CXCR3.1 and CXCR3.2 Differentially Contribute to Macrophage Polarization in Teleost Fish

This information is current as Xin-Jiang Lu, Qiang Chen, Ye-Jing Rong, Feng Chen and of September 29, 2021. Jiong Chen J Immunol 2017; 198:4692-4706; Prepublished online 12 May 2017; doi: 10.4049/jimmunol.1700101

http://www.jimmunol.org/content/198/12/4692 Downloaded from

Supplementary http://www.jimmunol.org/content/suppl/2017/05/12/jimmunol.170010 Material 1.DCSupplemental http://www.jimmunol.org/ References This article cites 69 articles, 17 of which you can access for free at: http://www.jimmunol.org/content/198/12/4692.full#ref-list-1

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision by guest on September 29, 2021 • No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2017 by The American Association of Immunologists, Inc. All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

CXCR3.1 and CXCR3.2 Differentially Contribute to Macrophage Polarization in Teleost Fish

Xin-Jiang Lu, Qiang Chen, Ye-Jing Rong, Feng Chen, and Jiong Chen

The study of multiple copies of chemokine receptor genes in various teleosts has long appealed to investigators seeking to understand the evolution of the immune system. The CXCR CXCR3 gene has two isoforms, CXCR3.1 and CXCR3.2, which are both expressed in macrophages. The distinct roles of teleost CXCR3s have not been identiﬁed previously. In this article, we found that CXCR3.1 and CXCR3.2 differentially contributed to macrophage polarization in the teleosts: ayu (Plecoglossus altivelis), grass carp (Ctenopharyngodon idella), and spotted green pufferﬁsh (Tetraodon nigroviridis). In ayu macrophages, the P. altivelis CXCR3.1 (PaCXCR3.1) gene was constitutively expressed, whereas the P. altivelis CXCR3.2 (PaCXCR3.2) gene was induced postinfection with Escherichia coli. Upon E. coli infection, PaCXCR3.1+ and PaCXCR3.2+ macrophages showed an M1 and an M2 phenotype, respectively. CXCL9–

11-like proteins mediated M1 and M2 polarization by interacting with the PaCXCR3.1 and PaCXCR3.2 proteins on macrophages, Downloaded from respectively. The transcription factors P. altivelis STAT1 and P. altivelis STAT3 were activated in PaCXCR3.1+ and PaCXCR3.2+ macrophages, respectively. Furthermore, the prognosis of septic ayu adoptively transferred with PaCXCR3.2+ macrophages was improved. Our data reveal a previously unknown mechanism for macrophage polarization, suggesting that redundant genes may regulate crucial functions in the teleost immune system. The Journal of Immunology, 2017, 198: 4692–4706.

n mammals, the chemokine receptor CXCR3 exists as a polarization in a mouse cancer model (7). Two CXCR3 isoforms http://www.jimmunol.org/ single gene and is preferentially expressed on immune cells are found in poikilothermic vertebrates, including teleosts (bony I to aid in cell migration to the sites of inflammation (1). The fishes), amphibians, and reptiles (8, 9). CXCR3.1 (CXCR3b) and primary ligands of CXCR3 are CXCL9, CXCL10, and CXCL11 CXCR3.2 (CXCR3a) are found in zebrafish, Japanese ricefish, and (2, 3). CXCL10/CXCR3 interactions drive the proinflammatory tetraodon (10). In Japanese ricefish, CXCR3.2 is a marker of in- effector Th1 polarization of T cells, whereas CXCL11/CXCR3 nate immune cells (10). In zebrafish, CXCR3.2 mediates macro- binding induces an immune-tolerant state that is characterized phage chemotaxis to the sites of bacterial infection (11). Similarly, by anti-inflammatory Th2 polarization (4, 5). CXCR3 has been in rainbow trout, CXCR3.1 and CXCR3.2 are mainly expressed in detected in a small subset of peripheral blood monocytes and a macrophages (12). We ask why two CXCR3 genes exist in teleost high percentage of monocytes recruited to inflammatory sites (6). fish. The cost–benefit trade-off of a given trait is optimized by by guest on September 29, 2021 Although CXCR3 is not a classic marker for macrophage polari- maximizing the benefit while minimizing the cost (13). Thus, we zation, its deficiency has been shown to promote macrophage M2 hypothesized that, in teleosts, the two isoforms of CXCR3 may confer a specific benefit by regulating macrophage function. Macrophages are present in most vertebrate tissues and perform Laboratory of Biochemistry and Molecular Biology, Ningbo University, Ningbo functions encompassing host defense and tissue homeostasis (14, 315211, People’s Republic of China 15). The inflammatory state of macrophages plays an important Received for publication January 23, 2017. Accepted for publication April 18, 2017. role in pathogen clearance (16, 17), and their polarization to dif- This work was supported by the Program for the National Natural Science Founda- ferent inflammatory phenotypes depends on environmental cues or tion of China (Grants 31372555 and 31472300), the Zhejiang Provincial Natural Science Foundation of China (Grant LZ13C190001), the Young Talent Cultivation pathophysiologic conditions (18). The classically activated mac- Project of the Zhejiang Association for Science and Technology (Grant 2016YCGC003), rophages (M1 type) are induced by LPS and IFN-g and express National 111 Project of China, the LiDakSum Marine Biopharmaceutical Development Fund, and the K.C. Wong Magna Fund at Ningbo University. proinflammatory mediators. The alternatively activated macrophages (M2 type) are induced by IL-4 and IL-13 and express high The sequences presented in this article have been submitted to GenBank (https:// www.ncbi.nlm.nih.gov/genbank/) under accession numbers JP725619, KY081643, levels of anti-inflammatory mediators. Macrophage polarization KJ130413, KU362928, KU362929, JP742610, and JP722452. is also regulated by soluble proteins, intracellular signals, and Address correspondence and reprint requests to Prof. Jiong Chen, Ningbo University, transcription factors. Galectin-dependent regulatory signaling 818 Fenghua Road, Ningbo 315211, Zhejiang, People’s Republic of China. E-mail stimulates M2-type macrophage polarization (19). TLR signaling address: [email protected] activates the STAT1 protein to skew macrophage function toward The online version of this article contains supplemental material. the M1 phenotype, whereas the activation of STAT3 by IL-4 and Abbreviations used in this article: CiCXCR3.1, C. idella CXCR3.1; CiCXCR3.2, C. idella CXCR3.2; iNOS, inducible NO synthase; IsoIgG, mouse isotype IgG; MEGA, IL-13 skews macrophage function toward the M2 phenotype (20). molecular evolutionary genetics analysis; MOI, multiplicity of infection; NAC, N-acetyl The ablation of protein kinase Ba (Akt1) and protein kinase Bb cysteine; PaCSFR1, P. altivelis M-CSFR; PaCXCL9–11l1, P. altivelis CXCL9–11l1; (Akt2) differentially affects macrophage polarization (21). Similar PaCXCL9–11l2, P. altivelis CXCL9–11l2; PaCXCL9–11l3, P. altivelis CXCL9–11l3; PaCXCR3.1, P. altivelis CXCR3.1; PaCXCR3.2, P. altivelis CXCR3.2; PaIL-10, to those in mammals, macrophages in teleost fish are differen- P. altivelis IL-10; PaIL-12p40a, P. altivelis IL-12p40a; PaIL-12p40c, P. altivelis IL- tiated from hematopoietic stem/progenitor cells and perform 12p40c; PaIL-1b, P. altivelis IL-1b;PaSTAT1,P. altivelis STAT1; PaSTAT3, P. altivelis STAT3; PaTGF-b, P. altivelis TGF-b;PaTNF,P. altivelis TNF; P/S, penicillin/streptomycin; functions such as phagocytosis, bacterial killing, and cytokine RNAi, RNA interference; ROS, reactive oxygen species; RT-qPCR, real-time quantitative production (22–24). However, they possess additional functions PCR; siRNA, small interfering RNA; TnCXCR3.1, T. nigroviridis CXCR3.1; TnCXCR3.2, and regulatory mechanisms that differ from those in mammals. T. nigroviridis CXCR3.2. For example, teleost macrophages express novel cytokines to re- Copyright Ó 2017 by The American Association of Immunologists, Inc. 0022-1767/17/$30.00 gulate inflammatory reactions (25, 26) and are poorly activated by www.jimmunol.org/cgi/doi/10.4049/jimmunol.1700101 The Journal of Immunology 4693

TNF (27). Head kidney–derived macrophages in teleosts show transcriptomic data were deposited in the Gene Expression Omnibus functional polarization upon differential stimulation (28, 29). (http://www.ncbi.nlm.nih.gov/geo/) under accession number GSE40221 Furthermore, teleost macrophages have been shown to polarize (22). PCR, cloning, and sequencing were used to confirm the authenticity of these sequences. The similarity between the obtained sequences and into a continuum of different activation states ranging from innate other known sequences was analyzed using the basic local alignment search to alternative activation (30). Thus, it is essential to understand the tool (http://blast.ncbi.nlm.nih.gov/blast.cgi). Multiple sequence alignments signaling mechanisms that regulate macrophage phenotypes, be- were generated using ClustalW (http://clustalw.ddbj.nig.ac.jp/). Phylogenetic cause macrophage polarization plays an important role in systemic and molecular evolutionary analyses were conducted using the molecular evolutionary genetics analysis (MEGA) tool (version 5). inflammatory responses (31). In vertebrate evolution, gene duplications produce redundant genes, Primary culture of fish cells which are associated with adaptive radiations in teleosts (32, 33). Ayu macrophages were isolated and cultured as described previously (22). These redundant genes may result in subfunctionalization, as in the Briefly, ayu were killed by an overdose of anesthetic (0.03% [v/v] eth- case of sea bass (Dicentrarchus labrax), in which different hepcidins ylene glycol monophenyl ether). Head kidneys were removed immedi- exhibit different roles (34). In this study, we found that CXCR3.1 and ately to isolate the leukocytes. Head kidney leukocyte–enriched fractions CXCR3.2 were expressed in the macrophages of the teleosts: ayu were obtained using a Ficoll density gradient (Invitrogen, Shanghai, China). Ayu macrophages were purified by flow cytometry after staining (Plecoglossus altivelis), grass carp (Ctenopharyngodon idella), and with anti-ayu M-CSFR (P. altivelis M-CSFR [PaCSF1R]) IgG (35). At spotted green pufferfish (Tetraodon nigroviridis). This offered a least 10,000 events were collected from the macrophage gate. The unique opportunity to uncover the functional divergence of redundant macrophages were maintained at 24˚C in RPMI 1640 medium containing chemokine receptor genes in the teleost immune system. We further 5% FBS, 5% ayu serum, and 1% penicillin/streptomycin (P/S). Grass carp macrophages were separated using Ficoll density gradient (Invi-

+ + Downloaded from found that CXCR3.1 and CXCR3.2 macrophages exhibited M1 and trogen), in combination with centrifugation. Nonadherent cells were washed M2 polarization, respectively. Taken together, these data highlight a off, and the attached cells were incubated in complete medium (RPMI 1640, new mechanism for macrophage polarization in teleost ﬁsh, 5% grass carp serum, 5% FBS, and 1% P/S) at 24˚C with 5% CO2.Spotted suggesting an evolutionary differentiation in the immune response of green pufferﬁsh macrophages were separated as described above and cultured vertebrates. in RPMI 1640 supplemented with 10% FBS and 1% P/S at 28˚C. Macro- phages were stimulated with live E. coli (DH5a) at the indicated multiplicities of infection (MOIs). In preliminary experiments, no CFU was detected in the

