The neuropeptide neuromedin U promotes inflammation by direct activation of mast cells

NMU-KO mice exhibit obesity and reduced nociceptive reflexes, which are reflective of the functions of NMU in the nervous system (8,14). Although NMU-R1 is expressed in a variety of tissues, including the gastrointestinal tracts and immune cells such as T cells, NK cells, and monocytes (9), the role of NMU in the periphery has not been clarified. Therefore, we first compared inflammatory responses using NMU-KO mice.

Intraplantar injection of CFA induced edema, neutrophil infiltration, and paw swelling in WT C57BL6/J mice (WT and KO; Fig. 1, A and B). NMU mRNA expression levels were extremely high in the paws of WT mice, being comparable or even higher than those in the intestine, which has been thought to be the most abundant source of NMU (Fig. 1 C). In contrast, paw edema and neutrophil infiltration did not occur in NMU-KO mice 3 h after intraplantar CFA injection (WT and KO; Fig. 1, A and B). In WT mice, injection of CFA resulted in a rapid induction of proinflammatory cytokines including TNF-α, IL-6, macrophage inflammatory protein 2, and leukocyte adhesion molecules, such as intracellular cell adhesion molecule 1, on the surface of endothelial cells (WT; Fig. 1 D). Induction of these acute inflammatory factors, however, did not occur in NMU-KO mice (KO; Fig. 1 D), indicating that NMU plays a critical role in the early inflammatory responses induced by CFA.

The early inflammation induced by CFA has been shown to be mast cell–dependent (15). Indeed, WBB6F₁-W/W^v mice in which mast cells are deficient did not exhibit any paw swelling, infiltration of neutrophils, or the induction of proinflammatory cytokines and intracellular cell adhesion molecule 1 by CFA injection, whereas control WBB6F₁-+/+ mice developed inflammation in a manner similar to WT mice (WBB6F₁-+/+ and –W/W^v; Fig. 1, A, B, and D). BM-derived mast cells (BMMCs) expressed relatively high levels of NMU-R1 mRNA, compared with BM or spleen cells (Fig. 1 C). Reconstitution of skin mast cells in WBB6F₁-W/W^v mice with BMMCs from -+/+ mice (W/W^v -+/+) by adoptive transfer restored the CFA-induced inflammatory responses (W/W^v -+/+; Fig. 1, A and B), confirming that the early phase of CFA-induced inflammation is dependent on mast cells.

The number and degranulation of skin mast cells were confirmed by toluidine blue (TB) staining. The total numbers of mast cells were not significantly different in the paws of WT, NMU-KO, and WBB6F₁-+/+ mice (Fig. S1 A), suggesting that the NMU deficiency did not alter mast cell development in NMU-KO mice. However, intense degranulation of mast cells was observed 3 h after CFA application in the paws of WT and WBB6F₁-+/+ mice but not in those of NMU-KO mice (Fig. S1 B). These data suggest a strong connection between NMU and the activation of mast cells in CFA-induced peripheral inflammation.

We also confirmed that CFA-induced mast cell degranulation and paw edema occurred in WBB6F₁-W/W^v mice reconstituted with either WT or NMU-KO BMMCs (Fig. S2). Thus, the lack of mast cell–mediated inflammation following CFA injection in NMU-KO mice results from the absence of NMU release, not from an intrinsic defect in mast cell effector function. These results suggest that NMU release from the environment after CFA treatment activates mast cells. To identify potential sources of the neuropeptide, we examined the localization of NMU in peripheral tissues by immunostaining. Specific recognition of NMU by our antibody was confirmed by a lack of significant reactivity in NMU-KO mice (Fig. 2).

In WT mice, intense NMU immunoreactivity was observed in the epithelial layer of the hindpaw, suggesting that keratinocytes may be a major source of this peptide. Intraplantar injection of CFA into the dermis induced a rapid reduction of NMU reactivity lasting up to 24 h. No destruction of the epithelial layers was observed. These data suggest that NMU was released rapidly into the subcutaneous region after CFA injection. Because mast cells are localized in close proximity to the epithelial surface, expression of NMU in epithelial tissues may ensure the immediate activation of mast cells against environmental insults such as infection, chemical damage, and mechanical and neurogenic stresses.

