Basophil tryptase mMCP-11 plays a crucial role in IgE-mediated, delayed-onset allergic inflammation in mice

Basophils are the rarest type of granulocyte and represent less than 1% of peripheral blood leukocytes. They were often considered erroneously a minor relative or blood-circulating precursor of tissue-resident mast cells, which they resemble.^1-3  It is now well accepted that basophils and mast cells are distinct cell lineages.⁴  Recent studies have identified crucial roles for basophils, distinct from those played by mast cells, in immune responses including allergy and protective immunity.^5-8  In response to various stimuli, basophils can readily secrete large quantities of the cytokine interleukin-4 (IL-4), which plays a versatile role in immunity through regulating other cells.⁹  In contrast to IL-4, functional roles of other molecules produced by basophils in immune responses remain ill defined.

Both basophils and mast cells contain cytoplasmic granules that are stained violet with basic aniline dyes, as described first by Paul Ehrlich more than 130 years ago. These granules store preformed materials such as histamine and proteases, and release them upon activation.^1,2  Ten serine proteases have been identified in murine mast cells, including 6 chymases (mouse mast cell protease-1 [mMCP-1], mMCP-2, mMCP-4, mMCP-5, mMCP-9, mMCP-10) and 4 tryptases (mMCP-6, mMCP-7, mMCP-11, mouse transmembrane tryptase).^10-14  mMCP-11 is the most recently discovered member of the mouse mast cell tryptase family and the most distantly related to other members in terms of the amino acid sequence.¹⁵  We previously demonstrated that mMCP-11 is preferentially expressed by basophils rather than mast cells, whereas mMCP-6 and mMCP-7 are expressed by mast cells but not basophils.¹⁶  None of the mast cell chymases are expressed by basophils, whereas the expression of granzyme B–like protease mMCP-8 is confined to basophils.^16,17  Thus, even though basophils and mast cells possess apparently similar basophilic granules, the repertoire of serine proteases stored in the granules is distinct between them, suggesting that mMCP-8 and mMCP-11 might be involved in nonredundant functions of basophils. We previously reported that the subcutaneous injection of recombinant mMCP-11 induces edematous skin swelling with increased microvascular permeability, suggesting that mMCP-11 may contribute to the development of basophil-mediated inflammatory responses.¹⁸ 

The nonredundant role of basophils in vivo has been clearly demonstrated in previous studies.^19-21  A single intradermal administration of allergens in the ear skin of IgE-sensitized mice induced 3 consecutive waves of ear swelling. The first 2 waves (early- and late-phase ear swelling on day 1) were mast cell–dependent, immediate allergic reactions. The third wave of ear swelling, which started on postchallenge day 2 and peaked on day 4, was accompanied by massive infiltration of eosinophils, neutrophils, monocytes, and macrophages in the ear skin lesion. We designated this delayed-onset response immunoglobulin E–mediated chronic allergic inflammation (IgE-CAI).¹⁹  Mast cells are dispensable for IgE-CAI, whereas basophil depletion prevents the development of IgE-CAI, even though basophils represented only 1% or 2% of infiltrates in the postchallenge skin lesion. Thus, basophils but not mast cells play an essential role in IgE-CAI.¹⁹  However, it remains to be determined how basophils elicit allergic inflammation in the IgE-CAI response.

In the present study, we generated a line of mMCP-11–deficient knockout (KO) mice to clarify the functional role of mMCP-11 in vivo. The KO mice showed an ameliorated IgE-CAI response, with reductions of swelling, vascular permeability, and leukocyte infiltration in the skin lesion, even though their basophil development remained unimpaired. The results of in vitro experiments suggested that mMCP-11–mediated proteolytic products of serum proteins, rather than mMCP-11 itself, trigger migration of leukocytes, including basophils, via 1 or more G protein–coupled receptors. Thus, the present study uncovered the crucial role of basophil tryptase mMCP-11 in allergic inflammation.

C57BL/6 mice (7-9 weeks old) were purchased from Japan SLC. To generate mMCP-11–deficient mice, the targeting vector for replacing the exons 1, 2, and 3 of the Mcpt11 (Prss34) gene with the loxP-PGK-gb2-neo-loxP cassette (Gene Bridges) was constructed by using Red/ET recombination (Gene Bridges) and transfected into C57BL/6 ES cells (UNITECH). The resultant knockout (KO) mice were further crossed with CAG-Cre mice (provided by J. Miyazaki, Osaka University)²²  to remove the loxP-flanked cassette from the Mcpt11 allele. All animal studies were approved by the Institutional Animal Care and Use Committee of Tokyo Medical and Dental University.