Materials and Methods macrophages postinfection with E. coli at the indicated MOIs. Macrophages http://www.jimmunol.org/ Animals were also infected with Listonella anguillarum atanMOIof2.N-acetyl cysteine (NAC) was added to cells at a final concentration of 20 mM in this Ayu (Zhemin No. 1, weighing 40 6 5 g each) were used in all experiments. This study. strain has undergone seven successive generations of mass selection, with no graft-versus-host effects (23). The fish were kept in freshwater tanks at 20–22˚C Flow cytometry in a recirculating system using filtered water. Grass carp, with a body weight of The following peptides corresponding to the indicated amino acid residues for 100–120 g, were obtained from a commercial farm (Hunan, China). The fish each protein were chemically synthesized for mAb production (GL Biochem, were kept in freshwater tanks at 20–22˚C in a recirculating system. One-year- Shanghai, China): PaCXCR3.1 (aa 1–59), PaCXCR3.2 (aa 1–44), grass carp old wild-type spotted green pufferfish, weighing 3–6 g, were kept in recircu- CXCR3.1 (CiCXCR3.1, aa 1–62), CiCXCR3.2 (aa 1–48), spotted green lating water at 26–28˚C. They were held in the laboratory for $2 wk, with pufferfish CXCR3.1 (TnCXCR3.1, aa 1–53), and TnCXCR3.2 (aa 1–46). by guest on September 29, 2021 healthy appearance and normal activity, prior to use in experiments. All fish Flow cytometry was used to validate the specificity of these Abs (Supple- used in this study were healthy and without any pathological signs. Bacterial mental Fig. 1A). HEK-293T transfectants stably expressing PaCXCR3.1, CFU were not detected in the blood of healthy ayu. The experimental conditions PaCXCR3.2, CiCXCR3.1, CiCXCR3.2, TnCXCR3.1, and TnCXCR3.2 were and procedures were approved by the Ningbo University Institutional Animal generated as described previously (36). The primer sequences for the target Care and Use Committee and were carried out in compliance with the National genes are listed in Supplemental Table I. Fragments of PaCXCR3.1, Guide for the Care and Use of Laboratory Animals Institutes of Health’s . PaCXCR3.2, CiCXCR3.1, CiCXCR3.2, TnCXCR3.1, and TnCXCR3.2 Characterization of gene cDNA were ligated into pcDNA3.1Zeo (+) for expression. The reconstructed plasmids were transfected into HEK-293T cells using Lipofectamine plus The cDNA sequence of the ayu CXCR3.1 (P. altivelis CXCR3.1 (Invitrogen), according to the protocol provided by the manufacturer. [PaCXCR3.1]) gene was obtained from the transcriptome analysis of ayu Stable cell lines for expression of CXCR3.1 and CXCR3.1 were selected monocytes/macrophages, and transcriptomic data were deposited in the with Zeocin (Invitrogen) and screened for flow cytometry. Cells expressing Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo/) under ac- PaCXCR3.1, PaCXCR3.2, CiCXCR3.1, CiCXCR3.2, TnCXCR3.1, and cession number GSE40221 (22). The cDNA sequence of ayu CXCR3.2 (P. TnCXCR3.2 were washed with FACS buffer and resuspended at a con- altivelis CXCR3.1 [PaCXCR3.2]) was obtained from the transcriptome centration of 2 3 107 cells per milliliter. Cells (2 3 106) were incubated analysis of mixed ayu tissues, and transcriptomic data were deposited in with 2 ml each of anti-CXCR3.1 and anti-CXCR3.2 IgG for 0.5 h at 4˚C. the Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo/) under After washing, the cells were incubated with the secondary Ab, PerCP- accession number GSE73321. The cDNA sequences of PaCXCR3.1 and conjugated donkey anti-mouse IgG (Jackson ImmunoResearch Laborato- PaCXCR3.2 were deposited in GenBank under accession numbers ries, West Grove, PA) for 0.5 h at 4˚C. For double staining, anti-PaCXCR3. JP725619 and KY081643, respectively. The cDNA sequence of grass carp 1 IgG and anti-PaCXCR3.2 IgG were labeled with FITC and R-PE, re- CXCR3.1 (C. idella CXCR3.1 [CiCXCR3.1]) was deposited in GenBank spectively, using the standard Ab Labeling Kit (Thermo Fisher Scientific), under accession number AY744916. The partial cDNA sequence of grass according to the manufacturer’s instructions. Cells (2 3 106) were incu- carp CXCR3.2 (C. idella CXCR3.2 [CiCXCR3.2]) was obtained from bated with 2 ml each labeled anti-CXCR3.1 IgG and anti-CXCR3.2 IgG for the genome data of grass carp (https://www.ncbi.nlm.nih.gov/bioproject/? 0.5 h at 4˚C. The stained cells were then washed and analyzed using the term=PRJEB5920) under accession PRJEB5920. The full-length cDNA Gallios Flow Cytometer (Beckman Coulter, Miami, FL). At least 10,000 sequence of CiCXCR3.2 was subsequently determined using the RACE events were collected from the macrophage gate. Data were analyzed using method. The cDNA sequence of CiCXCR3.2 was deposited in GenBank Kaluza software (Beckman Coulter). under accession number KY081642. The cDNA sequences of spotted CXCR3.1+ and CXCR3.2+ macrophages were also analyzed and sorted green pufferfish CXCR3.1 (T. nigroviridis CXCR3.1 [TnCXCR3.1]) and T. using flow cytometry. Flow cytometry sorting of CXCR3.1+ and CXCR3.2+ nigroviridis CXCR3.2 (TnCXCR3.2) were obtained from the genome data macrophages was performed by gating on the forward and side scatter analysis of spotted green pufferfish. cDNA sequences of ayu IL-12p40a (P. signals, as well as fluorescence signals, using an MoFlo XDP Cell Sorter altivelis IL-12p40a [PaIL-12p40a]), IL-12p40c (P. altivelis IL-12p40c (Beckman Coulter). At least 10,000 events were collected from the macro- [PaIL-12p40c]), STAT1 (P. altivelis STAT1 [PaSTAT1], accession num- phage gate. The sorted macrophages were incubated in complete medium at ber JP742610), STAT3 (P. altivelis PaSTAT3 [PaSTAT3], accession num- 24˚C with 5% CO2 for subsequent use. ber JP722452), CXCL9–11l1 (P. altivelis CXCL9–11l1 [PaCXCL9–11l1]), To remove surface-bound anti-CXCR3.1 and anti-CXCR3.2 IgG, the CXCL9–11l2 (P. altivelis CXCL9–11l2 [PaCXCL9–11l2]), and CXCL9– cells were incubated with a mixture of 200 mM acetic acid and 150 mM 11l3 (P. altivelis CXCL9–11l3 [PaCXCL9–11l3]) genes were obtained NaCl (pH 2.8) for 5 min. Acid stripping was terminated by transferring cells from the transcriptome analysis of ayu monocytes/macrophages, and to prewarmed complete medium. 4694 TWO TELEOST CXCR3s CONTRIBUTE TO MACROPHAGE POLARIZATION

Ab blockade use in experiments. Mouse isotype IgG (IsoIgG, 250 mg/ml) was used as a control. The peptides corresponding to PaCXCR3.1 (aa 1–59) and PaCXCR3.2 (aa 1–44) were synthesized for polyclonal Ab production (GL Biochem). The Real-time quantitative PCR speciﬁcity of polyclonal anti-PaCXCR3.1 and anti-PaCXCR3.2 IgG was validated by Western blot (Supplemental Fig. 1B). For in vitro Ab blockade We detected the expression levels of PaCXCR3.1 and PaCXCR3.2 mRNA in macrophages, sorted macrophages were incubated with anti-PaCXCR3.1 in ayu tissues after L. anguillarum infection. Ayu in the “infected” group IgG (250 mg/ml) and anti-PaCXCR3.2 IgG (250 mg/ml) for 30 min prior to were injected i.p. with L. anguillarum (1.2 3 104 CFU per ﬁsh in 100 mlof Downloaded from http://www.jimmunol.org/ by guest on September 29, 2021

FIGURE 1. Multiple alignment of the amino acid sequences of PaCXCR3s and other related CXCR3s. Similar residues are shaded in gray, identical residues are shaded in black, and alignment gaps are represented by hyphens (-). The threshold for shading is .60%. The GenBank accession numbers for the amino acid sequences are listed in Table II. EL1–3, extracellular loop 1–3; IL1–3, intracellular loop 1–3; TM1–7, transmembrane 1–7. The Journal of Immunology 4695 Downloaded from http://www.jimmunol.org/

FIGURE 2. Phylogenetic analysis of two teleost CXCR3 genes and their expression in macrophages. (A) Phylogenetic (neighbor-joining) analysis of the complete amino acid sequence of teleost and mammalian CXCR3s using the MEGA5.0 program. Node values represent the percentage bootstrap conﬁ- dence derived from 1000 replicates. The source sequences are listed in Table II. (B) mRNA expression of PaCXCR3s in ayu macrophages at different time by guest on September 29, 2021 points following E. coli infection at an MOI of 2 (n = 6). (C) CXCR3.1 and CXCR3.2 expression on the surface of ayu macrophages following E. coli infection at an MOI of 2. (D) The mRNA levels of PaCXCR3.1 and PaCXCR3.2 in cell populations separated using anti-PaCXCR3.1 and anti-PaCXCR3.2 IgG. n = 5. Null means no expression. Control, macrophages stained with Ab without infection; Infection, macrophages stained with Ab postinfection; IsoIgG, macrophages stained with isotype IgG.

PBS), whereas those in the control group were injected with PBS. Liver, expression of the target genes was normalized against that of 18S rRNA spleen, gill, kidney, head kidney, blood, and brain samples were collected using the 22DD cycle threshold method. at 4, 8, 12, and 24 h postinfection, snap-frozen immediately in liquid ni- Western blotting trogen, and preserved at 280˚C until further use. Total RNA was extracted from fish tissues and macrophages using The cells were pelleted and lysed in a buffer (20 mM HEPES, 1.5 mM RNAiso (Takara, Dalian, China). After treatment with DNase I, first-strand MgCl2, 0.2 mM EDTA, 100 mM NaCl, 0.2 mM DTT, 0.5 mM sodium cDNA was synthesized using AMV reverse transcriptase (Takara). Real- orthovanadate, and 0.4 mM PMSF [pH 7.4]) containing phosphatase in- time quantitative PCR (RT-qPCR) was performed on an ABI StepOne hibitor (Phosphatase Inhibitor Cocktail; Sigma, St. Louis, MO). The pro- Real-Time PCR System (Applied Biosystems, Foster City, CA) using SYBR tein concentration was measured in each soluble fraction using the Premix Ex Taq II (Takara). The primer sequences for the target genes are Bradford method. For Western blot analysis, the proteins were resolved listed in Supplemental Table I. Amplifications were carried out in a 25-ml using SDS-PAGE, transferred to membrane, incubated with specific Abs, reaction volume containing the sample, primers, and SYBR Premix Ex Taq and visualized using an ECL Western blotting detection system. Ayu TNF II. The reaction mixture was incubated for 300 s at 95˚C, followed by 40 (P. altivelis TNF [PaTNF]) Abs were prepared using a recombinant PaTNF amplification cycles of 30 s at 95˚C, 30 s at 60˚C, and 30 s at 72˚C. mRNA protein in our laboratory (23). Ayu IL-10 (P. altivelis IL-10 [PaIL-10]) Abs

Table I. mRNA expression of PaCXCR3s in ayu tissues after L. anguillarum challenge

Protein Time (h) Liver Spleen Gill Kidney Head Kidney Blood Brain CXCR3.1 4 — ↑↑———— 8 ↑↑↑↑ ——— 12 — ↑↑↑ ——— 24 — — ↑↑ ——— CXCR3.2 4 — ↑↑— ↑ —— 8 ↑↑↑↑ ↑ —— 12 ↑↑↑↑ ↑ — ↑ 24 ↑↑↑↑ ↑ — ↑ ↑, upregulation; ↓, downregulation; ―, unchanged. 4696 TWO TELEOST CXCR3s CONTRIBUTE TO MACROPHAGE POLARIZATION were prepared using a peptide derived from PaIL-10 protein (Lys137– of killed bacteria in the “killing” group plate compared with the “uptake” Phe168; GL Biochem). PaSTAT1 (aa 684–699), Tyr694-phosphorylated group plate, as described previously (37). PaSTAT1 (aa 684–699), PaSTAT3 (aa 696–712), and Tyr708-phosphorylated PaSTAT3 (aa 696–712) peptides were synthesized to generate the RNA interference respective mAbs (GL Biochem). The specificity of anti–PaIL-10, anti- Stealth RNA interference (RNAi) duplexes against PaSTAT1 and PaSTAT3 PaSTAT1, anti–p-PaSTAT1, anti-PaSTAT3, and anti–p-PaSTAT3 IgG was and a stealth RNAi negative control duplex were purchased from Invitrogen. validated using Western blot (Supplemental Fig. 1C). The mature peptide The sequences of the stealth RNAi duplexes used for the knockdown of (aa 23–182) of PaIL-10, fragment (aa 494–745) of PaSTAT1, and fragment PaSTAT1 and PaSTAT3 were 59-CAAGAUGAUCAUGACUUCAAG- (aa 535–767) of PaSTAT3 were amplified. The primer sequences for the CUAA-39,and59-CCAACAAAGUCAGAUUACUGGUAAA-39, respectively. target genes are listed in Supplemental Table I. The amplicons with the A scrambled small interfering RNA (siRNA) (59-GAGACACAGGCUCGU- expected size were inserted into the pET-28a (+) vector. The recombinant UAAUAGGAGU-39) was used as the negative control. Transfection of cells E. coli plasmids were constructed and transformed into BL21 (DE3) for with siRNA was performed using Lipofectamine 2000 transfection reagent overexpression. PaIL-10, PaSTAT1, and PaSTAT3 were expressed as an (Invitrogen), according to the manufacturer’s protocol. After incubation for inclusion body and were purified using an Ni-NTA column (QIAGEN, 5.5 h, the cell culture medium was changed to complete medium, and cells Shanghai, China), according to the manufacturer’s instructions. The were cultured for 48 h before collection for subsequent use. RT-qPCR was specificity of anti-PaSTAT1 and anti-PaSTAT3 IgG was tested in recombinant performed to confirm the knockdown. proteins and native proteins from ayu macrophages by Western blot analysis, as described above. The specificity of anti–PaIL-10 IgG was tested in Chemotaxis recombinant PaIL-10 proteins and native PaIL-10 from ayu serum by Western blot analysis, as described above. Sorted macrophages (5 3 105 cells per chamber) were seeded in the upper chambers of Transwell plates (5-mm pore size), and chemokine- Arginase and inducible NO synthase activity assay containing medium was added to the bottom chambers. After 4 h of

