CCL28-Deficient Mice Have Reduced Iga Antibody–Secreting Cells And

The Journal of Immunology

CCL28-Deﬁcient Mice Have Reduced IgA Antibody– Secreting Cells and an Altered Microbiota in the Colon

Kazuhiko Matsuo,* Daisuke Nagakubo,† Shinya Yamamoto,* Akiko Shigeta,‡ Shuta Tomida,x Mitsugu Fujita,‡ Takako Hirata,† Ikuo Tsunoda,‡ Takashi Nakayama,*,1 and Osamu Yoshie‡,1,2

CCL28 induces the migration of IgA Ab-secreting cells (ASCs) via CCR10 and also displays a potent antimicrobial activity in vitro. To explore the role of CCL28 in vivo, we generated CCL28-deficient mice. The mice exhibited a significant reduction and abnormal distribution of IgA ASCs in the lamina propria of the colon. The concentrations of total and Ag-specific IgA in the fecal extracts of CCL28-deficient mice were also drastically reduced. The average amount of IgA secreted by a single IgA ASC derived from the colon was also substantially reduced in CCL28-deficient mice. Furthermore, CCL28 was found to significantly increase the average amount of IgA secreted by a single IgA ASC derived from the colon in vitro. In contrast, the generation of IgA ASCs in Peyer’s and cecal patches was not significantly impaired in CCL28-deficient mice. We also found a relative increase in the Class Bacilli in the fecal extracts of CCL28-deficient mice and demonstrated a potent antimicrobial activity of CCL28 against Bacillus cereus and Enterococcus faecalis, both of which belong to Class Bacilli. Thus, CCL28 may also suppress the outgrowth of some bacterial species by its direct antimicrobial activity. Finally, CCL28-deficient mice exhibited a highly aggravated dextran sodium sulfate–induced colitis that was ameliorated by pretreatment with antibiotics. Collectively, CCL28 plays a pivotal role in the homing, distribution, and function of IgA ASCs in the colon and may also affect the intestinal microbiota through its direct antimicrobial activity. The Journal of Immunology, 2018, 200: 800–809.