The following antibodies were purchased from BioLegend: biotinylated anti-CD49b; fluorescein isothiocyanate (FITC)–conjugated anti-CD49b, anti-F4/80; phycoerythrin (PE)–conjugated anti-FcεRIα, anti-CD45 (30-F11); allophycocyanin (APC)–conjugated anti-CD200R3, anti-Ly-6G; Pacific Blue–conjugated anti-CD117, anti-CD11b. PE-conjugated anti-Siglec-F (E50-2440) from BD Biosciences also was used. Anti-mMCP-11 and anti-mMCP-8 antibodies were obtained as reported previously.¹⁶  IgE (IGEL b4) specific to 2,4,6-trinitrophenol (TNP) and anti-CD16/anti-CD32 (2.4G2) were prepared in our laboratory. Recombinant mMCP-11 and its protease-dead mutant (in which Ala was substituted for Ser²⁰⁰) were prepared as described previously.¹⁸ 

Mouse bone marrow–derived basophils (BMBAs) were generated as described previously.¹⁶  Briefly, bone marrow cells were cultured in the presence of 300 pg/mL recombinant mouse IL-3 (BioLegend) for 7 days, followed by purification of CD49b⁺ cells by using the IMag cell separation system (BD Biosciences) with biotinylated anti-CD49b (monoclonal antibody DX5) and streptavidin-conjugated magnetic particles. The purity of c-kit^-CD49b⁺CD200R3⁺ BMBAs was more than 95% after magnetic purification.

IgE-CAI was induced as described previously.¹⁹  In brief, mice were sensitized with an intravenous injection of 300 μg anti-TNP IgE and challenged on the following day with an intradermal injection of 10 μg TNP-ovalbumin (OVA) into the right ear and OVA alone (the control) into the left ear. Ear thickness was measured at 24-hour intervals for 7 days, and thickness difference between the right and left ears was quantified (Δ = right ear − left ear). In some experiments, mice were treated with topical administration of 200 μg SC560, 200 μg meloxicam, 100 μg celecoxib, or vehicle (10 μL ethanol) alone once a day during the development of IgE-CAI. All 4 substances were administered topically. For Evans blue dye leakage analysis, mice were intravenously injected with 0.5% Evans blue dye in 100 μL phosphate-buffered saline (PBS) on day 3 of the IgE-CAI response. Two hours after the dye injection, ears were excised and incubated overnight in 0.7 mL formamide at 63°C to extract the dye that had leaked into the skin. The amount of dye in the extract was determined by a spectrophotometer set at 620 nm. Passive cutaneous anaphylaxis was induced as previously described with some modifications.²³  Briefly, mice were locally sensitized with intradermal injections of 100 ng anti-TNP IgE and control PBS in the right and left ear, respectively. On the following day, mice were challenged with an intravenous injection of 200 μg TNP-OVA in 200 μL of PBS containing 0.25% Evans blue dye. Thirty minutes later, the treated ears were excised, and the dye was extracted and measured.

Single cell suspensions were obtained from ear skins by treatment with 125 U/mL collagenase (Wako) at 37°C for 2 hours. After pretreatment with anti-CD16/32 (2.4G2 clone) and normal rat serum to avoid the nonspecific binding of irrelevant antibodies, cells were stained with the indicated antibody combination on ice for 30 minutes, and analyzed with a benchtop cytometer (FACSCanto, BD Biosciences). Cell lineages were defined as follows: basophils (c-kit⁻CD49b⁺CD200R3⁺), eosinophils (Siglec-F⁺SSC^high), neutrophils (Ly-6G⁺), monocyte-macrophages (F4/80⁺CD11b⁺ among cells in which both eosinophils and neutrophils were excluded). For histopathological examination, ear specimens were fixed and embedded in paraffin, and sections were stained with hematoxylin and eosin. For western blot analysis, BMBAs were lysed with a radioimmunoprecipitation assay buffer–containing protease inhibitor cocktail (Thermo Scientific), and total cell lysates were analyzed with sodium dodecyl sulfate–polyacrylamide gel electrophoresis, followed by immunoblot analysis with the indicated antibodies.