To investigate the induction of mast cell–dependent inflammatory responses by NMU in vivo, we injected NMU subcutaneously into the paws of mice. SP, a potent neurogenic inflammatory peptide, was used for comparison because it has also been shown to induce degranulation of mast cells (13). A single injection of NMU induced progressive vasodilation in WT, NMU-KO, and WBB6F₁-+/+ mice, whereas NMU-induced vasodilation occurred only marginally in WBB6F₁-W/W^v mice (Fig. 3 A). Measurement of plasma extravasation using Evans blue outflow revealed that NMU injection into WT and WBB6F₁-+/+ mice, but not into WBB6F₁-W/W^v mice, induced two- to threefold higher extravasation than injection of vehicle (Fig. 3 B). Subsequent paw edema was also observed in WT and WBB6F₁-+/+ mice, but not in WBB6F₁-W/W^v mice (Fig. 3 C). NMU induced these early inflammatory responses in W/W^v -+/+ mice, in which BMMCs from WBB6F₁-+/+ mice were transferred adoptively into the paw (W/W^v -+/+; Fig. 3). These data indicate that NMU directly induces mast cell–dependent responses, such as vasodilation, plasma extravasation, and paw edema. In contrast to responses induced by CFA injection, NMU induced responses in NMU-KO mice similar to those seen in WT mice (KO; Fig. 3). Thus, NMU-induced mast cell inflammation can occur normally in NMU-KO mice, indicating that CFA-induced mast cell–mediated inflammation in the paw is largely dependent on the release of NMU in response to CFA.

In agreement with previous reports, SP injection into hindpaw induced extravasation and paw swelling in the hindpaws of WT, NMU-KO, and WBB6F₁-+/+ mice, but not in WBB6F₁-W/W^v mice (Fig. 3 B b, and C, bottom). Interestingly, plasma extravasation and paw edema induced by SP were less intense than those after NMU injection. Furthermore, the extravasation induced by SP was reduced in NMU-KO mice from the levels observed in WT mice (Fig. 3 B b, bottom), suggesting that SP-induced mast cell–dependent extravasation may be dependent at least partially on NMU. Further studies will be necessary to define the functional relationship between NMU and SP.

We then examined whether NMU induces degranulation of tissue mast cells in vivo (Fig. 4). NMU injection induced a high level of degranulation of mast cells in WT and WBB6F₁-+/+ mice as well as in NMU-KO mice (Fig. 4, A and C). The number of paw mast cells was not significantly different among WT, NMU-KO, and WBB6F₁-+/+ mice irrespective of NMU or SP injection (Fig. 4 B). Moreover, like WBB6F₁-+/+ mice, W/W^v -+/+ (Fig. 4 A) and W/W^v mice reconstituted with WT BMMCs (W/W^v-WT; Fig. S3) or with NMU-KO BMMCs (W/W^v-KO; Fig. S3) showed NMU-induced degranulation of mast cells and subsequent mast cell–dependent inflammatory responses (Fig. S3 A).

These data indicate that NMU can activate mast cells directly, leading to mast cell–dependent inflammation such as vasodilation and edema.

SP injection into hindpaw also induced mast cell degranulation in WT, NMU-KO, and WBB6F₁-+/+ mice (SP; Fig. 4, A and C). However, SP induced mast cell degranulation at lower levels than NMU, which may account for lower extravasation and paw swelling induced by SP than by NMU (SP; Fig. 3, B b and C).

Next, we examined the effects of NMU on mast cell function using peritoneal mast cells (PMCs) derived from WT mice. First, we confirmed the expression of NMU-R1 in PMCs (Fig. 5 A). As shown in Fig. 5, B and C, NMU induced Ca²⁺ mobilization and degranulation of PMCs in a dose-dependent manner. Ca²⁺ mobilization and degranulation of PMCs were observed at more than 10⁻⁴ M and 10⁻⁵ M NMU, respectively. The levels of NMU-induced Ca²⁺ mobilization and degranulation of PMCs were almost comparable with those induced by IgE receptor cross-linking. The low levels of IgE-dependent degranulation of PMCs were probably observed because most IgE receptors on PMCs were already occupied with endogenous IgE, as previously described (16); therefore, we speculate that only a small percentage of IgE receptors are available to bind exogenous DNP-specific IgE.

In this study, we showed the expression of NMU-R1 in mast cells and demonstrated that NMU induces an early inflammatory response by activating mast cells. Furthermore, the subsequent infiltration of neutrophils and expression of proinflammatory cytokines were completely suppressed in NMU-KO mice. These data strongly support the idea that NMU is the major mast cell–activating proinflammatory mediator produced in the periphery, especially in the skin. It is interesting that a bioactive NMU peptide (NMU-23) is secreted from frog skin (17); this observation suggests that NMU is a highly conserved molecule functioning in host defense against microbes, external mechanical stresses, and injury.

The high expression of NMU in the dermal layer of the skin contrasts with the expression patterns of other neuropeptides, such as SP and neuropeptide Y, which are expressed primarily in neural cells and have been shown to be involved in neurogenic inflammation. Although NMU is involved in nociceptive transmission and appetite in the nerve system (2, 6–8), our study demonstrated that NMU is also involved in peripheral nonneurogenic inflammation that activates mast cells. Thus, we speculate that NMU functions as a link between neurogenic inflammatory stimuli and mast cell–mediated inflammation. In addition, the effect of SP on neurogenic inflammation may require the activation of NMU in the axonal reflex of the sensory pathway. In either case, the blockade of NMU-R1 activation by a specific antagonist in the periphery may be a novel strategy to control the severity of inflammation in conditions such as hyperalgesia.