hemokines are a large family of small structurally related are $18 typical and 5 atypical chemokine receptors in humans chemotactic cytokines that play key roles in innate and (1, 2). The typical chemokine receptors are divided into four C acquired immunity by inducing migration of various subfamilies based on the subfamily of their signaling ligands leukocytes through a group of seven transmembrane G protein– (1, 2). Chemokines are now known to be involved in the health coupled receptors (1, 2). Humans have $46 chemokines, which and disease of various body tissues, including the gastrointestinal are grouped into four subfamilies (CXC, CC, C, and CX3C) by the tract (1, 2). motifs of the N-terminal conserved cytokine residues. Chemokine Because the mucosal surface is the major entry site for incoming receptors are functionally subdivided into two categories, typical pathogens, it is protected by innate and adaptive immune responses and atypical, depending on their signaling properties (1, 2). There against invading pathogens (3, 4). For example, the intestinal mucosal surface is covered with a thick mucus layer that prevents luminal microbes from attaching to the mucosal surface (3). In *Division of Chemotherapy, Kindai University Faculty of Pharmacy, Higashi-Osaka, Osaka 577-8502, Japan; †Department of Fundamental Biosciences, Shiga University addition, the epithelial cells produce antimicrobial peptides, such of Medical Science, Otsu, Shiga 520-2192, Japan; ‡Department of Microbiology, as defensins and cathelicidins, which together with phagocytic Kindai University Faculty of Medicine, Osakasayama, Osaka 589-8511, Japan; and xDepartment of Biobank, Graduate School of Medicine, Dentistry and Pharmaceuti- cells, constitute important components of innate immunity (3, 5). cal Sciences, Okayama University, Kita-ku, Okayama 700-8558, Japan Furthermore, IgA is the major class of Ab found in mucosal se- 1T.N. and O.Y. are cosenior authors. cretions, including tears, saliva, and secretions from the intestinal 2Current address: The Health and Kampo Institute, Sendai, Miyagi, Japan. and respiratory tracts. Secretory IgA maintains homeostasis of ORCIDs: 0000-0002-2663-2124 (A.S.); 0000-0002-1740-6168 (T.H.); 0000-0003- commensal microbiota and prevents the invasion of incoming 1798-714X (I.T.). pathogens (3, 6, 7). Received for publication January 9, 2017. Accepted for publication November 14, IgA Ab-secreting cells (ASCs) are generated in mucosal- 2017. induction sites, such as Peyer’s patches, and migrate to the mu- This work was supported by a Grant-in-Aid from the Ministry of Education, Culture, cosal effector sites, the lamina propria. CCR9 and CCR10 are the Sports, Science and Technology, Japan (26460582 to O.Y.), by the Ministry of Ed- best-characterized chemokine receptors that direct IgA ASCs to ucation, Culture, Sports, Science and Technology–Supported Program for the Stra- tegic Research Foundation at Private Universities, 2014–2018 (S1411037 to T.N.), the lamina propria. CCL25, the ligand for CCR9, is selectively and by the Kindai University Fund for Antiaging Center Project (to T.N.). expressed in the small intestine, whereas CCR9 is expressed on K.M., D.N., S.Y., A.S., and M.F. performed experiments and created the figures; S.T. almost all T cells and a fraction of IgA ASCs in the small intestine analyzed sequencing data; T.N. and O.Y. conceived and organized the study and (8–10). Thus, the CCL25–CCR9 axis is considered to have a analyzed data; and K.M., T.H., I.T., T.N., and O.Y. wrote the manuscript. major role in the homing of T cells and IgA ASCs to the small Address correspondence and reprint requests to Prof. Takashi Nakayama, Division of Chemotherapy, Kindai University Faculty of Pharmacy, Kowakae 3-4-1, Higashi- intestine. In contrast, CCL28, the ligand for CCR10, is widely Osaka, Osaka 577-8502, Japan. E-mail address: [email protected] expressed in various mucosal tissues, including the gastrointestinal Abbreviations used in this article: ASC, Ab-secreting cell; DSS, dextran sodium sulfate; tract, whereas CCR10 is expressed on nearly all IgA ASCs, but r EGFP, enhanced GFP; Neo , neomycin resistance gene; poly(I:C), polyinosinic- not on T cells, in the mucosal tissues (9–11). Thus, the CCL28– polycytidylic acid; PPB, potassium phosphate buffer; WT, wild-type. CCR10 axis is considered to play a key role in the common Copyright Ó 2018 by The American Association of Immunologists, Inc. 0022-1767/18/$35.00 mucosal immune system by distributing IgA ASCs to various www.jimmunol.org/cgi/doi/10.4049/jimmunol.1700037 The Journal of Immunology 801 mucosal tissues (12). Of note, cutaneous lymphocyte Ag–positive T cells also express CCR10 but only migrate to the skin where relevant adhesion molecules and CCL27, another ligand for CCR10, are selectively expressed (13). We demonstrated previously, using mice orally immunized with cholera toxin, that the administration of anti-CCL25 Ab suppressed the homing of Ag-specific IgA ASCs to the small intestine, whereas anti-CCL28 Ab suppressed the homing of Ag-specific IgA ASCs to the small and large intestines (10). Similarly, Feng et al. (14) reported that coadministration of anti-CCL25 and anti-CCL28 Abs, but neither one alone, effectively suppressed the homing of total and Ag-specific IgA ASCs to the small intestine in rotavirus- infected suckling mice. These results supported a pivotal role for the CCL28–CCR10 axis in the homing of IgA ASCs to the gastrointestinal tract. Furthermore, we (15) and other investigators (16, 17) have shown that CCL28 has potent antimicrobial activity against a broad spectrum of microbes in vitro. However, Morteau et al. (18) found normal levels of IgA ASCs in the small and large intestines of CCR10-deficient mice and found only a significant reduction in IgA ASCs in the lactating mammary gland. These findings might have cast doubt upon the essential role of the CCL28–CCR10 axis in the recruitment of IgA ASCs to the gastrointestinal tract. Subsequently, however, Hu et al. (19) demonstrated that enhanced generation of IgA ASCs in local isolated lymphoid follicles compensated for the number of IgA ASCs in FIGURE 1. Generation of CCL28-deficient mice. CCL28-deficient mice the intestine of CCR10-deficient mice. However, the compensa- (CCL28-knockout/EGFP–knock-in mice) were generated on the C57BL/6 tory IgA ASCs in CCR10-deficient mice carried fewer hyper- background by replacing exon 1 of the CCL28 gene with a gene cassette r A mutations in their IgH alleles compared with those of wild-type encoding EGFP and Neo .( ) Targeting strategy: the murine genomic CCL28 locus, the targeting vector construct, and the targeted allele with a (WT) mice (19). Furthermore, IgA memory B cells could not be cassette encoding EGFP and Neor are shown. (B) Genomic PCR analysis. properly maintained in the intestine of CCR10-deficient mice (19). PCR was performed for CCL28 using genomic DNA prepared from the Thus, the CCL28–CCR10 axis may have a role in the recruitment tails of WT mice, heterozygous CCL28-deficient mice (CCL28+/2), and of IgA ASCs to the intestine, as well as in their survival and/or homozygous CCL28-deficient mice (CCL282/2). Representative results function. from two separate experiments are shown. (C) RT-PCR analysis. CCL28 In the current study, we generated CCL28-deficient mice to expression was determined by RT-PCR using cDNA samples derived from further explore the role of the CCL28–CCR10 axis in mucosal the indicated tissues. GAPDH was used as an internal control. Represen- immunity. In CCL28-deficient mice, IgA ASCs in the lamina tative results from two separate experiments are shown. propria of the colon were significantly decreased, and the fecal IgA levels were drastically reduced. IgA ASCs derived from the colon of CCL28-deficient mice also secreted less IgA than those overnight and boiling at 90˚C for 10 min, we used 2 ml of a 1:10 dilution of the supernatant directly for PCR. The WT allele was amplified by a of WT mice. In this context, we have found that CCL28 enhances primer pair of +59-GTGTGGCTTTTCAAACCTCAGAA-39 (forward) and IgA secretion by IgA ASCs derived from the colon in vitro. 259-GCAATGGAAAAGTCTTACCCAAG-39 (reverse), and the mutated allele CCL28-deficient mice also had an altered intestinal microbiota (neomycin cassette) was amplified by a primer pair of +59-CTTGGGTGGA- and suffered from a highly aggravated dextran sodium sulfate GAGGCTATTC-39 (forward) and 259-AGGTGAGATGACAGGAGATC-39 (DSS)-induced colitis that was ameliorated by a pretreatment with (reverse). antibiotics. Collectively, our findings reveal the pivotal role of RT-PCR CCL28 in the homing, distribution, and function of IgA ASCs in Conventional RT-PCR was performed using KOD FX DNA polymerase the colon. CCL28 may also affect the intestinal microbiota (Toyobo, Osaka, Japan), as described previously (20). The amplification through its potent antimicrobial activity. conditions were denaturation at 94˚C for 30 s (5 min for the first cycle), annealing at 60˚C for 30 s, and extension at 72˚C for 30 s (5 min for the Materials and Methods last cycle) for 30 cycles for CCL28 and 25 cycles for GAPDH. Amplifica- Mice tion products were electrophoresed on 2% agarose gel and stained with ethidium bromide. The following primer pairs were used: +59-CATA- CCL28-deficient mice were generated by replacing exon 1 of the CCL28 CTTCCCATGGCCTCC-39 (forward) and 259-GAGAGGCTTCGTGCCTGTG- gene with a cassette encoding enhanced GFP (EGFP) and the neomycin 39 (reverse) for CCL28 and +59-GAGAGGCTTCGTGCCTGTG-39 (forward) resistance gene (Neor) using BRUCE-4 embryonic stem cells derived from and 259-TCCACCACCCTGTTGCTGTA-39 (reverse) for GAPDH. the murine strain C57BL/6J (Merck Millipore, Billerica, MA) (Fig. 1A) and backcrossed with C57BL/6J mice for $12 generations. C57BL/6J Immunofluorescence staining mice were purchased from CLEA Japan (Tokyo, Japan). Mice were Immunofluorescence staining of tissues was conducted as described pre- maintained in specific pathogen–free conditions and used at 8–10 wk of viously (21). Briefly, colon and sublingual gland tissues were dissected, age. All animal experiments were approved by the Center of Animal fixed overnight in 1% paraformaldehyde in PBS, washed with 20% sucrose Experiments, Kindai University, and performed in accordance with insti- in PBS, embedded in OCT compound (Sakura Finetek Japan, Tokyo, Ja- tutional guidelines. pan), frozen, and stored at 280˚C until use. Sections were cut and rehy- Genotyping PCR drated in PBS. For CCL28 staining, sections were blocked with donkey IgG in 2% BSA in PBS, incubated at 4˚C with goat anti-CCL28 (R&D To prepare DNA, each tail tip was put in a solution consisting of 99 mlof Systems, Minneapolis, MN) overnight, and incubated with Alexa Fluor lysis buffer (150 mM NaCl, 10 mM Tris-HCl [pH 8], 10 mM EDTA, and 488–labeled donkey anti-goat IgG (Thermo Fisher Scientific, Waltham, 0.1% SDS) and 1 ml of proteinase K (20 mg/ml). After heating at 55˚C MA) for 1 h. Sections were blocked again with goat IgG in 2% BSA in 802 ROLE OF CCL28 IN THE COLON