Total RNA was extracted from tissues or isolated cells by using the RNeasy Mini Kit (QIAGEN), followed by complementary DNA (cDNA) synthesis with reverse transcription by using oligo-dT and random primers. Quantitative polymerase chain reaction (PCR) assay of cDNA was performed by using the following primer sets: 5′-CTGGCTCCTGTTCCTCAGTCT-3′ and 5′-GCTGTGCTCCATGTCGTAGAG-3′ for Mcpt11 (Prss34); 5′-CACTGGTCAATGACATCATGCTCC-3′ and 5′-GTCAGACAAAGTGCAATTGGCCAAC-3′ for Mcpt8; 5′-AGGTCGGTGTGAACGGATTTG-3′ and 5′-TGTAGACCATGTAGTTGAGGTCA-3′ for Gapdh. Relative gene expression levels were calculated by using standard curves generated by serial dilutions of each cDNA standard and normalized by Gapdh expression levels.

BMBAs were sensitized with 1 μg/mL anti-TNP IgE overnight and then stimulated with 300 ng/mL (for degranulation assay) or 400 ng/mL (for measurement of cytokines) TNP-OVA or control OVA. Degranulation of BMBAs was determined by measurement of β-hexosaminidase release as previously described.²⁴  For measurement of cytokine production, BMBAs sensitized with IgE were stimulated with antigens for 24 hours. The concentrations of IL-4 and IL-6 in the culture supernatants were measured by enzyme-linked immunosorbent assay (ELISA MAX, BioLegend) and bead-based immunoassay (Cytometric Bead Array, BD Biosciences), respectively.

Leukocytes used for this assay were prepared as follows. Neutrophils were prepared from the bone marrow by using 62% Percoll gradient separation as described previously.²⁵  Eosinophils were isolated from the peritoneum of mice that had been treated for 7 days with daily intraperitoneal administration of IL-5.²⁶  Macrophages were isolated from the peritoneum of mice that had been treated with intraperitoneal administration of 1 mL of 4% thioglycolate broth 3 days before. The basophils used were BMBAs. The transwell apparatus (Kurabo) consisted of upper and lower chambers separated by a membrane with a 3-μm (for neutrophils) or 5-μm (for other cell types) pore size. Leukocytes (5 × 10⁵ cells) were placed into the upper chamber while 10 μg/mL recombinant mMCP-11, its protease-dead mutant, control bovine serum albumin (BSA), or a chemokine (100 ng/mL macrophage inflammatory protein 2 [MIP-2], 200 ng/mL eotaxin, or 200 ng/mL monocyte chemoattractant protein 1 [MCP-1]) was included in culture medium in the lower chamber. After 90 minutes (for neutrophils) or 2 hours (for other cell types) of incubation at 37°C, the number of cells that had migrated into the lower chamber was counted. In some experiments, cells were pretreated with 200 ng/mL (for BMBAs) or 100 ng/mL (for other cell types) of pertussis toxin (PTX, Sigma-Aldrich) for 1 hour before the transwell migration assay. mMCP-11 was treated with 10 μg/mL nafamostat (Sigma-Aldrich) or control dimethyl sulfoxide before and after 2-hour incubation of mMCP-11 in culture medium prior to the transwell migration assay. Recombinant tetramerized human β-tryptase was purchased from Promega. Bovine trypsin was from Sigma-Aldrich.

Statistical analysis was performed with an unpaired Student t test. A P value of less than .05 was considered statistically significant.

To explore the role of mMCP-11 in vivo, we established mMCP-11–deficient KO mice by targeting the exons 1, 2, and 3 of the mMCP-11 coding gene Prss34 (Figure 1A). The number of c-kit⁻CD49b⁺CD200R3⁺FcεRIα⁺ basophils detected in the bone marrow, spleen, and peripheral blood of KO mice was similar to that in wild-type (WT) mice (Figure 1B-C). The number of other hematopoietic cells, including mast cells, was also comparable between WT and KO mice (supplemental Figure 1, available on the Blood Web site). When cultured with IL-3, bone marrow cells isolated from WT and KO mice generated comparable numbers of c-kit⁻CD49b⁺CD200R3⁺ basophils (BMBAs) (supplemental Figure 2), suggesting unimpaired development of basophils in KO mice. The lack of mMCP-11 expression in basophils from KO mice was confirmed by both messenger RNA (mRNA) and protein levels (Figure 1D-E) while the expression of basophil-specific mMCP-8 remained intact in basophils from KO mice (Figure 1D-E). When stimulated with IgE plus antigens, BMBAs from KO mice degranulated (Figure 1F) and produced cytokines IL-4 and IL-6 (Figure 1G) as much as did BMBAs from WT mice. Moreover, WT and KO mouse BMBAs could migrate toward chemokine MIP-2 to a comparable extent (Figure 1H). Thus, the deletion of the Prss34 gene did not seem to have any adverse effect on basophil homeostasis or the FcεRI- or chemokine-mediated activation of basophils.

