte-derived chemokine), CCL2, and interleukin-1β (IL-1β) concentrations were measured by commercially available substance P, mouse CXCL1, rat CCL2, and rat IL-1β enzyme-linked immunosorbent assay kits according to manufacturers' instructions (Bachem/Peninsula, London, United Kingdom; Biosource International, Nivelles, Belgium; and R&D, Minneapolis, MN, respectively). To measure CXCL1, a rat CXCL1 peptide standard was used as described.6Optical density was measured by Spectra Max (Molecular Devices, Ismaning, Germany). Data were analyzed by the Softmax program (Molecular Devices).

The expression of ICAM-1 and NK¹receptor in inflamed paw tissue was analyzed in rats (n = 5/group) treated with NK¹receptor antagonists immediately before induction of inflammation and in control animals. Twenty-four hours later, rats were deeply anesthetized with halothane and perfused transcardially (0.1 m phosphate-buffered saline [PBS], followed by fixative solution: PBS containing 4% paraformaldehyde, pH 7.4). The skin with adjacent subcutaneous tissue was removed from both hind paws, postfixed for 30 min at 4°C in the fixative solution, and cryoprotected overnight at 4°C in PBS containing 10% sucrose. The tissue was embedded in Tissue Tek compound (OCT; Miles, Elkhart, IN), frozen, cut into 7-μm sections, mounted onto gelatin-coated slides, and processed for immunofluorescence.33The sections were incubated with mouse anti–ICAM-1 (1:200) alone or in combination with rabbit polyclonal anti–NK¹receptor antibody (1:500) and then with secondary antibodies. After incubation with primary antibodies, the tissue sections were washed with PBS and then incubated with the appropriate secondary antibodies: a Texas red conjugated goat anti-rabbit antibody in combination with fluorescein isothiocyanate conjugated donkey anti-mouse. The sections were then washed with PBS, mounted in Vectashield or Vectashield mounting media containing 4′,6 diamidino-2-phenylindole (DAPI) (both Vector Laboratories, Burlingame, CA), and viewed under Zeiss LSM510META confocal laser scanning system (Carl Zeiss Imaging, Thornwood, NY). To demonstrate specificity of staining, the following controls were included as mentioned in details elsewhere10by omission of either the primary antisera/antibodies or the secondary antibodies. Cell types were identified based on morphologic criteria. The number of ICAM and NK¹receptor expression in endothelial cells was quantified by an observer blinded to the experimental protocol using a Zeiss microscope (objective 20×).10The mean number of ICAM– or NK¹receptor–positive vessels per section was calculated from counting 3 sections per animal and 5 squares (384 mm²).

Rats were injected intraperitoneally every 12 h with 20 mg/kg L-733,060 (dissolved in sterile water and further diluted in up to 500 μl normal saline; Tocris, Bristol, United Kingdom) or 10 mg/kg SR140333 (dissolved in dimethyl sulfoxide and further diluted in up to 500 μl normal saline; Sanofi, Paris, France). Doses were chosen based on previous publications34,35and our pilot data. Control animals received solvent only (sterile water or dimethyl sulfoxide in normal saline, respectively). CFA was injected intraplantarly after the first injection, and experiments were performed 24–48 h later.

Separate groups of rats were injected intrathecally with 0.5 mg/kg L-733,060 (20 μl every 12 h) followed by 10 μl NaCl (0.9%) flushes.36Doses were chosen in pilot experiments and did not produce any toxic effects. Control animals received solvent only. Immediately after the first injection, paw inflammation was induced by CFA, and tissue was harvested 24 h later.

Data are presented as raw values or percentage control of baseline (mean ± SEM). Missing values are due to injection-related hematoma formation distorting fluorescence-activated cell sorting quantification of leukocyte subpopulations or due to accidental death by isoflurane overdose. Normally distributed data were analyzed by t  test. Otherwise, the Mann–Whitney rank sum test was used. Multiple measurements were analyzed by one-way analysis of variance for normally distributed data and by two-way repeated-measures analysis of variance when dependent and independent factors were present. Post hoc  comparisons were performed by the Student-Newman-Keuls and Holm-Sidak method, respectively. Differences were considered significant if P < 0.05.