PBS, incubated overnight with anti-CD31 (MEC13.3; BD Biosciences, San Diego, CA), and incubated with Alexa Fluor 647–labeled goat anti-rat IgG (Thermo Fisher Scientific) and DAPI (Dojindo) for 1 h. For IgA staining in colon tissues, sections were blocked with anti-CD16/CD32 (2.4G2; BD Biosciences) in 2% BSA in PBS for 15 min and a Streptavidin/Biotin Blocking Kit (Vector Laboratories, Burlingame, CA) for 15 min. Sections were incubated with biotinylated rat anti-mouse IgA (RMA-1; BioLegend, San Diego, CA) at 4˚C overnight, washed, and incubated with Alexa Fluor 555–labeled streptavidin (Thermo Fisher Scientific) and DAPI for 1 h. For IgA staining of the sublingual gland, sections were blocked with goat IgG in 2% BSA in PBS. Sec- tions were incubated with biotinylated rat anti-mouse IgA at 4˚C overnight, washed, and incubated with Alexa Fluor 647–labeled goat anti-rat IgG and DAPI for 1 h. After being mounted with ProLong Gold Antifade Reagent (Thermo Fisher Scientific), sections were observed under a confocal laser scanning microscope (FV10000-D; Olympus, Tokyo, Japan). Immunization and sample collection WT and CCL28-deficient mice were inoculated intranasally with 10 mlof PBS containing 100 mg of OVA and 10 mg of polyinosinic-polycytidylic acid [poly(I:C); both from Sigma-Aldrich, St. Louis, MO] five times at 2-wk intervals. Mice were sacrificed 2 wk after the last sensitization. Sera and fecal extracts were collected to quantify total and Ag-specific IgA and IgG levels. Fecal extracts were obtained by adding weighed pellets to sterile PBS (100 mg of feces per milliliter). The suspensions were vortexed and centrifuged at 15,000 3 g for 20 min, and supernatants were used as fecal extracts. ELISA Total IgG and IgA concentrations in sera, fecal extracts, and culture supernatants were measured using mouse IgG or IgA ELISA kits (eBio- science, San Diego, CA), following the manufacturer’s instructions. Anti- OVA IgG and IgA titers were determined by ELISA, following previously described protocols (22). End point titers of OVA-specific IgG and IgA were expressed as the reciprocal log2 of the last dilution that had an absorbance . 0.1 after subtracting the background. Isolation of lymphoid cells from lamina propria Lymphoid cells were isolated from the lamina propria of small intestines or colons. In brief, small intestines were carefully cleared of Peyer’s patches. Small intestines and colons were cut open longitudinally, cut into 5-mm segments, and washed at room temperature with vigorous shaking four times in divalent cation–free HBSS supplemented with 5 mM EDTA and 25 mM HEPES to remove epithelial cells and intraepithelial lymphocytes until no more shedding occurred. The intestinal tissues were washed twice in RPMI 1640 containing 10% FBS, 15 mM HEPES, and antibiotics. Lamina propria lymphocytes were released from intestinal tissues by shaking for 40 min in RPMI 1640 containing 20% FCS, 25 mM HEPES, antibiotics, and 0.4 U/ml collagenase type D (Roche Diagnostics, Indi- anapolis, IN) three times. Floating cells obtained after each 40-min incubation were pooled and resuspended in RPMI 1640 containing 10% FBS and antibiotics. To isolate lymphocytes from Peyer’s patches, cecal patches, and spleens, tissues were gently squeezed between glass slides, and cells were passed through 70-mm nylon mesh. ELISPOT Lymphoid cells prepared from lamina propria were seeded onto nitrocel- lulose 96-well plates at 2.5 3 104 cells per well. After a 24-h incubation, the frequency of total IgA ASCs was evaluated using a Mouse IgA ELISpot kit (Mabtech, Nacka Strand, Sweden), following the manufacturer’s instructions. In some experiments, the 24-h incubation was performed in the presence of 300 nM recombinant mouse CCL28 (R&D Systems). The average amount of IgA secreted by a single IgA ASC was calculated as follows: IgA production (picogram per cell) = total

FIGURE 2. Immunofluorescence staining of CCL28 and IgA in the and sublingual gland from WT and CCL28-KO mice were stained for colon and sublingual gland. (A) Immunofluorescence staining of CCL28. CCL28 (green) and CD31 (red). White arrowheads indicate cells co- Tissue sections of the colon and sublingual gland from WT and CCL28- expressing CCL28 and CD31. Scale bars, 20 mm. Representative results from deficient (CCL28-KO) mice were stained for CCL28 (green) and DAPI three separate experiments are shown. (C) Immunofluorescence staining of (blue). Scale bars, 100 mm. Representative results from three separate IgA. Tissue sections of the colon and sublingual gland from WT and experiments are shown. (B) Double immunofluorescence staining of CCL28-KO mice were stained for IgA (red) and DAPI (blue). Scale bars, CCL28 and CD31. Tissue sections of the colon (Figure legend continues) 100 mm. Representative results from three separate experiments are shown. The Journal of Immunology 803 secreted IgA quantified by ELISA (picogram)/total number of IgA ASCs Antimicrobial assay enumerated by ELISPOT (cell). The spot number was counted, and the average spot area was quantified using a VHX-6000 digital microscope Bacillus cereus ATCC 14579 and Enterococcus faecalis ATCC 29212 (Keyence, Tokyo, Japan). were cultured with defibrinated sheep blood on nutrient agar plates and trypticase soy agar plates, respectively. Antimicrobial assays were con- Flow cytometry ducted in low-ionic conditions, essentially as described previously (15). In brief, microbial cells in midlogarithmic growth phase were resuspended at Cells were treated with anti-CD16/32 (clone: 2.4G2; BioLegend) in PBS 105 cells per milliliter in 1 mM potassium phosphate buffer (PPB; a containing 0.1% BSA to block Fcg receptor binding. After washing, cells mixture of 1 mM K2HPO4 and 1 mM KH2PO4; pH 7.2), mixed with an were incubated for 30 min with a mixture of PE-Cy7–conjugated anti- equal volume of 1 mM PPB, with or without recombinant mouse CCL28 B220 (clone RA3-6B2; eBioscience), allophycocyanin-conjugated (R&D Systems) in 2-fold serial dilutions, and incubated in a U-bottom CD138 (clone 281-2), and FITC-conjugated anti-IgA (clone C10-3) (BD 96-well microtest plate at room temperature for 2 h with brief mixing Biosciences) or with a mixture of PE-conjugated anti-CD4 (clone every 15 min. After appropriate dilutions with 1 mM PPB, cells were GK1.5) and allophycocyanin-conjugated anti-CD8 (clone 53-6.7; both plated on agar plates and grown at 37˚C overnight. Colonies were counted 6 from BioLegend). After washing, 1 3 10 cells were analyzed on a to obtain the percentage of viable cells. BD LSRFortessa (BD Biosciences) using FlowJo software (TreeStar, Ashland, OR). DSS-induced colitis Adoptive transfer WT and CCL28-deficient mice were given 2% DSS (molecular mass 36–50 kDa; MP Biomedicals, Strasbourg, France) in drinking water for 7 d, as Splenocytes from WT mice were labeled with CellTracker Orange CMTMR described previously (24). Control mice received drinking water without (Thermo Fisher Scientific) and transferred i.v. to WT and CCL28-deficient DSS. To deplete the intestinal commensal bacteria, mice were treated with recipient mice at 1 3 107 cells per mouse. After 24 h, mice were sacrificed, an antibiotic mixture containing 1 mg/ml ampicillin (Sigma-Aldrich), and large and small intestines were removed. Mononuclear cells were 500 mg/ml vancomycin, 1 mg/ml neomycin sulfate, and 1 mg/ml metro- prepared from the lamina propria and analyzed by flow cytometry as de- nidazole (all from Nacalai Tesque, Kyoto, Japan) in the drinking water for scribed above. The percentage of CellTracker Orange-labeled IgA ASCs 5 d prior to treatment with DSS, as described previously (25). Mice were was calculated. weighed daily. The length and histological scores of the colon were assessed on day 7, as described previously (26). In brief, the inflammatory 16S rRNA sequencing analysis of intestinal microbiota cell infiltration was scored as follows: 0, occasional inflammatory cells in To ensure the same environmental conditions, weaning WT and CCL28- the lamina propria; 1, increased numbers of inflammatory cells in the deficient mice were housed in the same cage until fecal samples were lamina propria; 2, confluence of inflammatory cells extending into the submucosa; and 3, transmural inflammation. The tissue damage was scored collected at 10 wk of age (23). Genomic DNA was extracted from fecal as follows: 0, no mucosal damage; 1, lymphoepithelial lesions; 2, surface samples of WT and CCL28-deficient mice using a QIAamp DNA Micro mucosal erosion or focal ulceration; and 3, extensive mucosal damage and Kit (QIAGEN, Hilden, Germany). Using primers designed to target the V3 extension into deeper structures of the bowel wall. The combined histo- and V4 regions of bacterial 16S rRNA, the bacterial populations of fecal samples were analyzed by paired-end reads of 300-bp amplicon se- logical scores ranged from 0 (no changes) to 6 (extensive cell infiltration quencing using a MiSeq platform (Illumina, San Diego, CA) and the and tissue damage). Histological scoring was done for each mouse by two investigators who were blinded to the treatment conditions. manufacturer’s protocol (http://www.illumina.com/content/dam/illumina- support/documents/documentation/chemistry_documentation/16s/16s- Gram staining metagenomic-library-prep-guide-15044223-b.pdf). After initial filtering by removing low-quality reads, the extracted 16S rRNA sequences were an- Colon tissues from DSS-treated WT and CCL28-deficient mice were fixed notated at the class level by the metagenomic workflow on the Illumina overnight in 1% paraformaldehyde at 4˚C, embedded in paraffin, and cut BaseSpace application (Illumina). into 5-mm-thick sections. Gram staining of colon tissues was performed