We previously reported that basophils crucially contribute to the development of IgE-CAI in the ear skin,¹⁹  and that basophil depletion abolishes the IgE-CAI response but not the mast cell–mediated immediate-type reaction.²⁰  However, it remained to be determined what molecules derived from basophils were responsible for the development of IgE-CAI. To examine the role of mMCP-11 in IgE-CAI, WT and KO mice were passively sensitized with haptenic TNP-specific IgE and either challenged with intradermal injection of the corresponding allergen TNP-OVA or given control OVA. In both KO and WT mouse strains, TNP-OVA induced ear swelling with a peak on day 4 postchallenge, but the extent of ear swelling in KO mice was approximately half that in WT mice (Figure 2A). Moreover, the increase of microvascular permeability in the skin lesion was significantly lower in KO mice (Figure 2B). The increased permeability in WT mice was suppressed by treatment with meloxicam (COX-2 inhibitor) but not SC560 (COX-1 inhibitor), whereas neither drug significantly affected the permeability in KO mice (Figure 2C), suggesting that mMCP-11 increases permeability in the skin lesion through COX-2. In accordance with this, IgE-CAI ear swelling in WT but not KO mice was significantly inhibited by treatment with a COX-2 inhibitor, either meloxicam or celecoxib (Figure 2D and supplemental Figure 3). The tryptase activity of mMCP-11 remained unaffected by these inhibitors (data not shown). Thus, basophil-derived mMCP-11 appeared to contribute to ear swelling in the IgE-CAI response in part by means of COX-2-mediated vascularhyperpermeability.

Histopathological examination of ear skin specimens indicated the reduction of cellular infiltration in the IgE-CAI skin lesion of KO mice compared with WT mice (Figure 3A). In accordance with this, flow cytometric analysis revealed that the number of total cells isolated from the skin lesion, mainly CD45⁺ hematopoietic cells, was reduced in KO mice to less than half that in WT mice (Figure 3B-C). Even though KO basophils by themselves showed unimpaired migration toward MIP-2 in vitro (Figure 1H), the number of basophils infiltrating the IgE-CAI skin lesion was significantly reduced in KO mice (Figure 3C). This was also the case for the numbers of eosinophils, neutrophils, and monocytes/macrophages in the skin lesion of KO mice (Figure 3C). Thus, mMCP-11 appeared to be involved in the infiltration and accumulation of these leukocytes in the IgE-CAI skin lesion.

We previously reported that 1 single cutaneous injection of 10 μg recombinant mMCP-11 elicited transient ear swelling with increased microvascular permeability, whereas no significant infiltration of leukocytes was detected in the injection site.¹⁸  We assumed that this difference between the IgE-CAI skin lesion and the mMCP-11–injected skin could be attributed to different duration of mMCP-11 action. In contrast to the single injection, mMCP-11 likely accumulates in the IgE-CAI skin lesion because basophils are continuously recruited there and activated by allergens to release mMCP-11. To mimic this situation in vivo, we treated WT mice with repeated intradermal injection of recombinant mMCP-11 in the ear skin. Three injections of mMCP-11 administered at 24-hour intervals elicited the infiltration and accumulation of neutrophils, eosinophils, basophils, and monocytes/macrophages (Figure 4) along with transient ear swelling induced by each single injection (supplemental Figure 4A). The severity of mMCP-11–induced ear swelling was not affected by depletion of basophils, demonstrating that exogenous mMCP-11 can induce this reaction in a basophil-independent manner (supplemental Figure 4A-B). It is important that mMCP-11 carrying a mutation in the active site of tryptase did not show such a leukocyte recruiting ability (Figure 4), indicating that the protease activity of mMCP-11 is essential for this function.

