Lymphocytes to Inflamed Liver + CD8 Role for CXCR6 in Recruitment Of

Role for CXCR6 in Recruitment of Activated CD8 + Lymphocytes to Inflamed Liver Tohru Sato, Henrik Thorlacius, Brent Johnston, Tracy L. Staton, Wenkai Xiang, Dan R. Littman and Eugene C. This information is current as Butcher of October 2, 2021. J Immunol 2005; 174:277-283; ; doi: 10.4049/jimmunol.174.1.277 http://www.jimmunol.org/content/174/1/277 Downloaded from

References This article cites 35 articles, 14 of which you can access for free at: http://www.jimmunol.org/content/174/1/277.full#ref-list-1 http://www.jimmunol.org/ Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average by guest on October 2, 2021

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 © 2005 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Role for CXCR6 in Recruitment of Activated CD8؉ Lymphocytes to Inﬂamed Liver1

Tohru Sato,2,3*†‡ Henrik Thorlacius,3,4*‡ Brent Johnston,5*‡ Tracy L. Staton,*‡ Wenkai Xiang,§ Dan R. Littman,§ and Eugene C. Butcher*‡

Hepatic infiltration of activated CD8 lymphocytes is a major feature of graft-vs-host disease (GvHD). Chemoattractant cytokines and their receptors are key regulators of lymphocyte trafficking, but the involvement of chemoattractant receptors in the physiologic recruitment of cells into the inflamed liver has not been defined. The present study examines the role of the chemokine receptor CXCR6, which is highly expressed by liver-infiltrating CD8 T cells. Hepatic accumulation of donor CD8, but not donor ؊/؊ b bxd CD4, lymphocytes was significantly reduced in GvHD induced by transfer of CXCR6 , H-2D lymphocytes into BDF1, H-2D recipients. To determine whether altered recruitment contributes to the reduced accumulation, CXCR6؊/؊ or wild-type splenic lymphocytes participating in an active GvHD response were isolated and transferred i.v. into secondary recipients with active Downloaded from (GvHD, and the short term (6-h) recruitment of transferred cells to the inflamed liver was assessed. CXCR6؊/؊ CD8 (but not CD4 cells displayed a significant (33%) reduction in liver localization, whereas frequencies in blood of CXCR6؊/؊ and wild-type CD8 cells were similar. Proliferation and apoptosis of liver-infiltrating donor CD8 cells were unaffected. We conclude that CXCR6 helps mediate the recruitment of activated CD8 lymphocytes in GvHD-induced hepatitis and may be a useful target to treat pathological inflammation in the liver. The Journal of Immunology, 2005, 174: 277–283.

onor T cell recruitment into target tissues is a prominent Ͼ20 chemokine receptors have been described (4). The role of http://www.jimmunol.org/ component of graft-vs-host disease (GvHD),6 and a ma- specific chemokine receptors in attracting activated lymphocytes D jor complication in allogenic bone marrow transplanta- appears to be both stimulus and organ dependent, contributing to tion (1). Recruitment of activated lymphocytes is a multistep pro- inflammatory and homeostatic tissue localization of lymphocyte cess in which the expression of specific adhesion molecules and subsets (5). In GvHD the liver is a key target organ, characterized chemokine receptors endows lymphocytes with the ability to ac- by massive lymphocyte accumulation (6), which may reflect en- cess extravascular tissues (2). Chemokine receptors play a central hanced recruitment and/or proliferation. role in lymphocyte recruitment by rapidly transmitting intracellu- Acute GvHD is initiated by donor T cells recognizing and re- lar signals and triggering increased integrin affinity to adhesion acting to host allo-Ags and is characterized by a Th1-polarized by guest on October 2, 2021 molecule ligands displayed by endothelial cells (3). At present, response (7). Different combinations of chemokine receptors are associated with Th1- vs Th2-dependent immune reactions (8). Ac- cordingly, it has been shown that the chemokine receptor CXCR6 *Laboratory of Immunology and Vascular Biology, Department of Pathology, and †Department of Medicine, Stanford University School of Medicine, Stanford, CA (Bonzo/STRL33/TYMSTR) is expressed on T cells of Th1 phe- 94305; ‡Center for Molecular Biology and Medicine, Veterans Affairs Palo Alto notype (9–13). CXCR6, originally described as a novel simian § Health Care System, Palo Alto, CA 94304; and Howard Hughes Medical Institute immunodeficiency virus receptor (14) and fusion cofactor for and Molecular Pathogenesis Program, Skirball, Institute, New York University School of Medicine, New York, NY 10016 HIV-1 strains (15), is expressed at very low levels in naive CD8 Received for publication March 23, 2004. Accepted for publication October 22, 2004. cells (16, 17), but can be up-regulated in T cells by activated den- The costs of publication of this article were defrayed in part by the payment of page dritic cells (DC) (9–11). Moreover, it has been shown that many charges. This article must therefore be hereby marked advertisement in accordance Th1-polarized cells (13) and in vitro-activated CD8 cells (16) are with 18 U.S.C. Section 1734 solely to indicate this fact. attracted by the CXCR6 ligand, CXC ligand 16 (CXCL16; SR- 1 This work was supported by grants from the National Institutes of Health, the De- PSOX/SexCKine). In target tissues, CXCR6-expressing cells are partment of Veterans Affairs (to E.C.B.), and the FACS Core Facility of the Stanford Digestive Diseases Center. T.S. was supported by a K08 grant from the National readily found in the liver of patients with alcohol- or hepatitis Institutes of Health. H.T. was supported by the Swedish Research Council (2001- C-induced liver cirrhosis (9, 18). CXCL16 is abundantly expressed 6576, 2002-955, 2002-8012, and 2003-4661), Crafoordska stiftelsen, Blanceflors stiftelse, Einar och Inga Nilssons stiftelse, Harald och Greta Jaenssons stiftelse, Greta och in lymphoid and nonlymphoid tissues, such as the liver (13, 16, Johan Kocks stiftelser, Fro¨ken Agnes Nilssons stiftelse, Magnus Bergvalls stiftelse, 19). Furthermore, the kinetics of DC-mediated up-regulation also Mossfelts stiftelse, Nanna Svartz stiftelse, Ruth och Richard Julins stiftelse, Svenska suggest a role of CXCR6 in facilitating effector T cell migration La¨karesa¨llskapet, and Teggers stiftelse. into sites of pathological inflammation; CXCR6 is up-regulated, 2 Address correspondence and reprint requests to Dr. Tohru Sato, Veterans Affairs Palo Alto Health Care System, 3801 Miranda Avenue, MC154B, Palo Alto, CA whereas CCR7, a chemokine receptor involved in homing to 94304. E-mail address: [email protected] lymph nodes, is down-regulated after prolonged incubation with 3 T.S. and H.T. contributed equally to this work. DC under Th1-polarizing conditions (11). However, the role of 4 Current address: Department of Surgery, Lund University, Malmo University Hos- CXCR6 in lymphocyte recruitment in vivo is not known. pital, S-205 02, Malmo, Sweden. The aim of this study was to assess the contribution of CXCR6 5 Current address: Dalhousie University, Sir Charles Tupper Medical Building, Room to the recruitment of activated lymphocytes into inflamed liver. For 7C, 5850 College Street, Halifax, Nova Scotia, Canada B3H 1X5. this purpose we developed a short term, in vivo recruitment assay 6 Abbreviations used in this paper: GvHD, graft-vs-host disease; 7-AAD, 7-amino- of activated lymphocytes into the inflamed livers of mice under- actinomycin D; CXCL, CXC ligand; DC, dendritic cell; LIL, liver-infiltrating lym- Ϫ Ϫ phocyte; WT, wild type. going active GvHD and compared the recruitment of CXCR6 /