incubation at 24˚C, the macrophages that had migrated through the Downloaded from Arginase activity was measured in cell lysates, as described previously (21), Transwell membrane to the bottom chambers were collected and counted and was expressed as micromoles of urea per milligram of protein. For the measurement of inducible NO synthase (iNOS) activity, the by flow cytometry. supernatants from macrophage cell cultures were collected, and the Griess Reactive oxygen species detection assay was performed. Eighty-microliter aliquots of the supernatants or sodium nitrate standards were combined with equal volumes of fresh Griess Macrophages were loaded with dihydrorhodamine 123 (5 mg/ml; Sigma) reagent (1% sulfanilamide, 0.1% naphthylethylenediamine dihydro- for 5 min at 24˚C. Cells were then washed and infected with E. coli or chloride, and 2.5% phosphoric acid). The samples were incubated at 24˚C L. anguillarum for 30 min at 24˚C. The cells in the control group were treated http://www.jimmunol.org/ for 10 min, and absorbance was measured at 540 nm. iNOS activity was with PBS. Reactive oxygen species (ROS) production was quantified via flow expressed as micromoles of nitrite per milligram of protein. cytometry by measuring intracellular rhodamine (Gallios Flow Cytometer; Beckman Coulter). ROS content was expressed as the mean fluorescence Phagocytosis and bacterial killing intensity of treated cells. Phagocytosis and bacterial killing by macrophages were measured as Recombinant protein preparation previously described (26). Briefly, E. coli (DH5a) were labeled with FITC (Sigma), henceforth called FITC-DH5a. Macrophages were treated with Recombinant PaCXCL9–11l1, PaCXCL9–11l2, and PaCXCL9–11l3 pro- FITC-DH5a at an MOI of 10. Cells were incubated for 30 min and ana- teins (identified as CXCL9–11-like proteins in ayu) were produced using lyzed by flow cytometry using a Gallios flow cytometer (Beckman Coul- CHO cells, as described previously (38). Briefly, PaCXCL9–11l1,

ter). The results were expressed as the relative mean fluorescence intensity PaCXCL9–11l2, and PaCXCL9–11l3 cDNA fragments were amplified by guest on September 29, 2021 in comparison with the control. Bacterial killing assays were quantified by using the primers listed in Supplemental Table I and cloned into the EcoRI measuring CFU. Macrophages were infected with E. coli at an MOI of 10. and KpnI sites of pcDL-SRa296 (provided by Biovector Science Lab, Bacterial phagocytosis was allowed to proceed for 30 min, and non- Beijing, China) to generate the corresponding expression plasmids. CHO internalized E. coli were removed by washing with sterile PBS. One set of cells were transfected with the expression plasmids and maintained in samples (the uptake group) was lysed in 1% Triton X-100 solution and Eagle’s MEM supplemented with 10% FBS. Recombinant proteins were plated onto solid Luria-Bertani agar medium to provide bacterial uptake purified from the cell culture supernatants and passed through an values. The cells were incubated for 1.5 h to allow bacterial killing to endotoxin-removal column (Pierce, Rockford, IL). The resulting endotoxin occur. Cell lysate–bacterial samples were collected, lysed in 1% Triton levels in the recombinant proteins were ,0.1 EU/mg. Concentrations of X-100 solution, and then plated onto Luria-Bertani agar medium. The purified proteins were measured using the Bradford method. Macrophages percentage of bacteria killing was determined by calculating the proportions were treated with the recombinant proteins at the indicated concentrations.

Table II. CXCR3 sequences used in this study

Species

Accession Number/Ensembl Identifier Latin Name Common Name Protein NM_001504 Homo sapiens Human CXCR3 XM_005593931 Macaca fascicularis Crab-eating macaque CXCR3 NM_009910 Mus musculus Mouse CXCR3 NM_001011673 Bos taurus Cattle CXCR3 XM_008123303 Anolis carolinensis Green anole lizard CXCR3 XM_005173488 Danio rerio Zebrafish CXCR3.1 NM_001124423 Oncorhynchus mykiss Rainbow trout CXCR3.1 AY744916 C. idella Grass carp CXCR3.1 GSTENG00015608001 T. nigroviridis Spotted green pufferfish CXCR3.1 ENSORLG00000013488 Oryzias latipes Japanese ricefish CXCR3.1 JP725619 P. altivelis Ayu CXCR3.1 NM_001007314 D. rerio Zebrafish CXCR3.2 NM_001124625 O. mykiss Rainbow trout CXCR3.2 KY081642 C. idella Grass carp CXCR3.2 GSTENG00015610001 T. nigroviridis Spotted green pufferfish CXCR3.2 ENSORLG00000013459 O. latipes Japanese ricefish CXCR3.2 KY081643 P. altivelis Ayu CXCR3.2 The Journal of Immunology 4697

Survival assay and bacterial burden Ayu were infused with 5 3 106 PaCXCR3.1+ and PaCXCR3.2+ macrophages immediately postinfection with L. anguillarum (1.2 3 104 CFU per fish in 100 ml of PBS). Morbidity was monitored for 96 h after challenge, and the results were recorded every 12 h. Peritoneal lavage fluids were harvested aseptically at 24 h postinfection, plated onto thiosulfate–citrate– bile salts agar plates, and incubated for 12 h at 28˚C. CFU were calculated by multiplying the number of colonies on the plate by the dilution factor. Statistical analysis Data are presented as mean 6 SEM. The number of biological repeats (n) is indicated. The mice used in the experiments were randomly chosen from our in-house colonies or suppliers. Animal experiments were performed by an observer blinded to the experimental conditions. The survival curves were analyzed using the Kaplan–Meier method. All other data were analyzed by one-way ANOVA. When variances were significantly different (p , 0.05), logarithmic transformation was used to stabilize the variance. For data that did not follow a normal distribution, statistical significance was evaluated using the Mann–Whitney U test (two-tailed). *, **, and *** represent p values ,0.05, 0.01, and 0.001, respectively. Downloaded from Results Characterization and expression of two teleost CXCR3 genes in macrophages To elucidate the expression and function of CXCR3 proteins in teleosts, we first obtained the PaCXCR3.1 and PaCXCR3.2 gene

sequences from the transcriptome data (22). Multiple sequence http://www.jimmunol.org/ alignments with other known CXCR3 amino acid sequences revealed that the N terminus, seven transmembrane domains, three extracellular loops, three intracellular loops, and the C terminus were conserved between teleosts and mammals (Fig. 1). Sequence comparisons revealed that PaCXCR3.1 and PaCXCR3.2 shared 31 and 36% amino acid identity, respectively, with their mouse counterparts. Phylogenetic analysis indicated that ﬁsh CXCR3.1 genes were grouped together to form a distinct cluster that differed

from the ﬁsh CXCR3.2 and mammalian CXCR3 clusters by guest on September 29, 2021 (Fig. 2A). Fish macrophages express high levels of CXCR3.1 and

FIGURE 3. Polarization and function of CXCR3.1+ and CXCR3.2+ macrophages. (A) Protocol for the preparation of resting PaCXCR3.1+ and PaCXCR3.2+ macrophages. (B–E) mRNA levels of cytokines in PaCXCR3.1+ and PaCXCR3.2+ macrophages. mRNA levels of PaIL-1b (B), PaTNF (C), PaIL-10 (D), and PaTGF-b (E) were detected in macrophages after E. coli treatment at an MOI of 2. iNOS (F) and arginase (G) activity in PaCXCR3.1+ and PaCXCR3.2+ macrophages. (H) PaTNF and PaIL-10 protein levels in PaCXCR3.1+ and PaCXCR3.2+ macrophages. Bacterial killing (I) and phagocytosis (J) in PaCXCR3.1+ and PaCXCR3.2+ macrophages. n =5.*p , 0.05, **p , 0.01, ***p , 0.001. PCR3.1, PaCXCR3.1; PCR3.2, PaCXCR3.2.

Receptor binding assay Competition binding assays were performed with CXCR3.1–HEK-293T or CXCR3.2–HEK-293T cells using [125I]-PaCXCL9–11l1, [125I]-PaCXCL9– 11l2, and [125I]-PaCXCL9–11l3 labeled by the Bolton–Hunter procedure (2). The maximal binding of labeled [125I]-PaCXCL9–11l1, [125I]-PaCXCL9– 11l2, and [125I]-PaCXCL9–11l3 was determined by measuring binding 3 6 FIGURE 4. Macrophage polarization upon stimulation with E. coli at at saturating concentrations. A total of 5 10 cells was incubated with + + A a labeled chemokine and increasing concentrations of polyclonal anti- different MOIs. Proportion of CXCR3.1 and CXCR3.2 macrophages ( B + C PaCXCR3.1 or anti-PaCXCR3.2 IgG (0.3–3000 nM) in 200 mlofRPMI and ), PaTNF expression in CXCR3.1 macrophages ( ), and PaIL-10 + 1640 medium containing HEPES (25 mM, pH 7.4), BSA (10 mg/ml), and expression in CXCR3.2 macrophages (D) after stimulation with E. coli at sodium azide (0.1%). The incubations were performed for 30 min, and cell the indicated MOIs. The total proportions represented PaCSF1R+ cells. n = radioactivity was measured by gamma counting. 5. **p , 0.01, ***p , 0.001. 4698 TWO TELEOST CXCR3s CONTRIBUTE TO MACROPHAGE POLARIZATION

PaCXCR3.1+ and PaCXCR3.2+ macrophages from ayu macrophage culture after in vitro E. coli infection at an MOI of 2. Simply, PaCXCR3.1+ PaCXCR3.22 macrophages were deﬁned as PaCXCR3.1+ macrophages, whereas PaCXCR3.2+ PaCXCR3.12 macrophages were deﬁned as PaCXCR3.2+ macrophages. We also obtained resting PaCXCR3.1+ and PaCXCR3.2+ macrophages by blocking PaCXCR3s with the respective IgG (Fig. 3A). Acid wash was used to remove the Abs bound to the cell surface. Because CXCR3 affects macrophage polarization in mammals (7), we measured the marker of macrophage polarization in PaCXCR3.1+ and PaCXCR3.2+ macrophages. mRNA levels of P. altivelis IL-1b (PaIL-1b)andPaTNF were upregulated in E. coli–treated PaCXCR3.1+ macrophages compared with the PBS control, whereas their levels were downregulated in E. coli–treated PaCXCR3.2+ macrophages compared with the PBS control (Fig. 3B, 3C). mRNA levels of PaIL-10 and PaTGF-b were upregulated in E. coli–treated PaCXCR3.2+ macrophages compared with the PBS control but were downregulated in E. coli–treated PaCXCR3.1+ macrophages compared with the PBS

control (Fig. 3D, 3E). Moreover, iNOS activity was upregulated in Downloaded from E. coli–treated PaCXCR3.1+ macrophages compared with the PBS control, whereas iNOS activity was downregulated in E. coli–treated PaCXCR3.2+ macrophages compared with the PBS control (Fig. 3F). Arginase activity was upregulated in E. coli–treated FIGURE 5. Expression of PaIL-12p40 subunit genes in ayu. (A) Phy- PaCXCR3.2+ macrophages compared with the PBS control but + logenetic (neighbor-joining) analysis of amino acid sequences of IL-12p40 was downregulated in E. coli–treated PaCXCR3.1 macrophages http://www.jimmunol.org/ using MEGA5.0, rooted with the human cytokine-like factor 1 (CLF1) compared with the PBS control (Fig. 3G). The specificity of sequence (NM_004750). Node values represent the percentage bootstrap anti–PaIL-10 IgG was validated using Western blot. PaTNF and confidence derived from 1000 replicates. The accession numbers of the IL-12p40 genes are AJ621425, AJ628699, and AJ628700 for carp (Cyprinus carpio) IL-12p40a, IL-12p40b, and IL-12p40c, respectively; XM_004073526, XM_004076147, and XM_004072267 for Japanese ricefish (Oryzias latipes) IL-12p40a, IL-12p40b, and IL-12p40c, respectively; JP722560 and KY081641 for ayu (P. altivelis) IL-12p40a and IL- 12p40c, respectively; BN000860 for zebrafish (Danio rerio) IL-12p40; NM_213571 for chicken (Gallus gallus) IL-12p40; AY008847 for human by guest on September 29, 2021 (Homo sapiens) IL-12p40; NM_001303244 for mouse (Mus musculus) IL- 12p40; and AF059293 for human (H. sapiens) CLF1. (B) mRNA expression of ayu IL-12p40a in CXCR3.1+ and CXCR3.2+ macrophages. (C) mRNA expression of ayu IL-12p40c in CXCR3.1+ and CXCR3.2+ macrophages. n = 5. ***p , 0.001.