FIGURE 3. Total and Ag-specific IgA and IgG levels in serum and fecal samples. (A and B) Serum and fecal samples were obtained from WT and CCL28-deficient (CCL28-KO) mice. Total IgA (A) and total IgG (B) concentrations were quantified by ELISA. The data are expressed as mean 6 SE of 10 mice. (C and D)WT and CCL28-deficient mice were inoculated intranasally with 10 ml of PBS containing 100 mg of OVA and 10 mg of poly(I:C) as an adjuvant five times at 2-wk intervals. Two weeks after the last immunization, serum and fecal samples were collected. OVA-specific IgA (C) and OVA-specific IgG (D) concentrations were quantitated by ELISA. The data are expressed as mean 6 SE of 10 mice. **p , 0.01, ***p , 0.001. 804 ROLE OF CCL28 IN THE COLON

using a Gram staining kit for tissue (MUTO PURE CHEMICALS, Tokyo, Japan), following the manufacturer’s instructions. Statistical analysis The Student t test was performed to analyze differences between the two groups. One-way ANOVA with the Holm post hoc test was performed for multiple groups. We considered p , 0.05 as statistically signiﬁcant.

Results Generation of CCL28-deficient mice We generated CCL28-deficient mice by replacing exon 1 of CCL28 with a cassette encoding EGFP and Neor (Fig. 1A). We confirmed deletion of exon 1 in genomic DNA by PCR (Fig. 1B) and the absence of CCL28 mRNA in the mucosal tissues of CCL28- deficient mice by RT-PCR (Fig. 1C). CCL28-deficient mice were born in a Mendelian ratio, developed normally, and had no obvious phenotypic abnormality in specific pathogen–free conditions. We backcrossed CCL28-deficient mice to C57BL/6J mice for $12 generations before use. We first examined EGFP expression but were unable to detect EGFP in the mucosal tissues of our CCL28-deficient mice. This led us to perform standard immunofluorescence staining for CCL28 in the mucosal tissues of WT and CCL28-deficient mice. CCL28 signals were detected in the epithelial cells of mucosal tissues, including the colon and sublingual gland, of WT mice but not CCL28-deficient mice (Fig. 2A). Because CCL28 was also shown to be present on the venular endothelial cells of the small intestine and colon (10), we further examined colocalization of CCL28 and CD31, an endothelial cell marker. Indeed, CCL28 signals were often colocalized with CD31+ vasculature in the colon and sublingual gland of WT mice (Fig. 2B). These results corroborated the previous observations on CCL28 expression (9–11, 15). Reduction in IgA ASCs in mucosal tissues of CCL28-deficient mice Using immunofluorescence staining of IgA, we next examined numbers of IgA ASCs in mucosal tissues of WT and CCL28- deficient mice. We identified IgA ASCs as IgA+ cells with a plasma cell–like morphology. As shown in Fig. 2C, CCL28- deficient mice had markedly reduced IgA ASCs in the colon and sublingual gland compared with WT mice. Furthermore, although IgA ASCs were localized in the area adjacent to the muscularis mucosae, as well as in the upper lamina propria of the colon in WT mice, they were primarily found in the area adjacent to the muscularis mucosae in CCL28-deficient mice (Fig. 2C). The results suggested that homing to the colon, as well as the distribution of IgA ASCs within the lamina propria, was substantially affected in the absence of CCL28.

and CCL28-KO mice, stained for CD4 or CD8, and analyzed by flow cytometry. The frequencies of CD4+ T cells and CD8+ T cells are expressed as mean 6 SE of four mice. (C) Enumeration of IgA+B220+ cells in Peyer’s patches and cecal patches. Mononuclear cells were isolated from the Peyer’s patches and cecal patches of WT and CCL28-KO mice, stained for surface B220 and IgA, and analyzed by flow cytometry. The percentage FIGURE 4. Decreased IgA ASCs in the colon of CCL28-deficient mice. frequency of IgA+B220+ cells is indicated as mean 6 SE of five mice. (A) Enumeration of IgA ASCs. Mononuclear cells were isolated from the (D) Adoptive transfer experiment. Splenocytes from WT mice were labeled lamina propria of the colon and small intestine of WT and CCL28-KO with CellTracker Orange and injected i.v. into WT and CCL28-KO mice at mice. IgA ASCs were enumerated by ELISPOT. Spot numbers were 1 3 107 cells per mouse. After 24 h, mononuclear cells isolated from the quantified using a VHX-6000 digital microscope (Keyence). The data are lamina propria of the colon and small intestine were stained for surface B220 expressed as mean 6 SE of six mice. (B) Enumeration of CD4+ and CD8+ and IgA and analyzed by flow cytometry. The percentage frequency of T cells in the lamina propria of the colon. Mononuclear cells were isolated CellTracker Orange+B220+IgA+ cells is indicated as mean 6 SE of five from the lamina propria of the colon of WT (Figure legend continues) mice. *p , 0.05, **p , 0.01. The Journal of Immunology 805