ϩ ϩ vs CXCR6 / lymphocytes. Our results show that CXCR6 has a CTG ACT ATG TGC-3Ј; ␤-actin (forward), 5Ј-ATG TTT GAG ACC TTC significant role in recruiting activated CD8, but not CD4, lympho- AAC ACC-3Ј; and ␤-actin (reverse), 5Ј-TCT CCA GGG AGG AAG AGG Ј cytes to the inflamed liver in GvHD-induced hepatitis. AT-3 . For real-time PCR of CXCR6 and CXCL16 from normal and GvHD liver, liver tissue was harvested, and total RNA was isolated as described above and converted to cDNA using the Reverse Transcription Materials and Methods System (Promega) following the manufacturer’s instructions. 15Sa was Animal model of GvHD used as an endogenous control for relative quantification using the com- parative threshold cycle method. All reactions were performed in triplicate b C57BL/6J (B6, H-2 ; The Jackson Laboratory), C57BL/6J- on an Mx3000P Real-Time PCR System (Stratagene) using 2ϫ SYBR a a a b Igh Thy1 Gpi1 (Thy1.1, H-2 ; The Jackson Laboratory), and CXCR6- Green and the following primers: 15S (forward), 5Ј-GAAGCACGGATA Ϫ/Ϫ GFP knockin mice (CXCR6 , backcrossed onto C57BL/6 more than CATTGGTGA3Ј; 15S (reverse), 5Ј-TGTTCTGCCATTTCTCTAGGT-3Ј; b eight generations, H-2 ) were used as donors. CXCR6-GFP mice have the CXCR6 (forward), 5Ј-CCCTGTACTTTATGCCTTTG-3Ј; CXCR6 (re- CXCR6-coding sequence replaced with a sequence for enhanced GFP, verse), 5Ј-CTTGGAACTGTCCTCAGAAG-3Ј; CXCL16 (forward), 5Ј- which is regulated through the CXCR6 promoter (17). B6D2F1/J mice CCTTGTCTCTTGCGTTCTTCC-3Ј; CXCL16 (reverse), 5Ј-TCCAAAG bxd (BDF1, H-2 ; The Jackson Laboratory) were used as recipients or as TACCCTGCGGTATC-3Ј; CXCL10 (forward), 5Ј-CCAAGTGCTGC donors for syngeneic control experiments. All mice were 8–12 wk old and CGTCATTTTC-3Ј; and CXCL10 (reverse), 5Ј-GGCTCGCAGGGATG housed under conventional conditions in the animal facility at the Veterans ATTTCAA-3Ј. The chemokine or chemokine receptor expression in Affairs Palo Alto Health Care System. All experiments were approved by normal liver was given an arbitrary value of 1, and their expression in the local ethics committee at Stanford University. Nonlethal GvHD was GvHD liver was expressed as fold up-regulation from normal. induced as previously described (20). Briefly, 50 ϫ 106 donor lymphocytes isolated from the spleen were transferred i.v. into recipient mice on day 0. Flow cytometry The splenocytes from C57BL/6 and CXCR6Ϫ/Ϫ were similar with respect 6 to the expression of activation markers (CD69, CD25, CD44, and Isolated cells (1–2 ϫ 10 ) were stained with the following rat anti-mouse