CXCR3.2 genes (8, 39); thus, we further investigated PaCXCR3.1 and PaCXCR3.2 expression in ayu macrophages after E. coli challenge (22). Although PaCXCR3.2 mRNA was upregulated in macrophages after E. coli infection, there was no change in the PaCXCR3.1 mRNA levels (Fig. 2B). Ayu macrophages were identified using anti- PaCSF1R IgG, as described previously (35) (Fig. 2C). The specificity of monoclonal anti-PaCXCR3.1 and anti-PaCXCR3.2 IgG was validated using flow cytometry. The surface expression of PaCXCR3.1 was not altered after E. coli treatment, whereas the surface expression of PaCXCR3.2 was increased (Fig. 2C). The mRNA analysis also confirmed the specificity of the anti-PaCXCR3.1 and anti-PaCXCR3.2 IgGusedinflowcytometry(Fig.2D). We further measured PaCXCR3.1 and PaCXCR3.2 mRNA expression in tissues and macrophages by RT-qPCR. PaCXCR3.1 and PaCXCR3.2 mRNA were expressed in all tissues tested. PaCXCR3.1 mRNA was upregulated in the liver, spleen, gill, and kidney after E. coli infection. Moreover, PaCXCR3.2 mRNA was upregulated in all tissues tested, with the exception of blood (Tables I, II). FIGURE 6. Polarization of CXCR3.1+ and CXCR3.2+ macrophages in + CXCR3.1+ and CXCR3.2+ macrophages show functional grass carp and spotted green pufferfish. (A) Polarization in CiCXCR3.1 + polarization and CiCXCR3.2 grass carp macrophages, indicated by mRNA expression levels of PaIL-1b, PaTNF, PaIL-10, and PaTGF-b, as well as iNOS and Having demonstrated that PaCXCR3.1 and PaCXCR3.2 are arginase activities. (B) Polarization in TnCXCR3.1+ and TnCXCR3.2+ surface-expressed in ayu macrophages, we asked whether these macrophages in spotted green pufferfish. n = 6. CR3.1, CXCR3.1; CR3.2, proteins exhibit subfunctionalization in macrophages. We isolated CXCR3.2. *p , 0.05, ***p , 0.001. The Journal of Immunology 4699

Furthermore, mRNA levels of PaTNF and PaIL-10 in PaCXCR3.1+ and PaCXCR3.2+ macrophages, respectively, were upregulated as the MOI increased from 0.625 to 2.5 (Fig. 4C, 4D). Because IL-12 is expressed in CXCR3.2 (CXCR3a)+ phagocytes in Japanese ricefish (10), we further investigated the expression of IL-12 genes in PaCXCR3.1+ and PaCXCR3.2+ macrophages in ayu. We first obtained the IL-12p40a and IL-12p40c gene sequences from the transcriptome data (22). PaIL-12p40a and PaIL-12p40c gene sequences were identified in ayu. Sequence comparisons revealed that PaIL-12p40a and PaIL-12p40c shared 42.2 and 38.4% amino acid identity, respectively, with their Jap- anese ricefish counterparts. Phylogenetic tree analysis grouped teleost IL-12p40a together in a cluster that was distinct from the mammal IL-12p40, teleost IL-12p40b, and IL-12p40c cluster (Fig. 5A). PaIL-12p40a was most closely related to IL-12p40a from Japanese ricefish (Fig. 5A). Phylogenetic tree analysis grouped teleost IL-12p40c together in a cluster that was distinct from the mammal IL-12p40, teleost IL-12p40a, and IL-12p40b

cluster (Fig. 5A). PaIL-12p40c was most closely related to Downloaded from IL-12p40c from Japanese riceﬁsh (Fig. 5A). PaIL-12p40a and PaIL-12p40c mRNA were expressed in PaCXCR3.1+ and PaCXCR3.2+ macrophages in ayu, and their expression levels were higher in PaCXCR3.1+ macrophages than in PaCXCR3.2+ macrophages (Fig. 5B).

Because CXCR3.1 and CXCR3.2 exist in various teleost fish, http://www.jimmunol.org/ we further asked whether the polarization of CXCR3.1+ and CXCR3.2+ macrophages takes place in other teleost fish. We investigated polarization in CXCR3.1+ and CXCR3.2+ macrophages in grass carp and spotted green pufferfish. The specificity of anti- CiCXCR3.1 and anti-CiCXCR3.2 IgG was validated using flow cytometry. CiCXCR3.1+ macrophages showed a high expression of IL-1b and TNF and increased iNOS activity, whereas CiCXCR3.2+ macrophages showed a high expression of IL-10 and TGF-b and increased arginase activity (Fig. 6A). Furthermore, by guest on September 29, 2021 + + FIGURE 7. L. anguillarum did not induce the expression of CXCR3.2 TnCXCR3.1 and TnCXCR3.2 macrophages also showed the in ayu macrophages. (A) Expression of CXCR3.1 and CXCR3.2 after M1 and M2 phenotype, respectively (Fig. 6B). The specificity of L. anguillarum treatment at an MOI of 2 (n = 6). (B) iNOS and arginase anti-TnCXCR3.1 and anti-TnCXCR3.2 IgG was validated using activities in different macrophages (n = 5). (C) mRNA levels of cytokines indicative of established macrophage M1 and M2 phenotypes, including PaIL-1b, PaTNF, PaIL-10, and PaTGF-b in L. anguillarum–treated PaCXCR3.1+ macrophages. PaCXCR3.1+ macrophages were isolated and treated with PBS, E. coli,orL. anguillarum (n = 6). **p , 0.01, ***p , 0.001. Control, macrophages stained with anti-PaCXCR3s IgG without infection; IsoIgG, macrophages stained with isotype IgG; L. anguillarum, macrophages stained with anti-PaCXCR3s IgG after L. anguillarum infection.

PaIL-10 protein expression levels were detected by Western blotting analysis in PaCXCR3.1+ and PaCXCR3.2+ macrophages, respectively (Fig. 3H). Furthermore, E. coli–treated PaCXCR3.1+ and PaCXCR3.2+ macrophages showed higher phagocytic activity and higher bacterial killing capability than did the PBS control (Fig. 3I, 3J). Bacterial killing was higher in CXCR3.1+ macrophages relative to CXCR3.2+ macrophages (Fig. 3I). Taken together, these data suggest that PaCXCR3.1+ and PaCXCR3.2+ macrophages exhibit the M1 and M2 phenotype, respectively. The observations described above revealed that E. coli infection FIGURE 8. ROS mediates the expression of PaCXCR3.1 or PaCXCR3.2. affected macrophage polarization. We further measured macro- (A) ROS production in E. coli–orL. anguillarum–stimulated macrophage polarization balance after stimulation with E. coli at dif- phages, as assessed by dihydrorhodamine 123 staining and ﬂow + ferent MOIs. PaCSF1R cells were isolated to measure the cytometry and presented as a time-response curve (n =5).(B and C) + + number of PaCXCR3.1 and PaCXCR3.2 macrophages. We ob- Levels of CXCR3.1 and CXCR3.2 mRNA in macrophages after E. coli served an increase in the number of PaCXCR3.2+ macrophages or L. anguillarum treatment, with or without NAC, an antagonist of ROS with an increase in the MOI from 0.625 to 2.5 (Fig. 4A, 4B). (n =6).**p , 0.01, ***p , 0.001. 4700 TWO TELEOST CXCR3s CONTRIBUTE TO MACROPHAGE POLARIZATION

PaCXCR3 ligands mediate macrophage polarization PaCXCR3.1+ M1 and PaCXCR3.2+ M2 macrophages were found to coexist in the macrophage cultures after E. coli infection. Thus, we speculated that PaCXCR3.1 and PaCXCR3.2 act as markers of macrophage polarization, as well as mediate the polarization process. The specificity of polyclonal anti-PaCXCR3.1 and anti- PaCXCR3.2 IgG was validated by Western blot. We first obtained resting PaCXCR3.1+ and PaCXCR3.2+ macrophages by blocking PaCXCR3s with the respective Abs. Acid wash was used to remove the Abs bound to the cell surface. Ab blockade was further used to investigate PaCXCR3.1 and PaCXCR3.2 function in macrophage polarization induced by E. coli. iNOS activity in PaCXCR3.1+ macrophages was downregulated after anti-PaCXCR3.1 IgG incubation comparedwiththoseincubated with isotype IgG (Fig. 9A). Arginase activity in PaCXCR3.2+ macrophages was downregulated after anti-PaCXCR3.2 IgG treatment compared with after isotype IgG treatment (Fig. 9A). mRNA levels of PaIL-1b and PaTNF in PaCXCR3.1+ macrophages Downloaded from were downregulated after anti-PaCXCR3.1 IgG incubation compared with after isotype IgG incubation (Fig. 9B). mRNA levels of PaIL-10 and PaTGF-b in PaCXCR3.2+ macrophages were downregulated after anti-PaCXCR3.2 IgG treatment compared with after isotype IgG treatment (Fig. 9B). We further investigated the ligands for PaCXCR3.1 and http://www.jimmunol.org/ PaCXCR3.2 in ayu. Three CXCL9–11-like proteins were identi- FIGURE 9. Effect of PaCXCR3 blockade on macrophage polarization. fied and named as PaCXCL9–11l1 (accession number KJ130413), (A) iNOS and arginase activities in the differently treated macrophages PaCXCL9–11l2 (accession number KU362928), and PaCXCL9– (n =5).(B) mRNA levels of cytokines indicative of the established M1 11l3 (accession number KU362929) using phylogenetic analysis and M2 macrophage phenotypes, including PaIL-1b,PaTNF,PaIL-10, of teleost and mammalian chemokine genes (Fig. 10). Sequence and PaTGF-b,inE. coli–treated PaCXCR3.1+ and PaCXCR3.2+ mac- + + comparisons revealed that PaCXCL9–11l1, PaCXCL9–11l2, and rophages (n = 6). Resting PaCXCR3.1 and PaCXCR3.2 macrophages PaCXCL9–11l3 shared 27.9, 21.1, and 38.1% amino acid identity, were obtained by blocking PaCXCR3s with the respective Abs. Acid respectively, with zebrafish CXCL11-1 (XM_001339271). Sequence wash was used to remove the Abs bound to the cell surface. Ab blockade was further used to investigate PaCXCR3.1 and PaCXCR3.2 function in comparisons also revealed that PaCXCL9–11l1, PaCXCL9–11l2, by guest on September 29, 2021 macrophage polarization induced by E. coli. IsoIgG, isotype IgG; R3.1 and PaCXCL9–11l3 shared 36.2, 21.1, and 23.3% amino acid IgG, anti-PaCXCR3.1 IgG; R3.2 IgG, anti-PaCXCR3.2 IgG. **p , 0.01, identity, respectively, with zebrafish CXCL11-2 (NM_001126413). ***p , 0.001. Interestingly, although these ligands in teleost fish are CXCL9–11- like proteins, they are not identical to the CXCL9-, CXCL10-, or CXCL11-like proteins. flow cytometry. These results suggest that functional polarization We also confirm the effect of Ab blockade by polyclonal anti- of CXCR3.1+ and CXCR3.2+ macrophages is a common phe- PaCXCR3.1 and anti-PaCXCR3.2 IgG in macrophages (Fig. 11). nomenon in teleosts. Polyclonal anti-PaCXCR3.1 IgG potently displaced 125I-labeled 125 ROS mediate PaCXCR3.2 induction after E. coli but not L. PaCXCL9–11l2 and I-labeled PaCXCL9–11l3 with IC50 val- anguillarum infection ues of 55.03 6 5.05 and 12.44 6 2.23 nm, respectively (Fig. 11A, 125 Next, we determined the mechanisms contributing to the regulation 11B). Anti-PaCXCR3.1 IgG inhibited the binding activity of I- + of PaCXCR3.1 and PaCXCR3.2 expression postinfection. We first labeled PaCXCL9–11l2 by 64% in CXCR3.1 macrophages 125 investigated the expression of PaCXCR3.1 and PaCXCR3.2 post- (Fig. 11A). Polyclonal anti-PaCXCR3.2 IgG potently displaced infection with L. anguillarum, a major ayu pathogen. L. anguillarum I-labeled PaCXCL9–11l1 and PaCXCL9–11l2 with IC50 values of did not induce PaCXCR3.2 expression in macrophages (Fig. 7A), 42.22 6 3.89 and 19.33 6 2.37 nm, respectively (Fig. 11C, 11D). 125 because we could not isolate PaCXCR3.2+ macrophages postinfec- Anti-PaCXCR3.2 IgG inhibited the binding activity of I-labeled + tion in resting macrophages. PaCXCR3.1+ macrophages showed the PaCXCL9–11l2 by 58% in CXCR3.2 macrophages (Fig. 11D). M1 phenotype after L. anguillarum infection (Fig. 7B, 7C). We could not detect the binding activity of PaCXCL9–11l1 in ROS play an important role in regulating cellular processes PaCXCR3.1+ macrophages or PaCXCL9–11l3 in PaCXCR3.2+ in macrophages and affect macrophage polarization (40, 41). macrophages. Furthermore, we used chemotaxis assays to confirm Therefore, we investigated whether ROS are implicated in the the specific ligands for PaCXCR3.1 and PaCXCR3.2. PaCXCL9– expression of PaCXCR3.1 and PaCXCR3.2. Infection with E. coli, 11l2 and PaCXCL9–11l3, but not PaCXCL9–11l1, induced the but not with L. anguillarum, enhanced the production of ROS migration of PaCXCR3.1+ macrophages (Fig. 12A). Maximal cell in ayu macrophages. We detected the ROS content in macro- migration was observed at 30 and 10 nM of PaCXCL9–11l2 and phages by flow cytometry. The level of ROS at 60 min post- PaCXCL9–11l3, respectively. PaCXCL9–11l1 and PaCXCL9– infection was 13.3-fold higher than the baseline level 11l2, but not PaCXCL9–11l3, induced the migration of PaCXCR3.2+ (Fig. 8A). NAC, a ROS antagonist, prevented the upregulation of macrophages (Fig. 12B). Maximal cell migration was observed PaCXCR3.2 mRNA postinfection (Fig. 8B, 8C). These results sug- at 30 nM for PaCXCRL9–11l1 and PaCXCL9–11l2. Anti- gest that ROS mediate PaCXCR3.2 expression in macrophages after PaCXCR3.1 IgG blockade inhibited the chemotaxis ability of E. coli infection. resting PaCXCR3.1+ macrophages induced by PaCXCL9–11l2 The Journal of Immunology 4701