CCL28-deficient mice exhibit a drastic reduction in total and Ag-specific IgA levels in feces We next determined total IgA levels in the sera and fecal samples of WT and CCL28-deficient mice. As shown in Fig. 3A, serum IgA concentrations were reduced by ∼50% in CCL28-deficient mice. Furthermore, fecal IgA levels were drastically reduced in CCL28- deficient mice. In contrast, serum IgG concentrations were increased by ∼2-fold in CCL28-deficient mice (Fig. 3B). IgG concentrations in fecal extracts of WT and CCL28-deficient mice were very low (Fig. 3B). We also quantified Ag-specific IgA and IgG concentrations in sera and fecal samples after intranasal sensitization with OVA using poly(I:C) as an adjuvant. Although the concentrations of serum OVA-specific IgA levels were similar between WT and CCL28-deficient mice, OVA-specific IgA concentrations in the fecal extracts were again drastically lower in CCL28- deficient mice (Fig. 3C). In contrast, there were no significant differences in the serum and fecal OVA-specific IgG concentrations between WT and CCL28-deficient mice (Fig. 3D). Collec- tively, total IgA and Ag-specific IgA were dramatically reduced in the feces of CCL28-deficient mice. Impaired homing of IgA ASCs to the colon of CCL28-deficient mice We next enumerated IgA ASCs in the lamina propria of the colon and the small intestine using IgA ELISPOT. As shown in Fig. 4A, IgA ASCs in the lamina propria of CCL28-deficient mice were reduced significantly (∼50%) in the colon but remained at normal levels in the small intestine compared with WT mice. These results are consistent with the roles of CCL25 and CCL28 in the homing of IgA ASCs in the small intestine and colon, respectively (8–11). In contrast, there were normal levels of CD4+ T cells and slightly, but not significantly, reduced levels of CD8+ T cells in the colon of CCL28-deficient mice (Fig. 4B). Furthermore, there were no significant differences in the frequency of IgA+ cells in total B220+ cells of Peyer’s and cecal patches (the mucosal-induction sites) between WT and CCL28-deficient mice (Fig. 4C). To further support the role of CCL28 in the homing of IgA ASCs to the colon, we performed an adoptive transfer experiment in which labeled WT splenocytes were transferred i.v. to WT and CCL28- deficient mice. As shown in Fig. 4D, CCL28-deficient mice had significantly fewer labeled IgA ASCs in the colon compared with WT mice. In contrast, similar numbers of labeled IgA ASCs were found in the small intestine of WT and CCL28-deficient mice. This supported the notion that the homing of IgA ASCs to the colon is selectively impaired in CCL28-deficient mice. Reduced secretion of IgA by IgA ASCs derived from the colon FIGURE 5. Reduced IgA secretion by IgA ASCs from the colon of of CCL28-deficient mice CCL28-deficient mice. (A) IgA ELISPOT assay. Mononuclear cells were isolated from the lamina propria of the colon and small intestine of WT and Using ELISPOT assays, we also noticed that IgA ASCs derived CCL28-KO mice. IgA ASCs were enumerated by ELISPOT. Representa- from the colon of CCL28-deficient mice developed spots that were tive images from a total of 12 assays are shown. Scale bars, 400 mm. (B) Quantitation of average spot area. The spot areas were quantified using a VHX-6000 digital microscope (Keyence). The data are expressed as mean 6 SE of six mice. (C) The average amount of IgA secreted by a single IgA CD138+B220+ IgA ASCs is shown as mean 6 SE of five mice. (E) Effect ASC. Mononuclear cells were prepared from the lamina propria of the colon of CCL28 on IgA secretion. Mononuclear cells were isolated from of WT and CCL28-KO mice. IgA ASCs were enumerated with ELISPOT. the lamina propria of the colon of WT and CCL28-deficient mice. The Secreted IgA concentrations were quantitated by ELISA. The average ELISPOT assay was performed with or without 300 nM CCL28 during the amount of IgA secreted by a single IgA ASC was calculated as follows: incubation period (24 h). Spot numbers and spot areas were quantified IgA production (picogram per cell) = total secreted IgA quantified by using a VHX-6000 digital microscope (Keyence). IgA ASCs were enu- ELISA (picogram)/total number of IgA ASCs enumerated by ELISPOT merated with ELISPOT, and secreted IgA concentrations were quantitated (cell). The data are expressed as mean 6 SE of six mice. (D) The surface by ELISA. The average amount of IgA secreted by a single IgA ASC was phenotype of IgA ASCs in the lamina propria of the colon. Mononuclear calculated as follows: IgA production (picogram per cell) = total secreted cells were prepared from the colon lamina propria of WT and CCL28-KO mice, IgA quantified by ELISA (picogram)/total number of IgA ASCs enumer- stained for surface IgA, CD138, and B220, and analyzed by flow cytometry. ated by ELISPOT (cell). The data are expressed as mean 6 SE of six mice. The percentage frequency of CD138+B2202 and (Figure legend continues) *p , 0.05, **p , 0.01. 806 ROLE OF CCL28 IN THE COLON less clear than those of WT mice (Fig. 5A). In contrast, IgA ASCs derived from the small intestine of WT and CCL28-deficient mice showed no such difference (Fig. 5A). Accordingly, IgA ASCs derived from the colon, but not from the small intestine, of CCL28-deficient mice had a significant reduction in the average spot area compared with those of WT mice (Fig. 5B). This led us to determine the average amount of IgA secreted by a single IgA ASC derived from the colon of WT and CCL28-deficient mice. As shown in Fig. 5C, the average amount of IgA secreted by a single ASC derived from CCL28-deficient mice was reduced by ∼70% compared with WT mice. Thus, the numbers of IgA ASCs, as well as the amounts of IgA secreted by individual IgA ASCs, were dramatically reduced in the colon of CCL28-deficient mice. Stimulatory effect of CCL28 on IgA secretion by IgA ASCs The terminal differentiation of IgA ASCs is accomplished in the intestinal lamina propria, where there are two major subsets of IgA ASCs: CD138+B220+ IgA ASCs and CD138+B2202 terminally differentiated IgA ASCs (27). Because the reduced secretion of IgA by IgA ASCs derived from the colon of CCL28-deficient mice might be due to impaired terminal differentiation of IgA ASCs in the absence of CCL28, we compared the surface phenotypes of IgA ASCs in the colon of WT and CCL28-deficient mice. As shown in Fig. 5D, the majority of IgA ASCs derived from the colon were CD138+B2202 terminally differentiated IgA ASCs in WT and CCL28-deficient mice. This suggested that the terminal differentiation of IgA ASCs was not significantly impaired in the colon of CCL28-deficient mice. This led us to examine the effect of CCL28 on IgA ASCs derived from the colon in vitro. The presence of CCL28 during the 24-h incubation period of the ELISPOT assay did not significantly increase the spot number of IgA ASCs derived from WT mice (Fig. 5E, left panel), but it increased their spot sizes. Accordingly, the average spot area (Fig. 5E, middle panel) and the average amount of secreted IgA (Fig. 5E, right panel) were significantly increased. Similarly, CCL28 did not affect the spot number of IgA ASCs derived from A CCL28-deficient mice (Fig. 5E, left panel), but it significantly FIGURE 6. Analysis of intestinal microbiota composition. ( ) Next- increased the average spot area (Fig. 5E, middle panel) and the generation sequencing-based 16S rRNA profiles of fecal samples from WT and CCL28-deficient (CCL28-KO) mice. Fecal samples from WT and average amount of secreted IgA (Fig. 5E, right panel). Thus, CCL28-KO mice were collected and analyzed by 16S rRNA sequencing. CCL28 seems to have a direct stimulating effect on IgA secretion Relative abundance of nine major bacterial classes is shown. The data are by IgA ASCs in the colon. expressed as mean 6 SE of results from five mice. (B) Antimicrobial activity of CCL28 against B. cereus and E. faecalis. Bacterial cells in the CCL28 deficiency affects intestinal microbiota composition midlogarithmic growth phase were washed and incubated with different We (15) and other investigators (16, 17) have shown that CCL28 concentrations of CCL28 in 1 mM potassium phosphate buffer (pH 7.2) at has a potent direct antimicrobial activity against a wide variety of room temperature for 2 h with brief mixing every 15 min. Viable cells were microbes in vitro. Therefore, we examined the intestinal micro- counted by CFU assay. All assays were done in triplicate. The error bars 6 biota of WT and CCL28-deficient mice by 16S rRNA sequencing indicate mean SE. Representative results from three separate experi- , of fecal samples using next-generation sequencing. As shown in ments are shown. *p 0.05. Fig. 6A, the abundance of Class Bacilli was significantly increased in CCL28-deficient mice compared with WT mice (p , 0.01). In (Fig. 7B, 7C), CCL28-deficient mice exhibited much more addition, although not statistically significant, Class a-proteobacteria, aggravated colitis than WT mice. Enhanced mucosal inflammation Class b-proteobacteria, Class Erysipelotrichi, and Class Mollicutes and tissue damage in the colon of CCL28-deficient mice were also were relatively increased in CCL28-deficient mice. These findings apparent by H&E staining (Fig. 7D) and by the assessment of prompted us to test the antimicrobial activity of CCL28 against histological scores (Fig. 7E). To examine the role of the intestinal the bacterial species belonging to Class Bacilli. As shown in microbiota in the aggravated colitis of CCL28-deficient mice, we Fig. 6B, CCL28 killed B. cereus and E. faecalis, both of which pretreated CCL28-deficient mice with a potent antibiotic mixture. belong to Class Bacilli, with a half-maximum killing concentra- As shown in Fig. 7F, the antibiotic pretreatment almost completely tion of 2.1 6 0.2 and 4.0 6 0.2 mM, respectively. ameliorated DSS-induced colitis in CCL28-deficient mice, as evidenced by maintenance of body weight at almost the normal Aggravated DSS-induced colitis in CCL28-deficient mice level. To further support the involvement of intestinal microbiota Because the intestinal IgA level and microbiota are both known to in the aggravated DSS colitis of CCL28-deficient mice, we per- affect the severity of DSS-induced colitis (28–30), we compared formed Gram staining of colon tissues obtained from WT and DSS-induced colitis in WT and CCL28-deficient mice. As shown CCL28-deficient mice treated with DSS for 2 d (Fig. 7G). Al- by the body weight loss (Fig. 7A), as well as by the colon length though only Gram-positive species were technically revealed, The Journal of Immunology 807