CD45RB) and adhesion molecules (L-selectin and ␣ ␤ ) and cell subset Abs conjugated to FITC, PE, PerCP, PE-indotricarbocyanine, allophyco- Downloaded from 4 7 d distribution (CD4, CD8, NK, NKT, and B cells; data not shown). On day cyanin, allophycocyanin-indotricarbocyanine, or biotin: H-2D (clone 34- 2-12), CD4 (clone RM4-5), CD8␣ (clone 53-6.7), CD44 (clone IM7), 7, recipient mice were killed by CO2 inhalation, and the liver, spleen, and blood were harvested. In indicated experiments, syngeneic lymphocytes CD45RB (clone 16A), CD25 (clone PC61), CD69 (clone H1.2F3), L-se- ␣ ␤ were labeled with 0.5 ␮M CFSE (Molecular Probes) at 37°C for 10 min lectin (clone Mel-14), 4 7 (clone DATK 32), Thy1.2 (clone 53-2.1), b before injection. CD19 (clone 1D3), IgD (clone 217-170), and NK1.1 (clone PK136). The above-mentioned Abs were obtained from BD Pharmingen. NeutrAvidin Short term recruitment assay Cascade Blue was purchased from Molecular Probes. To study proliferation, we used a BrdU Flow Kit (BD Biosciences, San Jose, CA) following http://www.jimmunol.org/ GvHD was induced in BDF1 mice by transferring lymphocytes from B6- Ϫ/Ϫ the manufacturer’s instructions. Briefly, 1 mg of BrdU was injected i.p. 2 h Thy1.2, congenic B6-Thy1.1, or CXCR6 -Thy1.2 mice as described before death, and the isolated lymphocytes were subsequently permeabil- above. This congenic model for identification of donor lymphocytes was Ϫ/Ϫ ized for detection of intracellular BrdU. Four-color flow cytometry was chosen because using CXCR6 mice expressing GFP as donors excluded performed using a FACSCalibur (BD Biosciences), and seven-color flow CFSE labeling of donor cells. On day 8 after transfer, lymphocytes were cytometry was performed with LSR II (BD Biosciences). The data were isolated from the spleen of GvHD mice induced by Thy1.2 cells (wild-type acquired and analyzed on CellQuest software (version 3.3) for FACSCali- Ϫ/Ϫ ϫ 6 (WT) or CXCR6 ), and 20 10 cells were subsequently transferred to bur and on FACSDiVa software (version 2.1) and CellQuest-Pro software GvHD mice induced by Thy1.1 cells (short term recruitment mice). Six (version 4.0) for LSR II. Data are presented as the mean Ϯ SEM. Student’s hours after the second transfer, the short term recruitment mice were killed, t test was used for between-group comparisons, and significance was set and cells from the liver, spleen, and blood were harvested. This 6-h win- Ͻ

at p 0.05. by guest on October 2, 2021 dow allowed us to compare the recruitment of Thy1.2 cells with or without CXCR6 into inflamed livers. The phenotypes of the 20 ϫ 106 lymphocytes Migration assay CXCR6Ϫ/Ϫ injected from WT and mice were similar with respect to the ϩ Ϫ expression of activation markers, adhesion molecules, and cell subset dis- LIL were isolated from GvHD livers induced by WT, CXCR6 / , and Ϫ/Ϫ tribution described above (data not shown). A liver recruitment ratio was CXCR6 splenocytes at day 7 of disease. LIL (0.5–1 ϫ 106) were placed calculated as the absolute number of cells recruited to the liver divided by in the upper chamber of Transwell inserts (5-␮m pore size; Corning the number of injected cells. Short term recruitment was determined for Costar) and migrated to chemokines at their optimal concentrations for 2 h CD8ϩH-2DdϪThy1.2ϩ, CD4ϩH-2DdϪThy1.2ϩ, and CD19ϩH-2DdϪIgDbϩ at 37°C as described previously (21). After migration, polystyrene beads cell subsets. Additionally, the level of defined donor cell subsets circulating (Polysciences) were added to each bottom well as an internal standard, and in the blood was analyzed. an aliquot was collected for analysis by flow cytometry for calculation of the percent migrated. Triplicate wells were pooled and stained with Abs to Isolation of liver-infiltrating lymphocytes (LIL) H-2Dd and CD8 for calculating donor CD8 migration. The data presented are the percent migration of the indicated subsets, determined by measuring After mice were killed, the liver was perfused through the portal vein with the number of cells of each subset in the starting and migrated populations. 5 ml of PBS. The liver was harvested, dispersed through a stainless steel ␮ The chemokine CXCL12 was purchased from PeproTech, and CXCL16 mesh (70- m pore size), and washed with HBSS containing 10% bovine was obtained from R&D Systems. calf serum. The dispersed cells were mixed with a 33% Percoll (Amersham Biosciences) solution containing 100 U/ml heparin (Sigma-Aldrich) and centrifuged for 20 min (800 ϫ g) at room temperature. Next, the cell pellet Results was resuspended in double-distilled water for 20 s to lyse RBC, and the CXCR6 is up-regulated in GvHD hepatitis lysis was stopped with 2ϫ PBS. The isolated lymphocytes were then resuspended in HBSS (2% bovine calf serum), counted in a hemocytometer, We initially asked whether CXCR6 and CXCL16 are relevant can- and kept on ice until staining for flow cytometric analysis. didates to study in liver homing by examining their expression under normal and inflamed conditions. We examined livers from RT-PCR and real-time PCR control and/or GvHD mice 7 days after lymphocyte transfer. Total RNA was extracted from whole liver tissue using an RNeasy kit CXCR6 and CXCL16 mRNA were amplified by PCR from total according to the manufacturer’s instructions (Qiagen). RNA concentration RNA, and we found clear evidence of CXCR6 and CXCL16 and purity were determined by measuring absorbance spectrophotometri- mRNA expression in both healthy controls and GvHD animals cally at 260 and 280 nm. RT-PCR was performed with ThermoScript RT- PCR System (Invitrogen Life Technologies). Mouse ␤-actin served as an (data not shown). We also quantitatively analyzed the expression internal control gene. The PCR profile was 40 cycles of denaturation at of CXCR6 and CXCL16 and found an ϳ2-fold increase in mRNA 95°C for 30 s, annealing at 55°C for 30 s, and extension at 72°C for 1 min, expression in GvHD-inflamed liver compared with normal liver, as followed by a final extension of 7 min at 72°C. The primer sequences of assessed by real-time PCR (mean Ϯ SEM fold increase: CXCR6, CXCR6, CXCL16, and ␤-actin were: CXCR6 (forward), 5Ј-TAC GAT Ϯ Ϯ ϭ GGG CAC TAC GAG GGA G-3Ј; CXCR6 (reverse), 5Ј-GCA AAG AAA 1.6 0.1; CXCL16, 1.9 0.1; Fig. 1a; n 4). Next, we asked CCA ACA GGG AGA CCA C-3Ј; CXCL16 (forward), 5Ј-GCT TTG GAC whether CXCR6 was expressed on transferred lymphocytes infil- CCT TGT CTC TTG C-3Ј; CXCL16 (reverse), 5Ј-GTG CTG AGT GCT trating the liver. For this purpose, we used heterozygous GFP The Journal of Immunology 279 Downloaded from http://www.jimmunol.org/