FIGURE 10. Phylogenetic analysis of CXCL9–11-like genes. Phy- logenetic (neighbor-joining) analysis of the complete amino acid sequence of teleost and mammalian CXC chemokine genes using the MEGA5.0 program. Node values represent the percentage bootstrap conﬁdence derived

from 1000 replicates. PaCXCL9–11l1, Downloaded from PaCXCL9–11l2, and PaCXCL9– 11l3 are marked with a black dot. The source sequences are listed in Supplemental Table II. http://www.jimmunol.org/ by guest on September 29, 2021 and PaCXCL9–11l3 (Fig. 12C). Anti-PaCXCR3.2 IgG blockade that was distinct from the mammal STAT3 and STAT1 cluster inhibited the chemotaxis ability of resting PaCXCR3.2+ macro- (Fig. 13). We further investigated the phosphorylation patterns phages induced by PaCXCL9–11l1 and PaCXCL9–11l2 (Fig. 12D). ofPaSTAT1andPaSTAT3inCXCR3.1+ and CXCR3.2+ mac- Recombinant PaCXCL9–11l1, PaCXCL9–11l2, and PaCXCL9– rophages. The specificity of anti-PaSTAT1, anti–p-PaSTAT1, 11l3 proteins were prepared to investigate the extracellular signals anti-PaSTAT3, and anti–p-PaSTAT3 IgG was validated using of PaCXCR3.1 and PaCXCR3.2. PaCXCL9–11l2 and PaCXCL9– Western blot. Western blot analysis indicated that the phos- 11l3 upregulated the iNOS activity of resting PaCXCR3.1+ macro- phorylation of PaSTAT1 was higher in PaCXCR3.1+ macrophages phages, whereas PaCXCL9–11l1 and PaCXCL9–11l2 upregulated the than in PaCXCR3.2+ macrophages, whereas the phosphorylation arginase activity of resting PaCXCR3.2+ macrophages (Fig. 12E–H). of PaSTAT3 was higher in PaCXCR3.2+ macrophages than in Taken together, our results demonstrate that PaCXCL9–11ls mediate PaCXCR3.1+ macrophages (Fig. 14A, 14B). Treatment of mac- macrophage polarization via PaCXCR3s. rophages with PaSTAT1 and PaSTAT3 siRNA significantly decreased the expression of the respective mRNAs (Fig. 14C, 14D). PaCXCR3s selectively activate PaSTAT1 and PaSTAT3 in In E. coli–infected PaCXCR3.1+ macrophages, iNOS activity was macrophages downregulated after PaSTAT1 knockdown, but it remained un- Having characterized the specific ligands of PaCXCR3.1 and changed after PaSTAT3 knockdown (Fig. 14E). In PaCXCR3.2+ PaCXCR3.2, we next investigated whether PaCXCR3.1 and macrophages, arginase activity was unaffected after PaSTAT1 PaCXCR3.2 also differed with regard to the transcription factors knockdown but was downregulated after PaSTAT3 knockdown that they activated in macrophages. STAT1 and STAT3 mediate (Fig. 14F). PaSTAT1 siRNA treatment decreased the levels of macrophage polarization in mammals (20). Thus, we investigated PaTNF mRNA in PaCXCR3.1+ macrophages, and PaSTAT3 siRNA the roles of PaSTAT1 and PaSTAT3 in ayu macrophage polari- treatment decreased PaIL-10 mRNA expression in PaCXCR3.2+ zation. We first cloned PaSTAT1 and PaSTAT3 from ayu macro- macrophages (Fig. 14G, 14H). These data suggest that STAT1 phages. Sequence comparisons showed that PaSTAT1 shared high and STAT3 mediate macrophage polarization in CXCR3.1+ and amino acid sequence identity with Atlantic salmon STAT1 (98%) CXCR3.2+ macrophages, respectively. and mouse STAT1 (68%). PaSTAT3 also shared high amino acid + + sequence identity with Atlantic salmon STAT3 (97%) and mouse Effect of PaCXCR3.1 and PaCXCR3.2 macrophage infusion STAT3 (87%). Phylogenetic tree analysis grouped teleost STAT1 on sepsis together in a cluster that was distinct from the mammal STAT1 To demonstrate the function of PaCXCR3.1+ and PaCXCR3.2+ and STAT3 cluster (Fig. 13). PaSTAT1 was most closely related to macrophages in vivo, we performed adoptive transfer of macro- STAT1 from rainbow trout and Atlantic salmon (Fig. 13). Phylo- phages, and 5 3 106 macrophages were injected into septic ayu (Fig. genetic tree analysis grouped teleost STAT3 together in a cluster 15A) because sepsis is associated with macrophage dysfunction (38). 4702 TWO TELEOST CXCR3s CONTRIBUTE TO MACROPHAGE POLARIZATION

We further investigated the function of CXCR3 in macrophages. Ayu was selected as the model organism because its macrophage functions are well characterized (22), and its body size allows for macrophage isolation and adoptive transfer (43). After E. coli infection in mice, macrophages exhibit an M1-like Downloaded from

FIGURE 11. The inhibition of binding activity for [125I]-PaCXCL9– 11ls by anti-PaCXCR3s IgG. Cells were incubated with 30 nM [125I]- PaCXCL9–11l1, [125I]-PaCXCL9–11l2, or [125I]-PaCXCL9–11l1. Dis- placement of [125I]-PaCXCL9–11l2 (A) and [125I]-PaCXCL9–11l3 (B)in the presence of increasing concentrations of anti-PaCXCR3.1 IgG. Dis- placement of [125I]-PaCXCL9–11l1 (C) and [125I]-PaCXCL9–11l2 (D)in http://www.jimmunol.org/ the presence of increasing concentrations of anti-PaCXCR3.2 IgG. The values were normalized by setting the speciﬁc binding of 30 nM [125I]- PaCXCL9–11l1, [125I]-PaCXCL9–11l2, or [125I]-PaCXCL9–11l3 to 100%.

The survival of septic ayu was improved by the infusion of PaCXCR3.2+, but not PaCXCR3.1+, macrophages (Fig. 15B). Sepsis is characterized by inefficient bacterial clearance and the overexpression of inflammatory cytokines (38, 42). Therefore, we ex- by guest on September 29, 2021 amined the effects of macrophage infusion on the bacterial burden and cytokine production during L. anguillarum–induced sepsis. Bacterial burden in the peritoneal cavity was lower in PaCXCR3.1+ and PaCXCR3.2+ macrophage-infused ayu compared with ayu infused with PBS-treated macrophages (Fig. 15C). PaCXCR3.1+ macrophage infusion induced the upregulation of PaTNF expression and downregulation of PaIL-10 expression in the spleen, kidney, and liver (Fig. 15D–I). PaCXCR3.2+ macrophage infusion downregulated PaTNF expression and upregulated PaIL-10 expression in the spleen, kidney, and liver (Fig. 15D–I). These results indicate that CXCR3.2+ ayu macrophages significantly improve bacterial sepsis outcomes in ayu (Fig. 16). FIGURE 12. PaCXCL9–11ls mediate macrophage chemotaxis and po- Discussion larization. (A) Effect of PaCXCL9–11ls on the chemotaxis ability of resting CXCR3.1+ macrophages (n = 4). ***p , 0.001, PaCXCL9–11l2 In this study, we detected two ayu CXCR3 genes (PaCXCR3.1 and versus PaCXCL9–11l1. ###p , 0.001, #p , 0.05, PaCXCL9–11l3 versus PaCXCR3.2) and three CXCL9–11 homologs (PaCXCL9–11l1, PaCXCL9–11l1. (B) Effect of PaCXCL9–11ls on the chemotaxis ability of PaCXCL9–11l2, and PaCXCL9–11l3). In mammals, CXCR3 is resting CXCR3.2+ macrophages (n = 4). ***p , 0.001, **p , 0.01, the sole receptor shared by three ligands: CXCL9, CXCL10, and PaCXCL9–11l1 versus PaCXCL9–11l3. ###p , 0.001, ##p , 0.01, CXCL11 (2). Hence, teleost CXCR3s and their ligands were PaCXCL9–11l2 versus PaCXCL9–11l3. (C) Effect of Ab blockade on the different from their mammalian counterparts. Moreover, chemotactic ability of resting CXCR3.1+ macrophages. PaCXCL9–11l2 PaCXCR3.1 was found to be constitutively expressed in macro- and PaCXCL9–11l3 were used at 30 and 10 nM, respectively. ***p , phages, whereas PaCXCR3.2 was induced after E. coli infection 0.001. (D) Effect of Ab blockade on the chemotactic ability of resting + by a ROS-dependent mechanism. There are two or more CXCR3 CXCR3.2 macrophages. PaCXCL9–11l1 and PaCXCL9–11l2 were used , E F genes in most teleosts, including zebrafish, Japanese ricefish, at 30 nM. ***p 0.001. ( and ) Effect of PaCXCL9–11ls on iNOS and arginase activity in resting CXCR3.1+ macrophages (n = 5). The horizontal and tetraodon (10). The expression patterns of PaCXCR3.1 and lines denote means. ***p , 0.001. (G and H) Effect of PaCXCL9–11ls on PaCXCR3.2 in ayu were similar to those in rainbow trout, where iNOS and arginase activity in resting CXCR3.2+ macrophages. The mac- CXCR3.2, but not CXCR3.1, is induced by inflammatory stim- rophages were treated with 30 nM PaCXCL9–11l1, PaCXCL9–11l2, or ulants (12). The expression pattern of teleost CXCR3s indicates PaCXCL9–11l3 (n = 5). The horizontal lines denote means. *p , 0.05, that they may play a crucial role in regulating macrophage **p , 0.01. L9–11l1, PaCXCL9–11l1; L9–11l2, PaCXCL9–11l2; L9– function. 11l3, PaCXCL9–11l3. The Journal of Immunology 4703

teleost ﬁsh. Nevertheless, our investigation demonstrates that ﬁsh pathogens may affect the balance between M1 and M2 macrophages to induce host death. In mammals, macrophage polarization is controlled by environmental cytokines (20), intracellular signaling (21), and metabolic pathways (52). In this study, we found that E. coli infection Downloaded from

FIGURE 13. Phylogenetic analysis of STAT1 and STAT3 genes. Phy- logenetic (neighbor-joining) analysis of the complete amino acid sequences of teleost and mammalian STAT1 and STAT3 genes using the http://www.jimmunol.org/ MEGA5.0 program. Node values represent the percentage bootstrap confidence derived from 1000 replicates. The source sequences are listed in Supplemental Table III. profile (44–46). In this study, PaCXCR3.1+ macrophages exhibited an M1-like profile after E. coli infection, whereas PaCXCR3.2+ macrophages exhibited an M2-like profile after E. coli infection. Uncontrolled M1 inflammation is associated with acute E. coli infection, and the M1–M2 switch may provide by guest on September 29, 2021 protection against uncontrolled inflammation (20, 47, 48). Hence, it is critical that M2 macrophages are produced postinfection. We further found that macrophage polarization was regulated by CXCR3.1 and CXCR3.2 in grass carp and spotted green pufferfish as well, suggesting that the induction of macrophage polarization by CXCR3s is conserved in teleost fish. To our knowledge, this is the first study that demonstrates that macrophage polarization in teleosts is regulated by chemokine receptors. Furthermore, we found that CXCR3.2 was induced in ayu macrophages upon E. coli infection. Bacterial infection activates ROS production in macrophages (49), which serves as a mechanism for bacterial killing (40). Furthermore, ROS is also implicated in the immune response and signal transduction (40). ROS production during infection is dependent on the bacterial species and hosts. For example, viable L. anguillarum does not stimulate ROS production in Crassostrea virginica,butdead L. anguillarum does (50). In this study, we found that E. coli FIGURE 14. CXCR3.1 and CXCR3.2 mediate macrophage polarization infection induced ROS production in ayu macrophages, which by inducing PaSTAT1- and PaSTAT3-dependent pathways. (A and B) mediated CXCR3.2 expression in macrophages. In contrast, Phosphorylation levels of PaSTAT1 and PaSTAT3 proteins in PaCXCR3.1+ L. anguillarum infection did not stimulate ROS production in and PaCXCR3.2+ macrophages. CXCR3.1+ and CXCR3.2+ cells were ayu macrophages. In teleost fish, L. anguillarum is a common subjected to Western blot analysis to compare the phosphorylation levels pathogen that infects fish in seawater (51) and freshwater (38), of PaSTAT1 and PaSTAT3 proteins (n = 3). (C and D) Bar graphs dis- and our results may explain why L. anguillarum has high path- playing the effect of PaSTAT1 and PaSTAT3 siRNA transfection on their mRNA expression by RT-qPCR analysis. Macrophages were transfected ogenicity in teleost fish. The pathogen may tilt the balance be- with PaSTAT1 siRNA (C) or PaSTAT3 siRNA (D) for 48, 72, and 96 h (n = tween the M1 and M2 macrophages in ayu toward the M1 5). (E and F) iNOS and arginase activity in different macrophages after phenotype via the inhibition of ROS production. However, high PaSTAT1 or PaSTAT3 siRNA treatment (n = 5). The horizontal lines de- levels of cytokines produced by the M1-phenotype macrophages note means. (G and H) mRNA levels of PaTNF and PaIL-10 in different in ayu may lead to death. Thus, further investigation is needed to macrophages after PaSTAT1 or PaSTAT3 siRNA treatment (n =6). elucidate whether L. anguillarum shows similar effects in other ***p , 0.001. 4704 TWO TELEOST CXCR3s CONTRIBUTE TO MACROPHAGE POLARIZATION Downloaded from