NMU-KO mice were generated by gene targeting as described previously (14). NMU-KO mice were backcrossed more than 10 times into a C57BL6/J background and compared with WT C57BL6/J mice purchased from Japan CLEA. WBB6F₁-+/+ and W/W^v mice were purchased from Japan SLC, Inc. All animal procedures were performed in accordance with the Japanese Physiological Society's guidelines for animal care.

Skin inflammation was produced by intraplantar injection of 10 μg CFA (in 20 μl mineral oil, 50%; Sigma-Aldrich). After measuring the CFA-induced paw edema with a micrometer, the hindpaws were removed, immersed in 4% paraformaldehyde (PFA) solution, and stained with hematoxylin and eosin (H&E). The Evans blue plasma extravasation assay was performed as described (11). In brief, after anesthesia of the mice, Evans blue (50 mg/kg) was injected intravenously into the jugular vein. 50 pmol of SP or NMU (NMU-23; Peptide Institute, Inc.) was injected into one hindpaw of the mouse, and the other paw was injected with vehicle. After 30 min, the paw was removed, weighed, and incubated in formamide for 24 h at 56°C. Extracted Evans blue was measured spectrophotometrically at 600 nm. For the assessment of vasodilation and degranulation of mast cells, hindpaws were removed 15 min after injection, fixed in 4% PFA, and stained with either H&E or TB.

The number of skin mast cells present at reaction sites was counted and expressed as the number of cells/mm² of dermis. Mast cells were classified into three categories: extensively granulated (50% of cytoplasmic granules exhibiting fusion), slightly to moderately degranulated (10–15% of granules exhibiting fusion or discharge), or normal as previously described (18).

For RT-PCR, total RNA was extracted using TRIzol reagent (Invitrogen). The synthesis of first-strand cDNA was performed using SuperScript II (Invitrogen) according to the manufacturer's instructions. The sequences of primer used are listed in Table S1.

NMU was detected as described (7). Briefly, hindpaws of mice were immersed with 4% PFA overnight at 4°C and paraffin sectioned to 5 μm in thickness. Sections were incubated with rabbit anti–rat NMU serum (donated by K. Mori, National Cardiovascular Center Research Institute, Osaka, Japan) overnight at 4°C, then with Alexa Fluor 488 (Molecular Probes, Inc.) diluted 1:200. Fluorescent images were obtained using a confocal laser-scanning microscope (LSM 5 Pascal; Carl Zeiss MicroImaging, Inc.).

BMMCs were maintained in RPMI 1640 supplemented with 5 ng/ml murine IL-3 (PeproTech), 8% FCS, 1% nonessential amino acids, 100 IU/ml penicillin, 100 μg/ml streptomycin, and 10 μM 2-mercaptoethanol for 4–6 wk as described (19). For the reconstitution of skin mast cells in WBB6F1-W/W^v mice, 10⁶ BMMCs in 20 μl of normal saline were injected intradermally into the paws. 4 wk after injection of BMMCs, mice were used for experiments.

Peritoneal mast cells were isolated from Sprague-Dawley rats (250–300 g) as previously described (20). For IgE-dependent activation of PMCs, cells were sensitized with various concentrations of DNP-specific IgE mAb (Sigma-Aldrich) for 1 h at 37°C before stimulation with 1 μg/ml DNP-BSA. Ca²⁺ mobilization assay was performed as described (21) using Fluo-4-AM (Invitrogen) with FLEXstation (Molecular Devices Corporation). Degranulation assay was performed as described (13).

Values are presented as mean ± SE. Differences among groups were analyzed using one-way analysis of variance with Scheffe's post hoc test. p-values <0.05 were considered to be significant.

Table S1 shows the specific primers used for RT-PCR. Fig. S1 shows the number of skin mast cells and the proportion of degranulation induced by CFA in WT, NMU-KO, and WBB6F1-+/+ mice. Fig. S2 shows the CFA-induced inflammation and degranulated mast cells in the paws of WBB6F₁-W/W^v mice reconstituted with BMMCs from WT and NMU-KO mice. Fig. S3 shows NMU-induced vasodilation and degranulated mast cells in the paws of WBB6F₁-W/W^v mice reconstituted with BMMCs from WT and NMU-KO mice.

This work was supported by the Program for the Promotion of Basic Research Activities for Innovative Biosciences and Grant-in-Aids for Scientific Research (B) from the Ministry of Education, Culture, Sports, Science and Technology of Japan (to M. Kojima) and MEXT Open Research Center Project (2004).

The authors have no conflicting financial interests.