bxd by guest on October 2, 2021 FIGURE 1. CXCR6/CXCL16 expression and in vitro migration to CXCL16. a, GVHD was induced in BDF1 (H-2 ) mice by the transfer of splenocytes b from C57BL/6 (H-2 ) mice. On day 7 of disease, livers were harvested from GvHD and normal BDF1 mice, mRNA was isolated, and real-time PCR was performed. Samples were normalized to 15Sa rRNA, and chemokine and chemokine receptor expression from normal liver was assigned a relative value Ϯ ϭ of 1.0. Data are expressed as the mean SEM (n 4). b, GvHD was induced in BDF1 (allogeneic) mice by transfer of splenocytes from heterozygous CXCR6-GFP (CXCR6ϩ/Ϫ, H-2b) mice. The same splenocytes were transferred into Thy1.1 (syngeneic, H-2b) mice as a negative control. GFP expression as a marker for CXCR6 expression was assessed by flow cytometry on donor LIL subsets. Representative histograms are shown. Histograms are scaled to more easily compare the percentage of positive lymphocytes. c, GFP (CXCR6) expression on donor lymphocyte subsets is compared in liver (u) and spleen (f). d, CXCR6 message was amplified from C57BL/6 (B6) and CXCR6-GFP (CXCR6Ϫ/Ϫ) livers. ␤-Actin message was amplified as a control. The primers used are described in Materials and Methods. e, Liver-infiltrating lymphocytes from day 7 of GVHD hepatitis induced by WT, CXCR6ϩ/Ϫ,or CXCR6Ϫ/Ϫ donor cells were migrated to CXCL16 and CXCL12. A representative experiment of three independent experiments is shown. knockin mice, in which one allele of the CXCR6 gene has been 1c), consistent with the possibility of up-regulation of CXCR6 in replaced by enhanced GFP, to examine the expression of CXCR6 secondary lymphoid organs and migration of these cells to effector in normal (syngeneic transfer) and inflamed (allogeneic transfer) sites. liver (17). We analyzed GFP expression in LIL from day 7 of GvHD induced by CXCR6-GFP (CXCR6ϩ/Ϫ) lymphocytes in al- Absence of CXCR6 results in reduced effector CD8 cell logeneic and syngeneic hosts. We observed an increase in the per- accumulation in liver centage of GFPhigh lymphocytes among donor CD4 and CD8 lym- We next asked whether the absence of CXCR6 would influence the phocytes in the liver of allogeneic vs syngeneic recipients. The efficiency of transferred effector T cells to accumulate in the liver percentages of CD8ϩGFPhigh donor lymphocytes among CD8 LIL in a model of GvHD. Liver infiltration of donor T cells in GvHD Ϯ Ϯ were 24 3 and 60 2% in syngeneic and allogeneic recipients, was induced by transfer of WT (B6 into BDF1) or CXCR6-defi- ϭ Ͻ Ϫ/Ϫ respectively (Fig. 1b; n 3–5; p 0.05). Similarly, the percent- cient (CXCR6 into BDF1) donor lymphocytes. We confirmed age of CD4ϩGFPhigh donor lymphocytes among CD4 LIL in- that these CXCR6 gene-targeted mice do not express CXCR6, as creased from 9 Ϯ 1% in syngeneic recipients to 42 Ϯ 2% in al- shown by RT-PCR (Fig. 1d), and thus constitute an ideal tool for logeneic recipients (Fig. 1b; n ϭ 3–5; p Ͻ 0.05). Because GFP is investigating the role of CXCR6. We also confirmed in vitro that under the control of the CXCR6 promoter, these findings suggest activated CXCR6Ϫ/Ϫ CD8 lymphocytes fail to migrate to the that donor lymphocytes up-regulate CXCR6 before or during ac- CXCR6 ligand CXCL16, but are still able to migrate to another cumulation in the liver. Furthermore, we observed that a larger chemokine, CXCL12, which was used as a positive control (Fig. percentage of the donor cells were GFPhigh in the GvHD liver as 1e). CXCR6ϩ/Ϫ mice have an intermediate phenotype (Fig. 1e). opposed to secondary lymphoid tissues such as the spleen (Fig. Subsequent comparisons were made between GvHD induced by 280 CXCR6 MEDIATES RECRUITMENT OF CD8ϩ LYMPHOCYTES