Neurokinin-1 receptor blockade did not significantly change baseline paw pressure threshold in inflamed (fig. 1) or noninflamed paws (data not shown). At 24 h of inflammation, CWS-induced antinociception was significantly reduced by both intraperitoneal L-733,060 and intraperitoneal SR140333 treatments (figs. 1A and B), but not by intrathecal L-733,060 (fig. 1C). Furthermore, systemic treatment with L-733,060 did not alter baseline hyperalgesia at 48 h of inflammation, whereas it significantly reduced CWS-induced antinociception (table 1).

Cold water swim stress–induced antinociception is mediated by opioid-containing leukocytes at the site of inflammation.8,30In the inflamed paw, 213 ± 11 × 10³3E7⁺CD45⁺opioid-containing leukocytes per paw were present at 24 h after CFA, as quantified by flow cytometry. At 24 h of inflammation, intraperitoneal L-733,060 or intraperitoneal SR140333 administration significantly reduced the number of opioid-containing leukocytes in the paw by 41% and 44%, respectively (figs. 2A and B), whereas no effect was observed after intrathecal L-733,060 treatment (fig. 2C). Systemic treatment (L-733,060 intraperitoneally) also significantly decreased the number of opioid-containing leukocytes in the paw at 48 h after CFA (control: 100 ± 7.4% vs.  intraperitoneal L-733,060: 55 ± 7.2%; P < 0.05, t  test).

Levels of substance P were low in noninflamed paws (fig. 3A). At 24 and 48 h of inflammation, substance P content in the paw was significantly increased, with a 17.5-fold increase at 24 h. NK¹receptor expression was observed on endothelial cells as shown by colocalization with ICAM-1, peripheral neurons, and on CD45⁺leukocytes in the inflamed paw (figs. 3B–D).

Intercellular adhesion molecule 1 and NK¹receptor immunoreactivities on endothelial cells were quantified. Both were significantly up-regulated at 24 h of inflammation (fig. 4, representative examples; and table 2). Blockade of NK¹receptors using intraperitoneal SR140333 treatment did not alter ICAM-1 or NK¹receptor expression at 24 h (table 2). At 24 h inflammation, intraperitoneal SR140333 treatment did not alter the content of CXCL1 (PMN specific), CCL2 (monocyte specific), or IL-1β (table 3).

At 24 h of CFA, similar numbers of monocytes/macrophages and PMNs per paw were detected by flow cytometry (245 ± 18 × 10³CD68⁺macrophages per paw and 219 ± 16 × 10³RP1⁺PMNs per paw; fig. 5). Treatment with intraperitoneal L-733,060 or intraperitoneal SR140333 significantly reduced infiltrating PMNs and macrophages (CFA 24 h: figs. 5A, B, D, and E; CFA 48 h: table 4). CD3⁺lymphocytes did not change (data not shown). Intrathecal administration of L-733,060 did not influence overall leukocyte migration or recruitment of leukocyte subpopulation at 24 h after CFA (figs. 5C and F).

In this study, we demonstrated that NK¹receptor antagonists (1) impaired stress-induced peripheral opioid-mediated antinociception but did not alter baseline inflammatory hyperalgesia and (2) reduced migration of opioid-containing leukocytes without changing expression of ICAM-1 on endothelial cells or local chemokine/cytokine production. Systemic administration of blood-brain barrier–penetrating (SR140333) and nonpenetrating (L-733,060) NK¹receptor antagonists was equally effective, but intrathecal injection (L-733,060) was ineffective. Taken together, NK¹receptor antagonists seem to act peripherally by directly inhibiting the recruitment of opioid-containing leukocytes to the site of inflammation.