FIGURE 15. Effect of macrophages on the survival rate of infected ayu. (A) Protocol for adoptive transfer. (B) Survival in ayu infused with PBS-treated or E. coli–treated CXCR3.1+ or CXCR3.2+ macrophages. Ayu were infused with 5 3 106 macrophages and then infected with L. anguillarum (1.2 3 104 + + 6 CFU per ﬁsh) (n = 30). (C) Effect of CXCR3.1 or CXCR3.2 macrophage infusion (5 3 10 ) on bacterial clearance in the peritoneal cavity of ayu infected http://www.jimmunol.org/ with L. anguillarum (n = 6). The horizontal lines denote means. RT-qPCR analysis of PaTNF and PaIL-10 mRNA expression in the spleen (D and E), kidney (F and G), and liver (H and I) of ayu infused with CXCR3.1+ or CXCR3.2+ macrophages (n = 5). *p , 0.05, **p , 0.01, ***p , 0.001. under the same environmental conditions upregulated the surface that other receptors may exist in macrophages for PaCXCL9–11l2 in expression of PaCXCR3.1 and PaCXCR3.2 in M1 and M2 mac- addition to PaCXCR3.1 and PaCXCR3.2. Further investigation is rophages, respectively. Macrophage polarization is dependent on the needed to identify all of the receptors for CXCL9–11ls in teleost. availability of different functional-demand signals (14). In our ex-

periments, the signals from the cell culture environment were the by guest on September 29, 2021 same. Thus, the different receptors (PaCXCR3.1 and PaCXCR3.2) may have induced different signals in macrophages. This suggests that PaCXCR3.1 and PaCXCR3.2 may provide signals for macrophage polarization, rather than only being surface markers of polarization. The CXCR3 ligands CXCL9–11 have been known to regulate the migration and activation of macrophages in mice (53, 54). The number of CXC chemokine genes in teleosts varies among the different species (55). At least two CXCL9–11 homologs exist in teleosts, whereas there are several CXCL9–11 homologs in zebrafish (56, 57). In this study, three CXCL9–11 homologs were identified in ayu. Phylogenetic analysis showed that they were not precisely identical to the CXCL9-, CXCL10-, or CXCL11-like proteins. Hence, we further illustrate the primary functions of three CXCL9–11-like proteins in ayu. PaCXCL9– 11l2 and PaCXCL9–11l3 induced chemotaxis and the M1 phenotype in PaCXCR3.1+ macrophages via interacting with PaCXCR3.1, whereas PaCXCL9–11l1 and PaCXCL9–11l2 induced chemotaxis and the M2 phenotype in PaCXCR3.2+ macrophages via interacting with PaCXCR3.2. These results suggest that CXCR3.1 and CXCR3.2 mediate macrophage polarization by interacting with their specific ligands in teleosts.Inmammals,CXCL9,CXCL10,andCXCL11induce different signaling pathways in immune cells (4). In our study, PaCXCL9–11l1, PaCXCL9–11l2, and PaCXCL9–11l3 FIGURE 16. Teleost CXCR3s regulate macrophage polarization. Three induced various degrees of chemotaxis in PaCXCR3.1+ and CXCL9–11-like proteins specifically interact with CXCR3.1 and CXCR3.2 in macrophages. PaCXCL9–11l2 and PaCXCL9–11l3 induce chemotaxis PaCXCR3.2+ macrophages in ayu. Thus, three ligands and two and the M1 phenotype in PaCXCR3.1+ macrophages, whereas PaCXCL9– CXCR3s in teleost fish may constitute a complex system that 11l1 and PaCXCL9–11l2 induce chemotaxis and the M2 phenotype in regulates macrophage chemotaxis and function. We also found PaCXCR3.2+ macrophages. STAT1 and STAT3 mediate the induction of M1 that anti-PaCXCR3.1 and anti-PaCXCR3.2 IgG could not com- and M2 phenotype macrophages, respectively. CXCR3.2+ macrophages, but pletely inhibit the binding activity of PaCXCL9–11l2, suggesting not CXCR3.1+ macrophages, improve the prognosis of septic fish. The Journal of Immunology 4705

In mice, STAT1 is identified as a crucial transcription factor 2. Clark-Lewis, I., I. Mattioli, J. H. Gong, and P. Loetscher. 2003. Structure- for M1 macrophage polarization (58, 59). In contrast, STAT3 is function relationship between the human chemokine receptor CXCR3 and its ligands. J. Biol. Chem. 278: 289–295. a crucial transcription factor for M2 macrophage polarization 3. Chen, L. C., J. Y. Chen, A. L. Hour, C. Y. Shiau, C. F. Hui, and J. L. Wu. 2008. (60). STAT1 and STAT3 are also implicated in the immune Molecular cloning and functional analysis of zebrafish (Danio rerio) chemokine genes. Comp. Biochem. Physiol. B Biochem. Mol. Biol. 151: 400–409. response in teleost fish (61, 62). In this study, we found that 4. Zohar, Y., G. Wildbaum, R. Novak, A. L. Salzman, M. Thelen, R. Alon, + + STAT1andSTAT3wereactivatedinCXCR3.1 and CXCR3.2 Y. Barsheshet, C. L. Karp, and N. Karin. 2014. CXCL11-dependent induction of macrophages, respectively, after E. coli treatment. This suggests that FOXP3-negative regulatory T cells suppresses autoimmune encephalomyelitis. J. Clin. Invest. 124: 2009–2022. CXCR3.1 and CXCR3.2 regulate macrophage polarization via ac- 5. Karin, N., G. Wildbaum, and M. Thelen. 2016. Biased signaling pathways via tivation of different transcription factors. Various transcription factors CXCR3 control the development and function of CD4+ T cell subsets. J. Leukoc. in mammals regulate macrophage polarization, including STAT1, Biol. 99: 857–862. 6. Janatpour, M. J., S. Hudak, M. Sathe, J. D. Sedgwick, and L. M. McEvoy. 2001. STAT3, STAT5, STAT6, cAMP-responsive element-binding protein, Tumor necrosis factor-dependent segmental control of MIG expression by high and IFN regulatory factors (59). Although the phenomenon of mac- endothelial venules in inflamed lymph nodes regulates monocyte recruitment. rophage polarization occurs in various teleost fish, including zebrafish J. Exp. Med. 194: 1375–1384. 7. Oghumu, S., S. Varikuti, C. Terrazas, D. Kotov, M. W. Nasser, C. A. Powell, (18, 63), the transcription factors involved in the regulation of mac- R. K. Ganju, and A. R. Satoskar. 2014. CXCR3 deficiency enhances tumor rophage polarization have not been elucidated. Our data revealed that progression by promoting macrophage M2 polarization in a murine breast cancer STAT1 and STAT3 regulate the M1 and M2 macrophage phenotypes, model. Immunology 143: 109–119. 8. Zou, J., A. K. Redmond, Z. Qi, H. Dooley, and C. J. Secombes. 2015. The CXC respectively. Thus, our results support the concept that the tran- chemokine receptors of fish: Insights into CXCR evolution in the vertebrates. scription factors for macrophage polarization are conserved in ver- Gen. Comp. Endocrinol. 215: 117–131. tebrates. Together, our work reveals that CXCR3.1 and CXCR3.2 9. Bajoghli, B. 2013. Evolution and function of chemokine receptors in the immune Downloaded from system of lower vertebrates. Eur. J. Immunol. 43: 1686–1692. mediate macrophage polarization via different signaling pathways. 10. Aghaallaei, N., B. Bajoghli, H. Schwarz, M. Schorpp, and T. Boehm. 2010. It is traditionally believed that, in mammals, acquired immunity Characterization of mononuclear phagocytic cells in medaka fish transgenic (with its features of high specificity, Ab maturation, immunological for a cxcr3a:gfp reporter. Proc. Natl. Acad. Sci. USA 107: 18079–18084. 11. Torraca, V., C. Cui, R. Boland, J. P. Bebelman, A. M. van der Sar, M. J. Smit, memory, and secondary responses) is so successful that it allows a M. Siderius, H. P. Spaink, and A. H. Meijer. 2015. The CXCR3-CXCL11 sig- reduction in the number of innate immune molecule variants naling axis mediates macrophage recruitment and dissemination of mycobac- terial infection. Dis. Model. Mech. 8: 253–269. originating from lower vertebrates (64, 65). However, the mam- http://www.jimmunol.org/ 12. Xu, Q., R. Li, M. M. Monte, Y. Jiang, P. Nie, J. W. Holland, C. J. Secombes, and malian immune system appears to be less efficient against sepsis, T. Wang. 2014. Sequence and expression analysis of rainbow trout CXCR2, an infection that evokes the systemic inflammatory reaction syn- CXCR3a and CXCR3b aids interpretation of lineage-specific conversion, loss drome (66). Innate immunity plays an important role in sepsis by and expansion of these receptors during vertebrate evolution. Dev. Comp. Immunol. 45: 201–213. recognizing microorganisms and initiating anti-infectious re- 13. Okin, D., and R. Medzhitov. 2012. Evolution of inflammatory diseases. Curr. sponses (67). In teleost fish, innate immunity is also the first line Biol. 22: R733–R740. 14. Okabe, Y., and R. Medzhitov. 2016. Tissue biology perspective on macrophages. of defense against pathogens (68). Although the adaptive immu- Nat. Immunol. 17: 9–17. nity of teleosts is weak, they possess more diverse innate immune 15. Katzenback, B. A., F. Katakura, and M. Belosevic. 2016. Goldfish (Carassius molecules than do mammals (34, 69). Macrophage dysfunction is auratus L.) as a model system to study the growth factors, receptors and tran-

scription factors that govern myelopoiesis in fish. Dev. Comp. Immunol. 58: 68– by guest on September 29, 2021 related to the survival of septic mice and teleost fish (38, 43). 85. Macrophages are the main cells producing the proinflammatory 16. Freedman, T. S., Y. X. Tan, K. M. Skrzypczynska, B. N. Manz, F. V. Sjaastad, cytokines TNF and IL-1b, which are overproduced in sepsis and H. S. Goodridge, C. A. Lowell, and A. Weiss. 2015. LynA regulates an inflammation-sensitive signaling checkpoint in macrophages. Elife 4: e09183. lead to deleterious effects, such as organ dysfunction (67). We 17. Hodgkinson, J. W., C. Fibke, and M. Belosevic. 2017. Recombinant IL-4/13A found that, after E. coli infection, the PaCXCR3.2 gene in ayu led and IL-4/13B induce arginase activity and down-regulate nitric oxide response of to the polarization of M2 macrophages that controlled the over- primary goldfish (Carassius auratus L.) macrophages. Dev. Comp. Immunol. 67: 377–384. production of proinflammatory cytokines to improve the survival of 18. Wiegertjes, G. F., A. S. Wentzel, H. P. Spaink, P. M. Elks, and I. R. Fink. 2016. septic ayu. Our results support the hypothesis that diverse innate Polarization of immune responses in fish: the ‘macrophages first’ point of view. immune genes in teleosts indeed enhance macrophage function. Our Mol. Immunol. 69: 146–156. 19. Blidner, A. G., S. P. Meńdez-Huergo, A. J. Cagnoni, and G. A. Rabinovich. study shows that insights gained from a comparison of the teleost 2015. Re-wiring regulatory cell networks in immunity by galectin-glycan in- and mammalian immune systems could be useful for the develop- teractions. FEBS Lett. 589: 3407–3418. ment of new therapeutic strategies for human immune disorders. 20. Sica, A., and A. Mantovani. 2012. Macrophage plasticity and polarization: in vivo veritas. J. Clin. Invest. 122: 787–795. In summary, our study characterized the expression and function of 21. Arranz, A., C. Doxaki, E. Vergadi, Y. Martinez de la Torre, K. Vaporidi, two CXCR3 genes in ayu, a teleost. PaCXCR3.1+ and PaCXCR3.2+ E. D. Lagoudaki, E. Ieronymaki, A. Androulidaki, M. Venihaki, A. N. Margioris, macrophages exhibited the M1 and M2 phenotype, respectively, after et al. 2012. Akt1 and Akt2 protein kinases differentially contribute to macrophage polarization. Proc. Natl. Acad. Sci. USA 109: 9517–9522. E. coli stimulation (Fig. 15). Although two CXCR3 genes in teleosts 22. Lu, X. J., X. Y. Hang, L. Yin, Y. Q. He, J. Chen, Y. H. Shi, and C. H. Li. 2013. have been well known for several years, to our knowledge, this is the Sequencing of the first ayu (Plecoglossus altivelis) macrophage transcriptome first study to reveal their differential functions in macrophages. and microarray development for investigation the effect of LECT2 on macrophages. Fish Shellfish Immunol. 34: 497–504. Knowledge of the innate immune system of organisms such as ayu 23. Lu, X. J., Q. Chen, Y. J. Rong, and J. Chen. 2016. Mobilisation and dysfunction may lead to a better understanding of the pathogenesis of infectious of haematopoietic stem/progenitor cells after Listonella anguillarum infection in ayu, Plecoglossus altivelis. Sci. Rep. 6: 28082. diseases, such as sepsis. Thus, we propose that further investigation 24. Lin, A. F., L. X. Xiang, Q. L. Wang, W. R. Dong, Y. F. Gong, and J. Z. Shao. into the evolution of the vertebrate immune system may be useful in 2009. The DC-SIGN of zebrafish: insights into the existence of a CD209 ho- developing therapeutic strategies for treating human diseases. mologue in a lower vertebrate and its involvement in adaptive immunity. J. Immunol. 183: 7398–7410. 25. Wang, T., S. Bird, A. Koussounadis, J. W. Holland, A. Carrington, J. Zou, and C. J. Secombes. 2009. Identification of a novel IL-1 cytokine family member in Disclosures teleost fish. J. Immunol. 183: 962–974. The authors have no financial conflicts of interest. 26. Hong, S., R. Li, Q. Xu, C. J. Secombes, and T. Wang. 2013. Two types of TNF-a exist in teleost fish: phylogeny, expression, and bioactivity analysis of type-II TNF-a3 in rainbow trout Oncorhynchus mykiss. J. Immunol. 191: 5959–5972. 27. Roca, F. J., I. Mulero, A. Lo´pez-Munõz, M. P. Sepulcre, S. A. Renshaw, References J. Meseguer, and V. Mulero. 2008. Evolution of the inflammatory response in 1. Bromley, S. K., T. R. Mempel, and A. D. Luster. 2008. Orchestrating the or- vertebrates: fish TNF-a is a powerful activator of endothelial cells but hardly chestrators: chemokines in control of T cell traffic. Nat. Immunol. 9: 970–980. activates phagocytes. J. Immunol. 181: 5071–5081. 4706 TWO TELEOST CXCR3s CONTRIBUTE TO MACROPHAGE POLARIZATION