WT and CXCR6Ϫ/Ϫ splenocytes. In our in vivo analysis on day 7, Decreased accumulation of CXCR6Ϫ/Ϫ cells in liver is due to both groups of mice displayed marked hepatitis; the numbers of decreased recruitment 6 6 LIL were 6.1 Ϯ 1.1 ϫ 10 and 6.8 Ϯ 1.9 ϫ 10 cells/liver in Ϫ Ϫ The decreased accumulation of antihost CXCR6 / CD8 T cells GvHD generated by WT and CXCR6Ϫ/Ϫ lymphocytes, respec- could reflect decreased recruitment to the liver; however, altered tively. In comparison, syngeneic controls had 1.9 Ϯ 0.3 ϫ 106 proliferation or survival could also play a role. To differentiate cells/liver. Importantly, there were far fewer donor CD8 T cells in Ϫ Ϫ between these possibilities, we developed a short term recruitment GvHD induced by CXCR6 / cells (54% decrease in donor-de- ϩ assay that minimizes the potential for proliferative or survival sig- rived CD8 lymphocytes compared with WT GvHD, 172 Ϯ 18 ϫ nals to influence the accumulation of donor lymphocytes in the 103 vs 79 Ϯ 3 ϫ 103 (mean Ϯ SEM); n ϭ 7; p Ͻ 0.05; Fig. 2a). No difference was observed in the number of donor cells infiltrat- liver. With this assay we were able to track donor cells infiltrating Ϯ ϫ 6 CXCR6Ϫ/Ϫ Ϯ ϫ 6 liver and lymphoid organs during a 6-h period after injection by ing the spleen (B6, 1.7 0.3 10 ; , 1.7 0.4 10 ; dϪ ϩ Fig. 2b), suggesting that this reduction in liver CD8 cells is not a determining the number of donor H-2D Thy1.2 cells in various generalized phenomenon characterized by decreased donor cells in compartments. We chose 6 h for the short term recruitment assay all compartments. The difference in liver extended to a reduction in based on initial experiments that demonstrated the ability to detect the percentage of donor CD8 cells (39% reduction) among total migrating, short term recruitment donor cells in the liver in the CD8 cells (i.e., (donor CD8)/(total CD8) ϫ 100; B6, 38 Ϯ 5%; midst of on-going GvHD hepatitis (data not shown). Importantly, b Ϫ/Ϫ CXCR6Ϫ/Ϫ,23Ϯ 4%; Fig. 2c; p Ͻ 0.05). However, no statistically essentially all donor CD8 T cells (H-2D ), whether CXCR6 or significant difference was found in donor CD4 cells (Fig. 2d). Con- WT, were of the previously activated phenotype, as shown by the sidered together, the decrease in donor CXCR6Ϫ/Ϫ CD8 lympho- expression of (CD44high, CD45RBlow, and L-selectinlow); thus, the cytes in percentage and absolute numbers indicate that there was a donor CD8 T cells appeared to share a similar phenotype and state Downloaded from selective decrease in the CXCR6Ϫ/Ϫ cell subset, not only an overall of activation. Spleen cells from day 8 GvHD mice were injected decrease in total LIL. These findings stress the importance of i.v. into recipients (H-2Dbxd) also undergoing day 8 GvHD. After CXCR6 in the accumulation of pathological antihost CD8 T cells collecting cells from various tissues, we calculated a liver recruit- to a target organ of GvHD. ment ratio, defined as the absolute number of CD8 or CD4 short http://www.jimmunol.org/ by guest on October 2, 2021

FIGURE 2. Lymphocyte accumulation in tissues. The absolute numbers of donor CD8 lymphocytes (CD8ϩH-2Dd Ϫ in liver (a) and spleen (b) 7 days Ϫ/Ϫ after adoptive transfer are shown. GvHD was induced in BDF1 mice by transfer of splenocytes from WT and CXCR6-deficient (CXCR6 ) C57BL/6 mice. Transfer of syngeneic CFSE-labeled BDF1 splenocytes into BDF1 mice served as the negative control (Control). The absolute number of donor lymphocytes was calculated by multiplying the percentage of CD8ϩH-2Dd Ϫ(allogeneic) and CD8ϩCFSEϩ (syngeneic) cells in the small lymphocyte gate by the absolute number of lymphocytes recovered. The percentages of CD8 (c) and CD4 (d) lymphocytes of donor origin in liver and spleen were calculated by dividing the number of CD8ϩH-2Dd Ϫand CD4ϩH-2Dd Ϫlymphocytes by the total number of CD8ϩ and CD4ϩ lymphocytes in the respective tissues. The percentage of syngeneic donor CD8 lymphocytes was calculated by dividing the number of CD8ϩCFSEϩ by the total number of CD8ϩ lymphocytes in the respective tissues. Data are shown as the mean Ϯ SEM (n ϭ 7). p Ͻ 0.05, WT vs CXCR6Ϫ/Ϫ groups. The Journal of Immunology 281 term recruitment donor cells (CD8ϩH-2DdϪThy1.2ϩ or CD4ϩH- deficiency (data not shown). These data suggest that the reduction 2DdϪThy1.2ϩ) harvested from a specific compartment divided by of CXCR6Ϫ/Ϫ CD8 T cells in the liver is due to decreased recruit- the absolute number of such donor cells injected multiplied by 100. ment rather than altered proliferation. By performing this short term recruitment assay using donor lym- Reduced hepatic recruitment of CXCR6Ϫ/Ϫ CD8 lymphocytes phocytes from WT GvHD and CXCR6-deficient GvHD, we found could also reflect decreased availability in the blood. To address that the liver recruitment ratio of donor CXCR6Ϫ/Ϫ CD8 cells was this, we examined the percentage of relevant cells circulating in the decreased by 33% (B6, 12 Ϯ 2%; CXCR6Ϫ/Ϫ,8Ϯ 1%; n ϭ 8; p Ͻ blood and found that the levels of CXCR6Ϫ/Ϫ vs WT CD8 cells 0.05; Fig. 3a), which corresponds well with the 39% decrease in (CD8 donor cells from the second transfer, i.e., CD8ϩH-2DdϪ the percentage of donor CD8 lymphocytes infiltrating the liver as Thy1.2ϩ; Fig. 3b; n ϭ 8) in the blood of recipient mice 6 h after described above (Fig. 2c). These findings suggest that CXCR6 injection (i.e., at the end of the short term recruitment assays) were supports recruitment, rather than proliferation, of activated CD8 similar, suggesting that CXCR6 deficiency did not significantly cells in GvHD. It is noteworthy that the liver recruitment ratios of alter availability in the blood. other CXCR6-deficient donor cells, such as Thy1.2ϩCD4ϩ T lym- Although little proliferation is likely to occur during the 6-h phocytes and IgDbϩCD19ϩ naive B lymphocytes, were not dif- assay, to rule out any possible effect of proliferation on cell num- ferent from those of WT animals (Fig. 3a). IgDbϩCD19ϩ naive B bers, we compared the proliferation of CXCR6Ϫ/Ϫ vs WT donor T lymphocytes were used as an internal negative control, because cell populations by measuring the incorporation of BrdU into LIL. previous reports, which we confirmed in this study, have shown We injected BrdU i.p. 2 h before harvesting LIL from day 7 WT that naive B lymphocytes do not express CXCR6 (16, 17). More- and CXCR6-deficient GvHD mice. As shown above (Fig. 2a), the over, because naive B lymphocytes are a relatively rare population absolute number of CXCR6Ϫ/Ϫ donor lymphocytes infiltrating the in GvHD liver, their localization may be considered a basal or liver was again decreased (Fig. 4a), but the percentage of WT Downloaded from background level. Of note, the recruitment of lymphocytes to re- donor vs CXCR6Ϫ/Ϫ CD8 lymphocytes incorporating BrdU was cipient spleen, including CD8 cells, was not affected by CXCR6 almost identical (WT B6, 14 Ϯ 1%; CXCR6Ϫ/Ϫ,16Ϯ 3%; Fig. 4b; n ϭ 4–5), suggesting that the absence of CXCR6 does not affect the proliferative capacity of the donor CD8 cells in the liver. A third potential contributing factor to the accumulation of cells in the liver is apoptosis. It has been well established that a large http://www.jimmunol.org/ number of activated cells programmed for apoptosis accumulate in the liver (22). To assess the possibility that CXCR6/CXCL16 may be solely providing a survival signal in the liver and thus accumulating more cells in WT GvHD, we examined donor LIL for annexin V and 7-aminoactinomycin D (7-AAD) staining to assess the number of cells undergoing apoptosis. With our method of isolating LIL, we found few of the harvested cells to be undergoing apoptosis or necrosis and little difference between the WT and by guest on October 2, 2021 CXCR6-deficient donor cells in terms of their apoptotic markers (annexin, 3 Ϯ 0.6% of WT B6; 2 Ϯ 0.5% of CXCR6Ϫ/Ϫ ( p ϭ 0.17); 7-AAD, 2 Ϯ 0.2% of WT B6 vs 2 Ϯ 0.4% of CXCR6Ϫ/Ϫ ( p ϭ 0.87); Fig. 4c; n ϭ 4–8). These data thus support an important role for CXCR6 in CD8 cells homing into the liver.