bacterial cells were frequently colonized in the lamina propria and also were found in the muscularis of the colon in CCL28-deﬁcient mice (Fig. 7G). In contrast, only a few bacterial cells were found in the colon lamina propria of WT mice (Fig. 7G). Thus, the colon mucosa of CCL28-deﬁcient mice was less well protected from bacterial invasion upon intestinal injury.

Discussion CCL28 is expressed by the epithelial cells of most mucosal tissues, including the lactating mammary gland, salivary gland, and gastrointestinal tract, and it induces homing of IgA ASCs to the mucosal effector sites (i.e., lamina propria) via CCR10 (9–11, 15). Accordingly, the CCL28–CCR10 axis is considered to play an important role in the common mucosal immune system (12). Furthermore, we (15) and other investigators (16, 17) have shown that CCL28 has a potent direct antimicrobial activity against a broad spectrum of microbes in vitro. Thus, CCL28 may have a unique role in mucosa tissues as a chemokine with a direct antimicrobial activity. In this study, we generated CCL28-deficient mice and examined the role of CCL28 in the colon. Our results demonstrated that the homing, distribution, and function of IgA ASCs in the colon are greatly compromised in the absence of CCL28. Furthermore, the intestinal microbiota is substantially altered in the absence of CCL28. Consequently, CCL28-deficient mice had highly aggravated DSS-induced colitis that could be ameliorated by antibiotics. Previous studies involving CCR10-deficient mice found normal levels of IgA ASCs in the colon (18, 19). The mice also displayed no reduction in fecal IgA levels (18, 19). In contrast, we found a substantial reduction in IgA ASCs in the colon of CCL28-deficient mice (Fig. 2). We further found a drastic decrease in total and Ag- specific IgA levels in the feces of CCL28-deficient mice (Fig. 3). Although we do not know the exact reason for such discrepancies between CCR10-deficient mice and CCL28-deficient mice, our following findings may explain the drastic decreases in the fecal IgA levels of CCL28-deficient mice (1). The number of IgA ASCs, as well as their distribution in the upper lamina propria, was markedly reduced in CCL28-deficient mice (Fig. 2). Because the localization of IgA ASCs in the upper lamina propria may be very important for the efficient transportation of secreted IgA from the lamina propria into the intestinal lumen, the abnormal distribution of IgA ASCs in the lamina propria may result in low IgA levels in the colonic lumen of CCL28-deficient mice (2). The average amount of IgA secreted by a single IgA ASC derived from the colon of CCL28-deficient mice was also decreased by ∼70% (Fig. 5). This may result, in part, from the fact that IgA ASCs