To explore mechanisms underlying the mMCP-11–mediated leukocyte recruitment, we used the transwell migration assay to examine the possibility that mMCP-11 can directly stimulate leukocyte migration. Basophils, eosinophils, neutrophils, and macrophages were individually placed in the upper chamber of the apparatus, while mMCP-11 was included in the lower chamber. Enzymatically active but not inactive mMCP-11 elicited transwell migration of eosinophils, basophils, and macrophages but not neutrophils (Figure 5A), demonstrating that the protease activity of mMCP-11 is essential for triggering their migration. In contrast, no significant migration of T or B cells was detected under these conditions (supplemental Figure 5). Moreover, the addition of meloxicam to the transwell culture produced no significant effect on mMCP-11–induced basophil migration (data not shown), suggesting that mMCP-11 elicited leukocyte migration in a COX-2–independent manner. Of note, the pretreatment of eosinophils, basophils, and macrophages with pertussis toxin almost completely abolished the mMCP-11–induced migration (Figure 5B). These results suggested that mMCP-11 functioned as a protease and triggered leukocyte migration through 1 or more G protein–coupled receptors.

To clarify how the protease activity of mMCP-11 contributes to leukocyte migration, we added a protease inhibitor, nafamostat, to the mMCP-11–containing culture medium at the different timing in the transwell migration assay (Figure 6A). As expected, the migration of basophils, eosinophils, and macrophages was strongly reduced to a level as low as 10% to 15% of the control level (mMCP-11 alone) when nafamostat was added simultaneously with mMCP-11 to the lower chamber (compare experiments 1 and 2 in Figure 6A). In contrast, when nafamostat was added to the lower chamber to inactivate mMCP-11 after the culture medium had been incubated with mMCP-11 at 37°C for 2 hours previous to the transwell migration assay, leukocyte migration was elicited at a level as much as 60% to 75% of the control standard (experiment 3 in Figure 6A). In addition, serum depletion from culture medium completely abolished the basophil migration induced by mMCP-11 but not that induced by MIP-2 (Figure 6B). These results strongly suggested that mMCP-11 acted on a serum protein or proteins present in culture medium, and its proteolytic product rather than mMCP-11 itself promoted leukocyte migration. The lesser extent of migration in experiment 3 compared with experiment 1 could be attributed to the degradation of proteolytic products and lack of new products during the transwell migration assay due to the nafamostat-mediated inactivation of mMCP-11 in experiment 3. Importantly, the leukocyte migration in experiment 3 was abolished by the pretreatment of leukocytes with pertussis toxin (Figure 6C), implying that proteolytic products generated by mMCP-11 induced leukocyte migration via a G protein–coupled receptor.

Previous study illustrated that basophils and mast cells display distinct repertoires of serine proteases stored in secretory granules.¹⁶  Among the tryptase members of the mMCP family, mMCP-11 is preferentially expressed by basophils instead of mast cells, whereas mMCP-6 and mMCP-7 are expressed only by mast cells in mice.¹⁶  Extensive investigation of mMCP-6 and mMCP-7 demonstrated that they play important roles in protective immunity against bacterial and parasitic infections^27,28  and also in the development of inflammatory diseases such as autoimmune arthritis^29,30  and colitis.³¹  In contrast, the functional significance of basophil tryptase mMCP-11 remained enigmatic. We showed in the present study that mMCP-11 is a crucial effector molecule in the development of IgE-CAI. This is the first demonstration that the basophil-derived protease plays a significant role in vivo.

mMCP-11–deficient KO mice showed much less severe IgE-CAI than did WT mice, characterized by milder ear swelling with reduced microvascular hyperpermeability and reduced leukocyte infiltration. We previously reported that basophil ablation during the progress of IgE-CAI resulted in drastic reduction of eosinophils and neutrophils infiltrating the skin lesion concomitantly with basophil elimination, suggesting the possible contribution of basophils to the eosinophil and neutrophil infiltration.^20,21  The present study clearly demonstrated that basophil-derived mMCP-11 induces the infiltration of leukocytes, including basophils, eosinophils, neutrophils, and macrophages. Basophil infiltration in the IgE-CAI skin lesion was observed even in KO mice, albeit to a lesser extent than in WT mice. Therefore, mMCP-11 is not essential for basophil infiltration in IgE-CAI. Our findings instead suggest the following scenario: when basophils migrate into the skin lesion and are activated by allergens, they release mMCP-11, which in turn recruits more and more basophils together with other leukocytes. If this is the case, the ablation of mMCP-11 in basophils would reduce the infiltration of leukocytes, including basophils, just as we observed in the KO mice.