Discussion In this study we examined the in vivo role of the chemokine receptor CXCR6 in a model of lymphocyte recruitment to the liver during GvHD hepatitis. First, we examined the expression of CXCR6 on LIL using mice heterozygous for a GFP knockin mu- tation; the endogenous CXCR6 promoter drives the expression of GFP, allowing the detection of CXCR6ϩ cells by green fluores- cence (17). Using these splenocytes as donors, we found few T cells infiltrating the liver under normal conditions over a 1-wk period, in contrast to the significant increase observed under inflamed conditions, concomitant with an increase in the percentage of lymphocytes expressing CXCR6. The increase in frequency of ϩ FIGURE 3. Short term recruitment of donor lymphocytes in the liver. CXCR6 T cells was most prominent in target tissues of inflam- GvHD was induced by adoptively transferring WT B6-Thy1.2, CXCR6Ϫ/Ϫ- mation (i.e., liver) compared with secondary lymphoid organs (i.e.,

Thy1.2 and B6-Thy1.1 (congenic) lymphocytes into separate BDF1 mice. On spleen). This is consistent with past findings showing that CXCR6 day 8 after adoptive transfer, lymphocytes from WT and CXCR6-deficient is up-regulated upon activation and is found in abundance on T Ϫ/Ϫ (CXCR6 ) GvHD mice were transferred into congenic GvHD mice, and cells in Th1-mediated inflammation as in hepatitis C-induced (9, liver lymphocytes were isolated 6 h later. A liver recruitment ratio was cal- 18) and alcohol-induced cirrhosis and rheumatoid arthritis (9). culated for different cell subsets (CD8, CD4, and CD19) in the liver: [cell Next, we examined expression of the ligand, CXCL16, and found subset]recruited/[cell subset]injected. b, Circulating levels of donor lymphocytes on day 8 after adoptive transfer. Blood was harvested by cardiac puncture in it under both normal and inflamed conditions in the liver by RT- anesthetized, short term recruitment mice, stained, and analyzed by flow cy- PCR. Our finding of CXCL16 up-regulation of mRNA expression tometry for the percentage of specific cell subsets (CD8ϩH-2DdϪThy1.2ϩ or during GvHD is consistent with the findings of others who have CD4ϩH-2DdϪThy1.2ϩ) among leukocytes. Data are shown as the mean Ϯ reported up-regulation during inflammation in the liver (16, 19) SEM (n ϭ 8). p Ͻ 0.05. and in heart valves (23). CXCL16 is also observed in normal bone 282 CXCR6 MEDIATES RECRUITMENT OF CD8ϩ LYMPHOCYTES

specific decrease in the pathologically relevant cell population. Be- cause this observation was made 1 wk after donor cell transfer, we could not exclude the possibility that CXCR6 was required for lymphocyte proliferation or amplification of the inflammatory reaction. We therefore examined the role of CXCR6 in recruitment more closely using a short term recruitment assay. By focusing our attention on a 6-h window, we were able to minimize the effect of proliferation and survival while concentrating on actual homing. With this short term recruitment assay we again found a subset- specific decrease in the donor CD8 T cells, demonstrating a specific role for CXCR6 in activated CD8 T cell localization from blood into the inflamed liver. A role for CXCR6 and CXCL16 in cell recruitment from blood has been suggested previously based on their ability to mediate chemotaxis (13, 16) and firm adhesion (24, 25) of different cell populations in vitro. In its soluble form, CXCL16 could be trans- ported and presented luminally by endothelial cells to mediate chemokine-dependent activation of surface integrins on lymphocytes. ␣ ␤ ϩ Indeed, CXCL16-dependent firm adhesion of 4 1 lymphocytes on VCAM-1ϩ endothelial cells has been suggested to be important Downloaded from in firm adhesion of lymphocytes to neocapillaries of infected cardiac heart valves (23). In its transmembrane form, CXCL16 may be able to function as a structural adhesion molecule, anchoring activated CD8 lymphocytes, which may promote their retention in