28. Joerink, M., C. M. Ribeiro, R. J. Stet, T. Hermsen, H. F. Savelkoul, and 49. West, A. P., I. E. Brodsky, C. Rahner, D. K. Woo, H. Erdjument-Bromage, G. F. Wiegertjes. 2006. Head kidney-derived macrophages of common carp P. Tempst, M. C. Walsh, Y. Choi, G. S. Shadel, and S. Ghosh. 2011. TLR sig- (Cyprinus carpio L.) show plasticity and functional polarization upon differen- nalling augments macrophage bactericidal activity through mitochondrial ROS. tial stimulation. J. Immunol. 177: 61–69. Nature 472: 476–480. 29. Montero, J., V. Go´mez-Abellań, M. Arizcun, V. Mulero, and M. P. Sepulcre. 50. Bramble, L., and R. S. Anderson. 1997. Modulation of Crassostrea virginica 2016. Prostaglandin E2 promotes M2 polarization of macrophages via a cAMP/ hemocyte reactive oxygen species production by Listonella anguillarum. Dev. CREB signaling pathway and deactivates granulocytes in teleost fish. Fish Comp. Immunol. 21: 337–348. Shellfish Immunol. 55: 632–641. 51. Gonza´lez-Stegmaier, R., A. Romero, A. Estepa, J. Montero, V. Mulero, and 30. Chadzinska, M., K. M. Leon-Kloosterziel, B. Plytycz, and B. M. Lidy Verburg- L. Mercado. 2015. Effects of recombinant flagellin B and its ND1 domain from van Kemenade. 2008. In vivo kinetics of cytokine expression during peritonitis Vibrio anguillarum on macrophages from gilthead seabream (Sparus aurata L.) in carp: evidence for innate and alternative macrophage polarization. Dev. Comp. and rainbow trout (Oncorhynchus mykiss, W.). Fish Shellfish Immunol. 42: 144– Immunol. 32: 509–518. 152. 31. Xu, J., F. Chi, T. Guo, V. Punj, W. N. Lee, S. W. French, and H. Tsukamoto. 52. Izquierdo, E., V. D. Cuevas, S. Fernańdez-Arroyo, M. Riera-Borrull, E. Orta- 2015. NOTCH reprograms mitochondrial metabolism for proinflammatory Zavalza, J. Joven, E. Rial, A. L. Corbi, and M. M. Escribese. 2015. Reshaping of macrophage activation. J. Clin. Invest. 125: 1579–1590. human macrophage polarization through modulation of glucose catabolic path- 32. Berthelot, C., F. Brunet, D. Chalopin, A. Juanchich, M. Bernard, B. Noe¨l, ways. J. Immunol. 195: 2442–2451. P. Bento, C. Da Silva, K. Labadie, A. Alberti, et al. 2014. The rainbow trout 53. Corbera-Bellalta, M., E. Planas-Rigol, E. Lozano, N. Terrades-Garcıá, genome provides novel insights into evolution after whole-genome duplication M. A. Alba, S. Prieto-Gonza´lez, A. Garcıá-Martıńez, R. Albero, A. Enjuanes, in vertebrates. Nat. Commun. 5: 3657. G. Espı´gol-Frigole´, et al. 2016. Blocking interferon-g reduces expression of 33. Brunet, F. G., J. N. Volff, and M. Schartl. 2016. Whole genome duplications chemokines CXCL9, CXCL10 and CXCL11 and decreases macrophage infil- shaped the receptor tyrosine kinase repertoire of jawed vertebrates. Genome tration in ex vivo cultured arteries from patients with giant cell arteritis. Ann. Biol. Evol. 8: 1600–1613. Rheum. Dis. 75: 1177–1186. 34. Neves, J. V., C. Caldas, I. Vieira, M. F. Ramos, and P. N. Rodrigues. 2015. 54. Tomita, K., B. L. Freeman, S. F. Bronk, N. K. LeBrasseur, T. A. White, Multiple hepcidins in a teleost fish, Dicentrarchus labrax: different hepcidins for P. Hirsova, and S. H. Ibrahim. 2016. CXCL10-mediates macrophage, but not different roles. J. Immunol. 195: 2696–2709. other innate immune cells-associated inflammation in murine nonalcoholic 35. Chen, Q., X. J. Lu, M. Y. Li, and J. Chen. 2016. Molecular cloning, steatohepatitis. Sci. Rep. 6: 28786. Downloaded from pathologically-correlated expression and functional characterization of the col- 55. Baoprasertkul, P., C. He, E. Peatman, S. Zhang, P. Li, and Z. Liu. 2005. Con- ony stimulating factor 1 receptor (CSF-1R) gene from a teleost, Plecoglossus stitutive expression of three novel catfish CXC chemokines: homeostatic che- altivelis. Dongwuxue Yanjiu 37: 96–102. mokines in teleost fish. Mol. Immunol. 42: 1355–1366. 36. Filardo, E., J. Quinn, Y. Pang, C. Graeber, S. Shaw, J. Dong, and P. Thomas. 56. Chen, J., Q. Xu, T. Wang, B. Collet, Y. Corripio-Miyar, S. Bird, P. Xie, P. Nie, 2007. Activation of the novel estrogen receptor G protein-coupled receptor 30 C. J. Secombes, and J. Zou. 2013. Phylogenetic analysis of vertebrate CXC (GPR30) at the plasma membrane. Endocrinology 148: 3236–3245. chemokines reveals novel lineage specific groups in teleost fish. Dev. Comp. 37.Fairbairn,I.P.,C.B.Stober,D.S.Kumararatne, and D. A. Lammas. 2001. ATP- Immunol. 41: 137–152.

mediated killing of intracellular mycobacteria by macrophages is a P2X(7)-dependent 57. Fu, Q., Q. Zeng, Y. Li, Y. Yang, C. Li, S. Liu, T. Zhou, N. Li, J. Yao, C. Jiang, http://www.jimmunol.org/ process inducing bacterial death by phagosome-lysosome fusion. J. Immunol. 167: et al. 2017. The chemokinome superfamily in channel catfish: I. CXC subfamily 3300–3307. and their involvement in disease defense and hypoxia responses. Fish Shellfish 38. Lu, X. J., J. Chen, C. H. Yu, Y. H. Shi, Y. Q. He, R. C. Zhang, Z. A. Huang, Immunol. 60: 380–390. J. N. Lv, S. Zhang, and L. Xu. 2013. LECT2 protects mice against bacterial 58. Baer, C., M. L. Squadrito, D. Laoui, D. Thompson, S. K. Hansen, A. Kiialainen, sepsis by activating macrophages via the CD209a receptor. J. Exp. Med. 210: S. Hoves, C. H. Ries, C. H. Ooi, and M. De Palma. 2016. Suppression of 5–13. microRNA activity amplifies IFN-g–induced macrophage activation and pro- 39. Chang, M. X., B. J. Sun, and P. Nie. 2007. The first non-mammalian CXCR3 in a motes anti-tumour immunity. Nat. Cell Biol. 18: 790–802. teleost fish: gene and expression in blood cells and central nervous system in the 59. Lawrence, T., and G. Natoli. 2011. Transcriptional regulation of macrophage grass carp (Ctenopharyngodon idella). Mol. Immunol. 44: 1123–1134. polarization: enabling diversity with identity. Nat. Rev. Immunol. 11: 750–761. 40. Lambeth, J. D. 2004. NOX enzymes and the biology of reactive oxygen. Nat. 60. Mandal, P., B. T. Pratt, M. Barnes, M. R. McMullen, and L. E. Nagy. 2011. Rev. Immunol. 4: 181–189. Molecular mechanism for adiponectin-dependent M2 macrophage polarization:

41. Zhang, Y., S. Choksi, K. Chen, Y. Pobezinskaya, I. Linnoila, and Z. G. Liu. 2013. link between the metabolic and innate immune activity of full-length adipo- by guest on September 29, 2021 ROS play a critical role in the differentiation of alternatively activated macro- nectin. J. Biol. Chem. 286: 13460–13469. phages and the occurrence of tumor-associated macrophages. Cell Res. 23: 898– 61. Eslamloo, K., X. Xue, M. Booman, N. C. Smith, and M. L. Rise. 2016. Tran- 914. scriptome profiling of the antiviral immune response in Atlantic cod macro- 42. Lu, X. J., Q. Chen, Y. J. Rong, G. J. Yang, C. H. Li, N. Y. Xu, C. H. Yu, phages. Dev. Comp. Immunol. 63: 187–205. H. Y. Wang, S. Zhang, Y. H. Shi, and J. Chen. 2016. LECT2 drives haemato- 62. Huang, Y., X. Huang, Y. Yang, W. Wang, Y. Yu, and Q. Qin. 2015. Involvement poietic stem cell expansion and mobilization via regulating the macrophages and of fish signal transducer and activator of transcription 3 (STAT3) in nodavirus osteolineage cells. Nat. Commun. 7: 12719. infection induced cell death. Fish Shellfish Immunol. 43: 241–248. 43. Chen, J., Q. Chen, X. J. Lu, and C. H. Li. 2014. LECT2 improves the outcomes 63. Nguyen-Chi, M., B. Laplace-Builhe, J. Travnickova, P. Luz-Crawford, in ayu with Vibrio anguillarum infection via monocytes/macrophages. Fish G. Tejedor, Q. T. Phan, I. Duroux-Richard, J. P. Levraud, K. Kissa, G. Lutfalla, Shellfish Immunol. 41: 586–592. et al. 2015. Identification of polarized macrophage subsets in zebrafish. Elife 4: 44. Liang, M. D., A. Bagchi, H. S. Warren, M. M. Tehan, J. A. Trigilio, e07288. L. K. Beasley-Topliffe, B. L. Tesini, J. C. Lazzaroni, M. J. Fenton, and 64. Boehm, T., and J. B. Swann. 2014. Origin and evolution of adaptive immunity. J. Hellman. 2005. Bacterial peptidoglycan-associated lipoprotein: a naturally Annu. Rev. Anim. Biosci. 2: 259–283. occurring toll-like receptor 2 agonist that is shed into serum and has synergy 65. Star, B., A. J. Nederbragt, S. Jentoft, U. Grimholt, M. Malmstrøm, T. F. Gregers, with lipopolysaccharide. J. Infect. Dis. 191: 939–948. T. B. Rounge, J. Paulsen, M. H. Solbakken, A. Sharma, et al. 2011. The genome 45. Pinheiro da Silva, F., M. Aloulou, D. Skurnik, M. Benhamou, A. Andremont, sequence of Atlantic cod reveals a unique immune system. Nature 477: 207–210. I. T. Velasco, M. Chiamolera, J. S. Verbeek, P. Launay, and R. C. Monteiro. 66. Levy, M. M., M. P. Fink, J. C. Marshall, E. Abraham, D. Angus, D. Cook, 2007. CD16 promotes Escherichia coli sepsis through an FcR g inhibitory J. Cohen, S. M. Opal, J. L. Vincent, and G. Ramsay, International Sepsis Defi- pathway that prevents phagocytosis and facilitates inflammation. Nat. Med. 13: nitions Conference. 2003. 2001 SCCM/ESICM/ACCP/ATS/SIS international 1368–1374. sepsis definitions conference. Intensive Care Med. 29: 530–538. 46. Benoit, M., B. Desnues, and J. L. Mege. 2008. Macrophage polarization in 67. Annane, D., E. Bellissant, and J.-M. Cavaillon. 2005. Septic shock. Lancet 365: bacterial infections. J. Immunol. 181: 3733–3739. 63–78. 47. Sica, A., M. Erreni, P. Allavena, and C. Porta. 2015. Macrophage polarization in 68. Riera Romo, M., D. Pe´rez-Martıńez, and C. Castillo Ferrer. 2016. Innate im- pathology. Cell. Mol. Life Sci. 72: 4111–4126. munity in vertebrates: an overview. Immunology 148: 125–139. 48. Forlenza, M., I. R. Fink, G. Raes, and G. F. Wiegertjes. 2011. Heterogeneity of 69. Sunyer, J. O. 2013. Fishing for mammalian paradigms in the teleost immune macrophage activation in fish. Dev. Comp. Immunol. 35: 1246–1255. system. Nat. Immunol. 14: 320–326. Supplemental Data