(D) H&E staining of colon on day 7. Representative histological images are shown. (E) Histology score. H&E-stained sections were scored blindly on a scale of 0 to 6 to generate a histological score. Each circle represents a single mouse. The scores were analyzed statistically using mean 6 SE of six mice. (F) Effect of antibiotic pretreatment on DSS-induced colitis. To deplete the intestinal commensal bacteria, CCL28-KO mice were treated or not with an antibiotic mixture (ampicillin, vancomycin, neomycin sulfate, and metronidazole) in their drinking water for 5 d. Then, colitis was induced by administration of 2% DSS in the drinking water for 7 d, fol- FIGURE 7. Severe DSS-induced colitis of CCL28-deficient mice. Co- lowed by regular drinking water. The body weight of each mouse was litis was induced in WT and CCL28-deficient (CCL28-KO) mice by ad- monitored daily. Each group consisted of six mice. (G) Gram staining. WT ministration of 2% DSS in the drinking water for 7 d, followed by regular and CCL28-KO mice were treated with 2% DSS in the drinking water for drinking water. (A) Body weight. Mice were weighed daily. (B) Macroscopic 2 d. Then, colons were removed, and tissue sections were Gram stained. images of the colon on day 7. (C) The average length of the colon on day 7 Representative histological images are shown. Higher-magnification im- (left panel). The length ratio of DSS(+)/DSS(2) is also shown (right panel). ages of the boxed areas are shown (a–d). The arrows indicate the presence The data are shown as mean 6 SE of six mice. (Figure legend continues) of bacterial cells. *p , 0.05, **p , 0.01, ***p , 0.001. 808 ROLE OF CCL28 IN THE COLON present in the colon of CCL28-deficient mice are primarily derived belonging to Class Gammaproteobacteria, which were shown to from local sources, producing low amounts of IgA (31–33). Fur- be highly susceptible to the antimicrobial activity of CCL28 in thermore, we have demonstrated that CCL28 itself has a signifi- previous studies, were not affected in CCL28-deficient mice. cant enhancing effect on IgA secretion by individual IgA ASCs Future studies using exogenous or transgenic CCL28 will be derived from the colon (Fig. 5). Because signaling via chemokine useful to explore the possible role of CCL28 as a direct antimi- receptors is known to enhance the secretion of various factors, crobial agent in vivo. including cytokines, chemical mediators, and perforin (2), the Secretory IgA and antimicrobial peptides are important for the presence of CCL28 in the lamina propria may be important for the homeostatic regulation of intestinal microbiota (3). A lack of IgA homing of IgA ASCs, as well as for their IgA secretion. Collec- has been shown to increase susceptibility to intestinal injury (34, tively, our findings demonstrate a pivotal role for CCL28 in the 39). Oral administration of multiple bacterial species–binding homing, distribution, and function of IgA ASCs in the colon. high-affinity IgA mAb ameliorated DSS-induced colitis in asso- The numbers of IgA ASCs and their secreted IgA levels were ciation with changes in the composition of intestinal microbiota normal in the small intestine of CCL28-deficient mice compared (29). In the current study, we observed a highly aggravated DSS- with WT mice (Fig. 4). This is consistent with the notion that IgA induced colitis in CCL28-deficient mice that was ameliorated by ASCs in the small intestine primarily use the CCL25–CCR9 axis pretreatment with antibiotics (Fig. 6). We demonstrated enhanced in their intestinal homing. The small contribution of IgA ASCs in bacterial invasion in the colon mucosa of CCL28-deficient mice the small intestine to the fecal IgA levels in CCL28-deficient mice after a short-term exposure to DSS (Fig. 7). The reduction in se- may be explained as follows. Although secreted IgA is transported cretory IgA levels in the lumen, as well as the lack of direct an- across epithelial cells into the intestinal lumen by the polymeric timicrobial activity of CCL28 in the colon, may be responsible for IgR, its level of expression in the intestinal tract is substantially the aggravated DSS-induced colitis in CCL28-deficient mice that higher in the colon than in the small intestine (34). Thus, IgA results from impaired protection of injured mucosa against the ASCs in the colon, but not those in the small intestine, may pri- intestinal microbiota. marily contribute to the fecal IgA levels. This is supported by the In conclusion, our results demonstrate that CCL28 plays a study that reported that CCR9-deficient mice had normal IgA pivotal role in colonic mucosal immunity, primarily as a chemokine levels in their feces (35). that recruits and stimulates IgA ASCs via CCR10 and possibly also Intestinal IgA ASCs are known to be induced in GALTs, such as as a direct antimicrobial agent in the mucosal secretion of the colon. Peyer’s patches, cryptopatches, and isolated lymphoid follicles Our results provide future directions for CCL28 research eluci- (31, 33). Recently, by using germ-free mice colonized with in- dating whether immune responses to intestinal infection are altered testinal bacteria after appendectomy, Masahata et al. (36) dem- in CCL28-deficient mice, as well as the roles of CCL28 in other onstrated that cecal patches are the major site for the induction of mucosal tissues (e.g., mammary gland and respiratory tract), as a IgA ASCs migrating to the small intestine and the colon, whereas tropic and stimulatory factor for IgA ASCs, and as an antimicrobial Peyer’s patches are the induction site for IgA ASCs that prefer- agent in the regulation of the intestinal microbiota. Such studies entially migrate to the small intestine. Accordingly, IgA ASCs in will be useful for our understanding of the mucosal homeostasis cecal patches express CCR10 at higher levels than those in Peyer’s and immune responses that are regulated by CCL28. The role of patches (36). We observed no significant difference in the fre- CCL28 in human diseases also needs to be explored. In this context, quencies of IgA+ B220+ cells in Peyer’s patches and cecal patches CCL28 expression has been shown to be increased in epithelial between WT and CCL28-deficient mice (Fig. 4). Thus, CCL28 cells by inflammatory stimuli in vitro and in the colonic epithelial deficiency may not significantly affect the generation of IgA ASCs cells of patients with Crohn’s disease (40). in intestinal induction sites. The reason for the discrepancy between total and OVA-specific IgA levels in the serum of CCL28- Acknowledgments deficient mice (Fig. 3) is not clear. Although OVA-specific IgA We thank Namie Sakiyama, Takahisa Yamawaki, Hashimoto Koji, and Moe levels reflect the ongoing IgA responses, total IgA levels may Sawaragi (Kindai University) for excellent technical assistance. We also represent the sum of long-term IgA ASCs. Because the intestine thank William Figoni (Kindai University) for helpful proofreading and com- provides a tissue microenvironment for the long-term survival of ments. IgA ASCs (37), CCL28 may indirectly support the long-term survival of IgA ASCs through their homing to the colon. Disclosures Antimicrobial peptides are known to play essential roles in The authors have no financial conflicts of interest. innate immunity against microbes (3, 5). Previously, we (15) and other investigators (16, 17) have shown that CCL28 has a potent References antimicrobial activity against a wide variety of microbes. The 1. Zlotnik, A., and O. Yoshie. 2012. The chemokine superfamily revisited. Immunity antimicrobial activity of CCL28 is also well conserved in different 36: 705–716. species, such as human, bovine, pig, and mouse (15–17). Like 2. Bachelerie, F., A. Ben-Baruch, A. M. Burkhardt, C. Combadiere, J. M. Farber, G. J. Graham, R. Horuk, A. H. Sparre-Ulrich, M. Locati, A. D. Luster, et al. most other antimicrobial peptides (38), CCL28 is inserted into the 2013. International union of basic and clinical pharmacology. [corrected]. bacterial membrane, leading to pore formation in target microbes LXXXIX. Update on the extended family of chemokine receptors and intro- ducing a new nomenclature for atypical chemokine receptors. [Published erra- (15). In this study, we found a relative increase in Class Bacilli in tum appears in 2014 Pharmacol. Rev. 66: 467.] Pharmacol. Rev. 66: 1–79. the fecal samples of CCL28-deficient mice (Fig. 6). We also 3. Chairatana, P., and E. M. Nolan. 2017. Defensins, lectins, mucins, and secretory demonstrated a potent direct killing activity of CCL28 against B. immunoglobulin A: microbe-binding biomolecules that contribute to mucosal immunity in the human gut. Crit. Rev. Biochem. Mol. Biol. 52: 45–56. cereus and E. faecalis, both of which belong to Class Bacilli 4. Perez-Lopez, A., J. Behnsen, S. P. Nuccio, and M. Raffatellu. 2016. Mucosal (Fig. 6). Thus, the absence of CCL28 may be responsible, in part, immunity to pathogenic intestinal bacteria. Nat. Rev. Immunol. 16: 135–148. for the outgrowth of some bacterial species, such as those be- 5. Diamond, G., N. Beckloff, A. Weinberg, and K. O. Kisich. 2009. The roles of antimicrobial peptides in innate host defense. Curr. Pharm. Des. 15: 2377– longing to Class Bacilli, in the colon. However, our results are far 2392. from sufficient to conclude the involvement of the direct anti- 6. Mantis, N. J., N. Rol, and B. Corthe´sy. 2011. Secretory IgA’s complex roles in immunity and mucosal homeostasis in the gut. Mucosal Immunol. 4: 603–611. microbial activity of CCL28 in the control of the intestinal micro- 7. Macpherson, A. J., M. B. Geuking, and K. D. McCoy. 2012. Homeland security: biota. It is also unclear why other bacterial species, such as those IgA immunity at the frontiers of the body. Trends Immunol. 33: 160–167. The Journal of Immunology 809