the site of inflammation (24, 25). CXCL16 joins fractalkine (CX3C ligand 1) as the only other transmembrane chemokine that can http://www.jimmunol.org/ function in both its soluble and its membrane-bound form to regulate cell migration and adhesion (26, 27). A role for CXCR6/CXCL16 in providing a survival signal could also be put forth, because CXCL16 is up-regulated on DC under activating conditions (16) with a concomitant increase in CXCR6 on cocultured T cells (9). Other chemokines, such as stromal cell- derived factor-1, have been shown to be survival factors for CD4 T cells (28). However, in our analysis of markers of proliferation by guest on October 2, 2021 (BrdU incorporation) and apoptosis (annexin V and 7-AAD) on donor CD8 LIL, we found no significant difference in the percentage of positive cells, suggesting no proliferative or survival ad- vantage of WT cells over CXCR6-deficient cells. Our current studies have concentrated on short term recruitment, but additional studies will be necessary to critically examine possible additional roles of CXCR6/CXCL16 in the liver. One factor we have not been able to address directly in the current studies is the rate of exit of FIGURE 4. Proliferation and apoptosis of donor lymphocytes in the homed cells from the liver. Detailed studies conducted in sheep liver. GvHD was induced by adoptively transferring wild-type B6-Thy1.2 Ϫ Ϫ estimate the total output of lymphocytes from the liver through the and CXCR6 / -Thy1.2 lymphocytes into separate BDF mice. On day 7, 1 ϳ ϫ 6 liver lymphocytes were isolated, and BrdU incorporation was determined afferent lymphatics to be 1 10 cells/h under normal condi- in LIL by flow cytometry. a, A representative histogram showing BrdU tions (29). However, under activated conditions (i.e., after the ad- staining of donor CD8 lymphocytes in the liver from wild-type (gray line) ministration of i.v. endotoxin), afferent lymphatic flow from the and CXCR6Ϫ/Ϫ (solid line) GvHD. b, Percentage of BrdU-positive donor liver to the liver-draining lymph node peaked at 6 h, but cellular CD8 lymphocytes is shown for WT (Ⅺ) and CXCR6Ϫ/Ϫ (f) lymphocytes. output decreased until 6 h and subsequently peaked by 18 h. Fur- c, Percentage of annexin V- and 7-AAD-positive donor CD8 lymphocytes thermore, cell transfer experiments of alloantigen-stimulated lym- Ϫ Ϫ is shown for WT (Ⅺ) and CXCR6 / (f) lymphocytes. Data are shown as phocytes into normal rats showed progressive accumulation of the mean Ϯ SEM (n ϭ 4–8). Annexin, p ϭ 0.17; 7-AAD, p ϭ 0.87. cells in the liver from 6–9 h, but peak accumulation in the liver- draining lymph node occurred later, at 15 h (30). The results of both studies suggest that at 6 h after injection, the liver is rapidly marrow, where it may participate in the homing of plasma cells to accumulating cells, and cellular exit is likely to have little effect on the bone marrow (24). cell numbers at this early time point. After confirming the expression of the receptor-ligand pair in The residual accumulation of CXCR6Ϫ/Ϫ CD8 cells in the liver GVHD hepatitis, we examined the functional significance of the indicates that other mechanisms compensate for the CXCR6 def- pair by focusing on the natural course of GvHD in mice induced by icit. Other investigators have suggested a role of CCR5 in medi- transferred CXCR6-deficient cells. Transferring splenocytes from ating recruitment of CD8 donor T lymphocytes to the liver in homozygous CXCR6-GFP mice, we examined the severity of dis- GvHD (31–33). Their studies examined total lymphocyte accumu- ease by analyzing the frequency of donor cells infiltrating the liver. lation at different time points, but did not exclude a potential con- We found that the lack of CXCR6 on donor cells resulted in fewer tribution of concomitant effects on proliferation; nevertheless, CD8 LIL, in both percentage and absolute numbers, suggesting a CCR5 and CXCR6 may both participate in the accumulation of The Journal of Immunology 283