SUPPLEMENTAL TABLE I. primers used in this work. Primer Gene Accession No./Ensemble ID Nucleotide sequence (5′→3′) PaTNF F TNF-α JP740414 ACATGGGAGCTGTGTTCCTC PaTNF R GCAAACACACCGAAAAAGGT PaIL-1βF IL-1β HF543937 TACCGGTTGGTACATCAGCA PaIL-1βR TGACGGTAAAGTTGGTGCAA PaIL-10F IL-10 KY799108 TGCTGGTGGTGCTGTTTATGTGT PaIL-10R AAGGAGCAGCAGCGGTCAGAA PaTGF-βF TGF-β JP742920 GATCCAGAACCTGAGGGACA PaTGF-βR CTGGAATGCCGAGAACAAAT PaCXCR3.1F CXCR3.1 JP725619 CTACTTCGGTCGAGTCCTGC PaCXCR3.1R GACGAAGAACACCAGAGCCA PaCXCR3.2F CXCR3.2 KY081643 CCTTCGCCTTATCGTCTCCC PaCXCR3.2R CTCCGTCACCATTGTCCCAA PaIL-12 p40aF IL-12 p40a JP722560 TGAGGATACGCACTGAGACG PaIL-12 p40aR TCTGTACTGCCACGAGCAAC PaIL-12 p40cF IL-12 p40c KY081641 CCCTGACTGTTTGGAGGGTA PaIL-12 p40cR CCTGAAGTATGAGCGGGTGT PaSTAT1F STAT1 JP742610 ATCTCCAATGTGAGCCAGCT PaSTAT1R GCTTGTCTGCCAGCATATCC PaSTAT3F STAT3 JP722452 AGGCTCGAGGAAGTTCAACA PaSTAT3R AAGGTGATGAGGTGGAGCTC Pa18S rRNAF 18S rRNA FN646593 GAATGTCTGCCCTATCAACT Pa18s rRNAR GATGTGGTAGCCGTTTCT CiTNF F TNF HQ696609 CATCCATTTAACAGGTGCATAC CiTNF R GCAGCAGATGTGGAAAGAGAC CiIL-1βF IL-1β JN705663 GATTCGAAAGTTCGATTCAATCT CiIL-1βR TTCAGTGACCTCCTTCAAGAC CiIL-10F IL-10 HQ388294 TGCTCATTTGTGGAGGGCTT CiIL-10R GGTCTCCAAGTAGAAGCGCA CiTGF-βF TGF-β EU099588 TGGACTGGAAGTGGATGCAT CiTGF-βR TCTAGCACTTGGGGTACACG CiCXCR3.1F CXCR3.1 AY744916 CCATCCGTGTCATCGTAGCC CiCXCR3.1R GCCGTCACAGGATGTTTGGT CiCXCR3.2F CXCR3.2 KY081642 GGTCAGTTGGGCGCCTTATA CiCXCR3.2R GAGCTCACGACGGAACTTCA Ci18S rRNAF 18S rRNA EU047719 ATTTCCGACACGGAGAGG Ci18S rRNAR CATGGGTTTAGGATACGCTC TnTNF F TNF CGGTGAGGTCGGCGTGTCA CAAE01014565 TnTNF R TTGGTCTCGGTCCACAGTTTGTG TnIL-1βF IL-1β AJ574910 TCAATCCAACCGTGAGGC TnIL-1βR TGTCGTGCTTGTAGAACAGAA TnIL-10F IL-10 AJ544915 CGGGAATACTACGAGGCAAAT TnIL-10R CGTCAGGGTGTTCATGTCAAA TnTGF-βF TGF-β CR700904 CAGACACCCACAGACGCTAG TnTGF-βR GTTTTGCGTAACGGCTGAGG Tn18S rRNAF 18S rRNA CR734162 AGCAACTTTAGTATACGCTATTG Tn18S rRNAR CCTGAGAAACGGCTACCACATCC PaCXCL9–11l1pF CXCL9-11l1 KJ130413 CCGGAATTCAGAGGAAGTGACGTCATGAAa PaCXCL9–11l1pR CGGGGTACCTCAGTACAGTGTCTTCAGTTa PaCXCL9–11l2pF CXCL9-11l2 KU362928 CCGGAATTCTCTGGGTCAGTTGGTGGTGa PaCXCL9–11l2pR CGGGGTACCTCAGACGTGACTGCTGCGTa PaCXCL9–11l3pF CXCL9-11l3 KU362929 CCGGAATTCCAGCTGGGTGGGCGCCAa PaCXCL9–11l3pR CGGGGTACCTCATTCCCTTTTCTCTTCTTGa PaIL-10pF IL-10 JP758157 CGGAATTCACTCGTGTGGTGTGTTCTGAb PaIL-10pR GCCTCGAGTCAATGTTTTCTTCGTTTAGATGCCb PaSTAT1pF STAT1 JP742610 CGGAATTCAAGGTGCTGAGCTGGCAGTTb PaSTAT1pR GCCTCGAGCTATTCAGCAAACATCATGTAGb PaSTAT3pF STAT3 JP722452 CGGAATTCGGACCTTGTGTGAACTACTCb PaSTAT3pR GCCTCGAGTTACATAGGAGAGGCCACATb PaCXCR3.1pF CXCR3.1 JP725619 CACCGAATTCAGAGACATGTATTGTTTACTCCAGAG PaCXCR3.1pR CAAGGCTTCTAGATCAGCAATGGCTGGACTGTTc PaCXCR3.2pF CXCR3.2 KY081643 CACCGAATTCAGAGACATGGATTCTTTCAGGGCAAC PaCXCR3.2pR CAAGGCTTCTAGATCACACCATGACAGAAAAGTATc CiCXCR3.1pF CXCR3.1 AY744916 CACCGAATTCAGAGACATGGATAATATAGCCTATGTT CiCXCR3.1pR CAAGGCTTCTAGATTAGATGAACTTCACCCCCAc CiCXCR3.2pF CXCR3.2 KY081642 CACCGAATTCAGAGACATGAAGAACAGCACGACAT CiCXCR3.2pR CAAGGCTTCTAGATCAGGCCATGACAGAGAAGTc TnCXCR3.1pF CXCR3.1 GSTENG00015608001 CACCGAATTCAGAGACATGGATAACATGACCACAGA TnCXCR3.1pR CAAGGCTTCTAGATCACACCATGACTGAGTTTAAGc TnCXCR3.2pF CXCR3.2 GSTENG00015610001 CACCGAATTCAGAGACATGGGAAACGTTCAGAAGA TnCXCR3.2pR CAAGGCTTCTAGATCCAAATCTGTCGCTTTCTGAc a The underlined nucleotides represent the restriction sites for EcoR I and Kpn I, respectively. b The underlined nucleotides represent the restriction sites for EcoR I and Xho I, respectively. c The underlined nucleotides represent the restriction sites for EcoR I and Xba I, respectively.

1 SUPPLEMENTAL TABLE II. CXCL8-13 sequences used in this study Species Accession no./Ensemble ID Protein Latin name English name NM 008176 Mus musculus house mouse CXCL1 NM_001511 Homo sapiens human CXCL1 NM_009140 Mus musculus house mouse CXCL2 NM_002089 Homo sapiens human CXCL2 NM_203320 Mus musculus house mouse CXCL3 NM_002090 Homo sapiens human CXCL3 NM_019932 Mus musculus house mouse CXCL4 NM_002619 Homo sapiens human CXCL4 NM_009141 Mus musculus house mouse CXCL5 NM_002994 Homo sapiens human CXCL5 NM_001081886 Equus caballus horse CXCL6 NM_002993 Homo sapiens human CXCL6 GU187911 Canis lupus familiaris dog CXCL7 BC028217 Homo sapiens human CXCL7 NM_213867 Sus scrofa pig CXCL8 NM_000584 Homo sapiens human CXCL8 AJ279069 Oncorhynchus mykiss rainbow trout CXCL8l HQ872500 Cynoglossus semilaevis tongue sole CXCL8 XM_004065727 Oryzias latipes Japanese ricefish CXCL8l NM_001113651 Danio rerio zebrafish CXCL8l-1 XM_001342570 Danio rerio zebrafish CXCL8l-2 NM_001140710 Salmo salar Atlantic salmon CXCL8l BT079497 Esox lucius northern pike CXCL8l AJ421443 Cyprinus carpio common carp CXCL8l NM_001303371 Larimichthys crocea large yellow croaker CXCL8l NM_002416 Homo sapiens human CXCL9 NM_008599 Mus musculus house mouse CXCL9 NM_001565 Homo sapiens human CXCL10 NM_021274 Mus musculus house mouse CXCL10 AY335951 Ictalurus furcatus blue catfish CXCL10 XM_690954 Danio rerio zebrafish CXCL10l XM_017684482 Pygocentrus nattereri red-bellied piranha CXCL10l XM_017462662 Ictalurus punctatus channel catfish CXCL10l XM_015602833 Astyanax mexicanus Mexican tetra CXCL10l-1 XM_007238695 Astyanax mexicanus Mexican tetra CXCL10l-2 XM_017451843 Ictalurus punctatus channel catfish CXCL11l KX451318 Ictalurus punctatus channel catfish CXCL11-1 AY335950 Ictalurus punctatus channel catfish CXCL11-2 AJ417078 Oncorhynchus mykiss rainbow trout CXCL11l BT078895 Esox lucius northern pike CXCL11 AB082985 Cyprinus carpio common carp CXCL11 XM_001339271 Danio rerio zebrafish CXCL11-1 NM_001126413 Danio rerio zebrafish CXCL11-2 NM_019494 Mus musculus house mouse CXCL11 NM_005409 Homo sapiens human CXCL11 AJ536027 Cyprinus carpio common carp CXCL12b AJ627274 Cyprinus carpio common carp CXCL12a NM_178307 Danio rerio zebrafish CXCL12 XM_010872749 Esox lucius northern pike CXCL12 AM850703 Oryzias latipes Japanese ricefish CXCL12-1 001104727 Oryzias latipes Japanese ricefish CXCL12-2 HE578135 Oncorhynchus mykiss rainbow trout CXCL12a HE578136 Oncorhynchus mykiss rainbow trout CXCL12b NM_021704 Mus musculus house mouse CXCL12 NM_000609 Homo sapiens human CXCL12 XM_005165388 Danio rerio zebrafish CXCL13 XM_017451844 Ictalurus punctatus channel catfish CXCL13 XM_004084737 Oryzias latipes Japanese ricefish CXCL13 HE578137 Oncorhynchus mykiss rainbow trout CXCL13 NM_001303343 Larimichthys crocea large yellow croaker CXCL13 XM_020104199 Paralichthys olivaceus Japanese flounder CXCL13 NM_018866 Mus musculus house mouse CXCL13 NM_006419 Homo sapiens human CXCL13 XM_004073010 Oryzias latipes Japanese ricefish CXCL14 NM_001281358 Oncorhynchus mykiss rainbow trout CXCL14 XM_010742927 Larimichthys crocea large yellow croaker CXCL14 NM_131627 Danio rerio zebrafish CXCL14 XM_019125536 Cyprinus carpio common carp CXCL14 NM_004887 Homo sapiens human CXCL14 NM_019568 Mus musculus house mouse CXCL14 KJ130413 Plecoglossus altivelis ayu CXCL9–11l1 KU362929 Plecoglossus altivelis ayu CXCL9–11l3 KU362928 Plecoglossus altivelis ayu CXCL9–11l2 NM_001113651 Danio rerio zebrafish CXCL19 NM_001115055 Danio rerio zebrafish CXCL20

2 SUPPLEMENTAL TABLE III. STAT1 and STAT3 sequences used in this study. Accession no. Species Protein Latin name English name NM_001124707 Oncorhynchus mykiss rainbow trout STAT1 NM_001123654 Salmon salar Atlantic salmon STAT1 JP742610 Plecoglossus altivelis ayu STAT1 CAG10270 Tetraodon nigroviridis spotted green pufferfish STAT1 XM_011481679 Oryzias latipes Japanese ricefish STAT1 NM_131480 Danio rerio zebrafish STAT1 KU508677 Ctenopharyngodon idella grass carp STAT1 NM_001205313 Mus musculus mouse STAT1 NM_007315 Homo sapiens human STAT1 BC151378 Bos taurus cattle STAT1 NM_213769 Sus scrofa pig STAT1 XM_843260 Canis lupus familiaris dog STAT1 NM_001124708 Oncorhynchus mykiss rainbow trout STAT3 XM_014204116 Salmon salar Atlantic salmon STAT3 JP722452 Plecoglossus altivelis ayu STAT3 AF307106 Tetraodon nigroviridis spotted green pufferfish STAT3 NM_001104838 Oryzias latipes Japanese ricefish STAT3 AB018219 Danio rerio zebrafish STAT3 KC978890 Ctenopharyngodon idella grass carp STAT3 U06922 Mus musculus mouse STAT3 NM_139276 Homo sapiens human STAT3 BC109481 Bos taurus cattle STAT3 NM_001044580 Sus scrofa pig STAT3 XM_014116558 Canis lupus familiaris dog STAT3

3 SUPPLEMENTAL FIGURE 1. Validation of the specificity of antibodies used in this study. (A) Flow cytometry analysis of HEK-293T cells expressing PaCXCR3.1, PaCXCR3.2, CiCXCR3.1, CiCXCR3.2, TnCXCR3.1, and TnCXCR3.2. NC: isotype antibody, HEK-293T-EV: HEK-293T cells were transfected with empty vector. HEK-293T-CXCR3s: HEK-293T cells were transfected with PaCXCR3.1, PaCXCR3.2, CiCXCR3.1, CiCXCR3.2, TnCXCR3.1, or TnCXCR3.2. (B) Validation of the specificity of polyclonal anti-PaCXCR3.1 and anti-PaCXCR3.2 IgG. Lane 1: negative control; Lane 2: total protein from macrophages; Lane 3: total protein from HEK-293T cells expressing PaCXCR3.1 or PaCXCR3.2. (C) Validation of the specificity of anti-PaIL-10, anti-PaSTAT1, anti-p-PaSTAT1, anti-PaSTAT3, and anti-p-PaSTAT3 IgG. For SDS-PAGE, Lane M: protein marker; Lane 1 and 2: protein from BL21 (DE3) transformed with the plasmid before and after IPTG induction; Lane 3: purified recombinant PaIL-10, PaSTAT1 or PaSTAT3; For Western blot with anti-PaIL-10, anti-PaSTAT1 or anti-PaSTAT3 IgG, Lane 4: negative control; Lane 5: purified recombinant PaIL-10, PaSTAT1 or PaSTAT3; Lane 6: ayu serum for PaIL-10, total protein from macrophages for PaSTAT1 and PaSTAT3; For Western blot with anti-p-PaSTAT1 or anti-p-PaSTAT3 IgG, Lane 7: negative control; Lane 8: purified recombinant PaSTAT1 or PaSTAT3; Lane 9: total protein from macrophages.