8. Bowman, E. P., N. A. Kuklin, K. R. Youngman, N. H. Lazarus, E. J. Kunkel, 22. Matsuo, K., Y. Ishii, Y. S. Quan, F. Kamiyama, Y. Mukai, Y. Yoshioka, N. Okada, J. Pan, H. B. Greenberg, and E. C. Butcher. 2002. The intestinal chemokine and S. Nakagawa. 2011. Transcutaneous vaccination using a hydrogel patch thymus-expressed chemokine (CCL25) attracts IgA antibody-secreting cells. induces effective immune responses to tetanus and diphtheria toxoid in hairless J. Exp. Med. 195: 269–275. rat. J. Control. Release 149: 15–20. 9. Kunkel, E. J., C. H. Kim, N. H. Lazarus, M. A. Vierra, D. Soler, E. P. Bowman, 23. Laukens, D., B. M. Brinkman, J. Raes, M. De Vos, and P. Vandenabeele. 2016. and E. C. Butcher. 2003. CCR10 expression is a common feature of circulating Heterogeneity of the gut microbiome in mice: guidelines for optimizing exper- and mucosal epithelial tissue IgA Ab-secreting cells. J. Clin. Invest. 111: 1001– imental design. FEMS Microbiol. Rev. 40: 117–132. 1010. 24. Wirtz, S., C. Neufert, B. Weigmann, and M. F. Neurath. 2007. Chemically in- 10. Hieshima, K., Y. Kawasaki, H. Hanamoto, T. Nakayama, D. Nagakubo, duced mouse models of intestinal inflammation. Nat. Protoc. 2: 541–546. A. Kanamaru, and O. Yoshie. 2004. CC chemokine ligands 25 and 28 play es- 25. Kirkland, D., A. Benson, J. Mirpuri, R. Pifer, B. Hou, A. L. DeFranco, and sential roles in intestinal extravasation of IgA antibody-secreting cells. F. Yarovinsky. 2012. B cell-intrinsic MyD88 signaling prevents the lethal dis- J. Immunol. 173: 3668–3675. semination of commensal bacteria during colonic damage. Immunity 36: 11. Lazarus, N. H., E. J. Kunkel, B. Johnston, E. Wilson, K. R. Youngman, and 228–238. E. C. Butcher. 2003. A common mucosal chemokine (mucosae-associated epi- 26. Siegmund, B., H. A. Lehr, G. Fantuzzi, and C. A. Dinarello. 2001. IL-1 beta- thelial chemokine/CCL28) selectively attracts IgA plasmablasts. J. Immunol. converting enzyme (caspase-1) in intestinal inflammation. Proc. Natl. Acad. Sci. 170: 3799–3805. USA 98: 13249–13254. 12. Kunkel, E. J., and E. C. Butcher. 2003. Plasma-cell homing. Nat. Rev. Immunol. 27. Fagarasan, S., and T. Honjo. 2003. Intestinal IgA synthesis: regulation of front- 3: 822–829. line body defences. Nat. Rev. Immunol. 3: 63–72. 13. Homey, B., W. Wang, H. Soto, M. E. Buchanan, A. Wiesenborn, D. Catron, ´ € 28. Kamada, N., S. U. Seo, G. Y. Chen, and G. Nun˜ez. 2013. Role of the gut A. Muller, T. K. McClanahan, M. C. Dieu-Nosjean, R. Orozco, et al. 2000. microbiota in immunity and inflammatory disease. Nat. Rev. Immunol. 13: Cutting edge: the orphan chemokine receptor G protein-coupled receptor-2 321–335. (GPR-2, CCR10) binds the skin-associated chemokine CCL27 (CTACK/ALP/ 29. Okai, S., F. Usui, S. Yokota, Y. Hori-I, M. Hasegawa, T. Nakamura, M. Kurosawa, ILC). J. Immunol. 164: 3465–3470. S. Okada, K. Yamamoto, E. Nishiyama, et al. 2016. High-affinity monoclonal 14. Feng, N., M. C. Jaimes, N. H. Lazarus, D. Monak, C. Zhang, E. C. Butcher, and IgA regulates gut microbiota and prevents colitis in mice. Nat. Microbiol. H. B. Greenberg. 2006. Redundant role of chemokines CCL25/TECK and 1: 16103. CCL28/MEC in IgA+ plasmablast recruitment to the intestinal lamina propria 30. Cao, A. T., S. Yao, B. Gong, C. O. Elson, and Y. Cong. 2012. Th17 cells up- after rotavirus infection. J. Immunol. 176: 5749–5759. regulate polymeric Ig receptor and intestinal IgA and contribute to intestinal 15. Hieshima, K., H. Ohtani, M. Shibano, D. Izawa, T. Nakayama, Y. Kawasaki, F. Shiba, M. Shiota, F. Katou, T. Saito, and O. Yoshie. 2003. CCL28 has dual homeostasis. J. Immunol. 189: 4666–4673. roles in mucosal immunity as a chemokine with broad-spectrum antimicrobial 31. Pabst, O. 2012. New concepts in the generation and functions of IgA. Nat. Rev. activity. J. Immunol. 170: 1452–1461. Immunol. 12: 821–832. 16. Berri, M., I. Virlogeux-Payant, C. Chevaleyre, S. Melo, G. Zanello, H. Salmon, 32. Suzuki, K., S. Kawamoto, M. Maruya, and S. Fagarasan. 2010. GALT: organi- and F. Meurens. 2014. CCL28 involvement in mucosal tissues protection as a zation and dynamics leading to IgA synthesis. Adv. Immunol. 107: 153–185. chemokine and as an antibacterial peptide. Dev. Comp. Immunol. 44: 286–290. 33. Sutherland, D. B., and S. Fagarasan. 2012. IgA synthesis: a form of functional 17. Pallister, K. B., S. Mason, T. K. Nygaard, B. Liu, S. Griffith, J. Jones, immune adaptation extending beyond gut. Curr. Opin. Immunol. 24: 261–268. S. Linderman, M. Hughes, D. Erickson, J. M. Voyich, et al. 2015. Bovine CCL28 34. Yel, L. 2010. Selective IgA deficiency. J. Clin. Immunol. 30: 10–16. mediates chemotaxis via CCR10 and demonstrates direct antimicrobial activity 35. Pabst, O., L. Ohl, M. Wendland, M. A. Wurbel, E. Kremmer, B. Malissen, and against mastitis causing bacteria. PLoS One 10: e0138084. R. Fo¨rster. 2004. Chemokine receptor CCR9 contributes to the localization of 18. Morteau, O., C. Gerard, B. Lu, S. Ghiran, M. Rits, Y. Fujiwara, Y. Law, plasma cells to the small intestine. J. Exp. Med. 199: 411–416. K. Distelhorst, E. M. Nielsen, E. D. Hill, et al. 2008. An indispensable role for 36. Masahata, K., E. Umemoto, H. Kayama, M. Kotani, S. Nakamura, T. Kurakawa, the chemokine receptor CCR10 in IgA antibody-secreting cell accumulation. J. Kikuta, K. Gotoh, D. Motooka, S. Sato, et al. 2014. Generation of colonic IgA- J. Immunol. 181: 6309–6315. secreting cells in the caecal patch. Nat. Commun. 5: 3704. 19. Hu, S., K. Yang, J. Yang, M. Li, and N. Xiong. 2011. Critical roles of chemokine 37. Xiong, N., and S. Hu. 2015. Regulation of intestinal IgA responses. Cell. Mol. receptor CCR10 in regulating memory IgA responses in intestines. Proc. Natl. Life Sci. 72: 2645–2655. Acad. Sci. USA 108: E1035–E1044. 38. Zasloff, M. 2002. Antimicrobial peptides of multicellular organisms. Nature 20. Nakayama, T., K. Hieshima, T. Arao, Z. Jin, D. Nagakubo, A. K. Shirakawa, 415: 389–395. Y. Yamada, M. Fujii, N. Oiso, A. Kawada, et al. 2008. Aberrant expression of 39. Wang, N., N. Shen, T. J. Vyse, V. Anand, I. Gunnarson, G. Sturfelt, S. Rantapaä¨- Fra-2 promotes CCR4 expression and cell proliferation in adult T-cell leukemia. Dahlqvist, K. Elvin, L. Truedsson, B. A. Andersson, et al. 2011. Selective IgA Oncogene 27: 3221–3232. deficiency in autoimmune diseases. Mol. Med. 17: 1383–1396. 21. Nagakubo, D., O. Yoshie, and T. Hirata. 2016. Upregulated CCL28 expression in 40. Ogawa, H., M. Iimura, L. Eckmann, and M. F. Kagnoff. 2004. Regulated pro- the nasal mucosa in experimental allergic rhinitis: implication for CD4(+) duction of the chemokine CCL28 in human colon epithelium. Am. J. Physiol. memory T cell recruitment. Cell. Immunol. 302: 58–62. Gastrointest. Liver Physiol. 287: G1062–G1069.