The detailed molecular mechanism by which mMCP-11 triggers leukocyte migration remains to be determined in future studies, as does the mechanism for mast cell tryptase– and chymase-induced chemotaxis of leukocytes.^32-35  Nevertheless, our findings in the present study showed that the protease activity of mMCP-11 is essential for leukocyte migration and that a G protein-coupled receptor or receptors expressed on leukocytes is also involved. mMCP-11 does not seem to directly act on leukocytes to induce their migration. Most likely, mMCP-11 cleaves a serum protein, and its proteolytic product attracts basophils, eosinophils, and macrophages via a G protein–coupled receptor. In contrast, no significant migration of neutrophils was induced by mMCP-11 in the transwell migration assay, whereas intradermal administration of mMCP-11 elicited neutrophil infiltration in the skin. In the latter case, quantitative PCR analysis showed slightly upregulated expression of neutrophil-attracting chemokines such as CXCL1 and CXCL2 but without statistical significance (data not shown). The possible contribution of mMCP-11–elicited production of chemokines in the skin lesion to the migration of neutrophils and other leukocytes needs to be further explored.

mMCP-11 is highly expressed by all basophils and by a small subset of mast cells to a much lesser extent.¹⁶  As for the case of basophils, the number of mast cells in the skin and peritoneum of KO mice was comparable to that of WT mice (supplemental Figure 1), suggesting the unimpaired development of mast cells in the absence of mMCP-11. Moreover, the extent of microvascular leakage in the IgE-mediated passive cutaneous anaphylaxis was comparable between KO and WT mice (supplemental Figure 6), indicating intact degranulation of mast cells in KO mice. In accordance with this finding, mast cell–dependent, immediate-type (early- and late-phase) ear swelling ahead of the IgE-CAI response was normally elicited in KO mice (data not shown), in contrast to the amelioration of IgE-CAI. Thus, mMCP-11 did not appear to contribute to the mast cell–mediated immediate-type reaction we observed. However, the functional significance of mMCP-11 in mast cell–mediated chronic inflammation, if any, remains to be investigated.

The hSPL-4 gene on human chromosome 16p13.3 has been identified as the human ortholog of the mMCP-11 gene located on mouse chromosome 17A3.3.¹⁵  However, the hSPL-4 gene possesses a premature translational-termination codon and, unlike the mMCP-11 gene, does not encode production of enzymatically active enzyme. Of note, human basophils express α- and β-tryptases, albeit to a lesser extent than do mast cells,^36,37  suggesting that α- and β-tryptases in human basophils might be functional orthologs of mMCP-11. Indeed, repeated intradermal injections of human β-tryptase elicited leukocyte infiltration into the skin of mice (supplemental Figure 7). In addition, β-tryptase induced transwell migration of basophils in vitro via a G protein–coupled receptors in a manner dependent on its protease activity and on the presence of serum (supplemental Figures 8 and 9), as did mMCP-11. It is noteworthy that trypsin, although it demonstrated higher activity than did β-tryptase and mMCP-11 when assessed by using a standard tryptase substrate (data not shown), showed little or no ability to induce basophil migration (supplemental Figure 8A), suggesting that not all tryptases share this property with mMCP-11.

In conclusion, the results of our study of mMCP-11–deficient mice demonstrate that basophil tryptase mMCP-11 plays a crucial role in the development of the IgE-CAI response through the promotion of microvascular leakage and leukocyte infiltration in a tryptase activity–dependent manner. This study has cast new light on the functional significance of proteases produced and released by basophils.

The authors thank J. Miyazaki (Osaka University) for providing CAG-Cre mice.

This work was supported by research grants from Japan Science and Technology Agency (1A145) (H.K.) and the Japanese Ministry of Education, Culture, Sports, Science and Technology (15H05786 [H.K.] and 15K20969 [Y.Y.]).

Contribution: M.I. and K.T. performed experiments; H.D. contributed to the initial study; M.F. and S.Y. generated mMCP-11–deficient mice; S.S. generated rmMCP-11; S.Y. provided helpful suggestions; M.I., Y.Y., and H.K. wrote the manuscript; and Y.Y. and H.K. designed and supervised the study.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Yoshinori Yamanishi, Department of Immune Regulation, Graduate School of Medical and Dental Sciences, Tokyo Medical and Dental University, 1-5-45 Yushima, Bunkyo-ku, Tokyo 113-8519, Japan; e-mail: yamanishi.mbch@tmd.ac.jp.