Baseline inflammatory hyperalgesia was not reduced by administration of NK¹receptor antagonists (figs. 1A and Band table 1). These findings are in line with previous animal studies15,37as well as studies in human pain patients.12Intrathecal application of NK¹receptor antagonists was also ineffective in our study (fig. 1C), but it has been reported to reduce hyperalgesia in other models.37–39In those models, antihyperalgesic effects were short lasting (maximum of 6 h after injection), whereas we measured nociceptive thresholds at 24–48 h of inflammation and 12 h after the last injection of the NK¹receptor antagonist. Long-lasting antihyperalgesic effects were only achieved by intrathecal injection of saporin-conjugated substance P that destroys NK¹receptor–expressing spinal neurons.40Inflammatory hyperalgesia is induced by pronociceptive mediators,2,41including cytokines (e.g. , IL-1β42) and chemokines (e.g. , CXCL143and CCL244,45). In our studies, NK¹receptor blockade did not change the local expression of representative hyperalgesic cytokines or chemokines (table 3), in line with the unaltered baseline nociceptive thresholds (fig. 1and table 1).

In contrast to this lack of change in baseline hyperalgesia, stress (CWS)–induced antinociception was reduced by NK¹receptor antagonists (figs. 1A and Band table 1). Our previous studies demonstrated that this form of antinociception is mediated by opioid peptide release from leukocytes at the site of inflammation and is fully blocked by opioid receptor antagonists in late stages of inflammation.8,30,46In the current study, we observed a comparable decrease in opioid-containing leukocytes in the inflamed paw using both blood-brain barrier–penetrating (L-733,060) and nonpenetrating (SR140333) NK¹receptor antagonists, whereas intrathecal injection (L-733,060) was ineffective. The intrathecally injected dose of L-733,060 was previously used in rats and was shown to effectively inhibit NR2B tyrosine phosphorylation in the spinal cord in the CFA model.36In addition to intrathecal injection of L-733,060, we also examined intrathecal injection of SR140333 (up to 6 μg/kg). This treatment did not reduce migration of opioid-containing leukocytes or stress-induced antinociception either, but we did not present the data because we observed neurotoxicity (e.g. , paralysis) at the highest doses. CWS-induced antinociception is mediated exclusively by locally secreted opioid peptides in subcutaneous inflammation.30,47In another paradigm, swim stress has been shown to involve IL-1 receptor activation,48but major differences exist between the two models: (1) presence versus  absence of inflammation, (2) species (rats vs.  mice), (3) type of swim stress (cold vs.  warm water), and (4) tests of nociception (Randall-Selitto vs.  hot plate). In particular, in our model CWS-induced antinociception cannot be blocked by an antibody against IL-1 or by an IL-1 receptor antagonist, indicating that this form of antinociception is not IL-1 dependent.49Taken together, our data show that NK¹receptor antagonist treatment correlates with impaired recruitment of opioid-containing leukocytes as well as a decrease in opioid-mediated antinociception. This seems to involve a peripheral site of action (i.e. , outside the central nervous system). Conceivably, peripheral NK¹receptors could be blocked on leukocytes, endothelial cells. or peripheral sensory neurons.

To further elucidate the mechanisms of reduced migration of opioid-containing leukocytes, we measured substance P content and NK¹receptor expression in the inflamed paw. Local substance P content increased during CFA-induced inflammation (fig. 3A). Potential sources of substance P are the peripheral sensory neurons and infiltrating leukocytes.13,50,51In addition to increased substance P amounts, we detected NK¹receptor expression on all infiltrating leukocytes, on neurons, and on endothelial cells (figs. 3B–D), confirming previous studies.16,17,20,52While neurogenic inflammation contributes to leukocyte infiltration, it is unclear whether blockade of NK¹receptors on peripheral sensory neurons influences leukocyte recruitment.53–55However, NK¹receptors have previously been shown to enhance leukocyte recruitment by increasing adhesion molecule expression on endothelial cells,17,23,24by augmenting chemokine production25or by directly inducing chemotaxis of leukocytes.19–22 

Adhesion molecules (i.e. , ICAM-1, selectins, and integrins) have previously been shown to mediate the recruitment of opioid-containing leukocytes, and their blockade can impair peripheral opioid-mediated antinociception during CFA-induced inflammation.9,11In the current study, expression of ICAM-1 on endothelial cells was increased at later stages of inflammation (24 h after CFA; table 2). Neither the CFA-induced up-regulation of ICAM-1 nor that of NK¹receptors on endothelial cells was altered by peripheral NK¹receptor blockade (fig. 4and table 2). Previous in vitro  studies have demonstrated that substance P increased the expression of ICAM-116,17as well as NK¹receptors.56,57In in vivo  models of inflammation, the effects of NK¹receptor antagonists vary between experimental models: They decreased NK¹receptor but not ICAM-1 expression in a mouse model of pancreatitis.58In contrast, they reduced ICAM-1 expression in experimental autoimmune encephalomyelitis.59In our model, NK¹receptor antagonists reduced the recruitment of opioid-containing leukocytes, but they did not alter the expression of ICAM-1 or NK¹receptors on the inflamed endothelium. We cannot fully exclude the possibility that NK1 receptor antagonists alter the expression of other adhesion molecules such as selectins or integrins and, thereby, impair peripheral antinociception.

As a second hypothesis, NK¹receptor antagonists might decrease local chemokine production and thereby impair selective leukocyte recruitment. Substance P has been shown to up-regulate chemokine production in leukocytes and epithelial cells in vitro .60,61In experimental pancreatitis, local chemokine expression was reduced by an NK¹receptor antagonist.62We examined a PMN- and a monocyte/macrophage-specific chemokine in our model (i.e. , CXCL163and CCL2,64respectively), and their expression was not altered by NK¹receptor antagonists (table 3). Multiple chemokines are responsible for PMN and monocyte/macrophage recruitment,65and alterations in other chemokines might occur after treatment with NK¹receptor antagonists. Although this possibility cannot be excluded, our previous study demonstrated that PMN-specific chemokines (i.e. , CXCR2 ligands) were regulated as a group and did not show relevant differences between individual chemokines.6Taken together, in our model NK¹receptor antagonists reduce the migration of opioid-containing leukocytes, but this effect does not seem to be mediated through alterations in adhesion molecule expression on endothelial cells or in local chemokine production.

Third, substance P binding to NK¹receptors on leukocytes might directly induce migration of opioid-containing leukocytes. This migration might be reduced by NK¹receptor blockade. Several studies demonstrated that NK¹receptor agonists induce chemotaxis of PMNs and monocytes/macrophages in vitro .19,66,67Previous studies have shown that PMNs and monocyte/macrophage migration to the site of inflammation is impaired after treatment with NK¹receptor antagonists or in NK¹receptor–deficient mice.20,24,68,69In line with these studies, the number of infiltrating macrophages and PMNs as well as of opioid-containing leukocytes was reduced in our model (figs. 2 and 5and table 4).

In conclusion, we have shown that NK¹receptor antagonists reduce the recruitment of opioid-containing leukocytes into the inflamed paw and thereby impair opioid-mediated antinociception in CFA inflammation. NK¹receptor agonists such as substance P seem to directly induce preferential recruitment of opioid-containing monocytes/macrophages while no indirect effects (i.e. , alterations in adhesion molecule expression or chemokine production) are observed. Most likely, NK¹receptor antagonists act directly on NK¹receptor–expressing, opioid-containing leukocytes, but the effects might also be mediated indirectly through NK¹receptors on peripheral neurons. Taken together, NK¹receptor agonists are important in the recruitment of opioid-containing leukocytes and the generation of peripheral opioid-mediated antinociception. The impairment of endogenous opioid-mediated peripheral analgesia might be an additional explanation for the lack of efficacy of NK¹receptor antagonist in human studies.12 

The authors thank Susanne Kotré and Katharina Hopp (both Technicians, Department of Anesthesiology, Charité, Berlin, Germany) for expert technical assistance.