CD8 lymphocytes in GvHD-induced hepatitis. Indeed, CCR5-, 14. Deng, H. K., D. Unutmaz, V. N. KewalRamani, and D. R. Littman. 1997. Ex- CXCR3-, and CXCR6-expressing lymphocytes have been found in pression cloning of new receptors used by simian and human immunodeficiency viruses. Nature 388:296. liver infiltrates of patients with hepatitis C (18, 34), and combina- 15. Liao, F., G. Alkhatib, K. W. Peden, G. Sharma, E. A. Berger, and J. M. Farber. tions of these three chemokine receptors can be coexpressed on 1997. STRL33, A novel chemokine receptor-like protein, functions as a fusion cofactor for both macrophage-tropic and T cell line-tropic HIV-1. J. Exp. Med. activated T cells (8, 17). Experiments have also supported an im- 185:2015. portant role of CXCR3 in many target organs of GvHD, including 16. Matloubian, M., A. David, S. Engel, J. E. Ryan, and J. G. Cyster. 2000. A the liver, as shown by improved liver histology in GvHD induced transmembrane CXC chemokine is a ligand for HIV-coreceptor Bonzo. Nat. Im- munol. 1:298. by CXCR3-deficient lymphocytes compared with WT (35). Be- 17. Unutmaz, D., W. Xiang, M. J. Sunshine, J. Campbell, E. Butcher, and yond being partially redundant for recruitment into the liver, mul- D. R. Littman. 2000. The primate lentiviral receptor Bonzo/STRL33 is coordi- tiple chemokine receptors may also play a role in navigation within nately regulated with CCR5 and its expression pattern is conserved between human and mouse. J. Immunol. 165:3284. the liver. Others have stressed the concept of a “symphony” of 18. Boisvert, J., E. J. Kunkel, J. J. Campbell, E. B. Keeffe, E. C. Butcher, and chemokines mediating an inflammatory process (36) and have em- H. B. Greenberg. 2003. Liver-infiltrating lymphocytes in end-stage hepatitis C phasized that combinations of chemokine receptors, rather than virus: subsets, activation status, and chemokine receptor phenotypes. J. Hepatol. 38:67. individual receptors, act together to define the trafficking of dif- 19. Dong, H., N. Toyoda, H. Yoneyama, M. Kurachi, T. Kasahara, Y. Kobayashi, ferentiated effector/memory lymphocyte populations (8). H. Inadera, S. Hashimoto, and K. Matsushima. 2002. Gene expression profile In conclusion, our novel data demonstrate that CXCR6 is an analysis of the mouse liver during bacteria-induced fulminant hepatitis by a cDNA microarray system. Biochem. Biophys. Res. Commun. 298:675. important mediator of the accumulation of activated CD8 T cells in 20. Via, C. S., S. O. Sharrow, and G. M. Shearer. 1987. Role of cytotoxic T lym- the liver in GVHD-induced hepatitis. The reduced accumulation of phocytes in the prevention of lupus-like disease occurring in a murine model of CXCR6Ϫ/Ϫ CD8 cells appears to reflect, at least in large part, a graft-vs-host disease. J. Immunol. 139:1840. 21. Campbell, J. J., E. P. Bowman, K. Murphy, K. R. Youngman, M. A. Siani, Downloaded from significant defect in their ability to migrate into inflamed liver from D. A. Thompson, L. Wu, A. Zlotnik, and E. C. Butcher. 1998. 6-C-kine (SLC), blood. Thus, we conclude that CXCR6 plays an important role in a lymphocyte adhesion-triggering chemokine expressed by high endothelium, is the hepatic homing and accumulation of activated CD8 T lympho- an agonist for the MIP-3␤ receptor CCR7. J. Cell Biol. 141:1053. 22. Huang, L., G. Soldevila, M. Leeker, R. Flavell, and I. N. Crispe. 1994. The liver cytes and may act in concert with other chemokine receptors to eliminates T cells undergoing antigen-triggered apoptosis in vivo. Immunity orchestrate pathological inflammation in the liver. 1:741. 23. Yamauchi, R., M. Tanaka, N. Kume, M. Minami, T. Kawamoto, K. Togi, T. Shimaoka, S. Takahashi, J. Yamaguchi, T. Nishina, et al. 2004. Upregulation http://www.jimmunol.org/ Acknowledgments of SR-PSOX/CXCL16 and recruitment of CD8ϩ T cells in cardiac valves during We thank C. Crumpton and L. Rott for their expertise at the flow cytometry inflammatory valvular heart disease. Arterioscler. Thromb. Vasc. Biol. 24:282. core facility, L. Ohler and G. Sperinde for their expertise in real-time PCR, 24. Nakayama, T., K. Hieshima, D. Izawa, Y. Tatsumi, A. Kanamaru, and O. Yoshie. 2003. Cutting edge: profile of chemokine receptor expression on human plasma and C. Arnold and G. Debes for their constructive criticism and sugges- cells accounts for their efficient recruitment to target tissues. J. Immunol. tions on the paper. 170:1136. 25. Shimaoka, T., T. Nakayama, N. Fukumoto, N. Kume, S. Takahashi, J. Yamaguchi, M. Minami, K. Hayashida, T. Kita, J. Ohsumi, et al. 2004. Cell References surface-anchored SR-PSOX/CXC chemokine ligand 16 mediates firm adhesion 1. Ferrara, J. L. M., H. J. Deeg, and S. J. Burakoff. 1997. Graft-vs.-Host Disease. of CXC chemokine receptor 6-expressing cells. J. Leukocyte Biol. 75:267. Marcel Dekker, New York. 26. Imai, T., K. Hieshima, C. Haskell, M. Baba, M. Nagira, M. Nishimura,

2. Butcher, E. C., and L. J. Picker. 1996. Lymphocyte homing and homeostasis. M. Kakizaki, S. Takagi, H. Nomiyama, T. J. Schall, et al. 1997. Identification and by guest on October 2, 2021 Science 272:60. molecular characterization of fractalkine receptor CX3CR1, which mediates both 3. Constantin, G., M. Majeed, C. Giagulli, L. Piccio, J. Y. Kim, E. C. Butcher, and leukocyte migration and adhesion. Cell 91:521. ␤ C. Laudanna. 2000. Chemokines trigger immediate 2 integrin affinity and mo- 27. Fong, A. M., L. A. Robinson, D. A. Steeber, T. F. Tedder, O. Yoshie, T. Imai, and bility changes: differential regulation and roles in lymphocyte arrest under flow. D. D. Patel. 1998. Fractalkine and CX3CR1 mediate a novel mechanism of leu- Immunity 13:759. kocyte capture, firm adhesion, and activation under physiologic flow. J. Exp. 4. Zlotnik, A., and O. Yoshie. 2000. Chemokines: a new classification system and Med. 188:1413. their role in immunity. Immunity 12:121. 28. Suzuki, Y., M. Rahman, and H. Mitsuya. 2001. Diverse transcriptional response 5. Kunkel, E. J., and E. C. Butcher. 2002. Chemokines and the tissue-specific mi- of CD4ϩ T cells to stromal cell-derived factor (SDF)-1: cell survival promotion gration of lymphocytes. Immunity 16:1. and priming effects of SDF-1 on CD4ϩ T cells. J. Immunol. 167:3064. 6. McDonald, G. B., H. M. Shulman, K. M. Sullivan, and G. D. Spencer. 1986. 29. Young, A. J., G. M. Hare, and J. B. Hay. 1994. Blood-to-lymph migration of Intestinal and hepatic complications of human bone marrow transplantation. Part small lymphocytes through the liver of the sheep. Hepatology 19:758. I. Gastroenterology 90:460. 30. Smith, M. E., A. F. Martin, and W. L. Ford. 1980. Migration of lymphoblasts in 7. Rus, V., A. Svetic, P. Nguyen, W. C. Gause, and C. S. Via. 1995. Kinetics of Th1 the rat: preferential localization of DNA-synthesizing lymphocytes in particular and Th2 cytokine production during the early course of acute and chronic murine ϩ lymph nodes and other sties. Monogr. Allergy 16:203. graft-versus-host disease: regulatory role of donor CD8 T cells. J. Immunol. 155:2396. 31. Murai, M., H. Yoneyama, T. Ezaki, M. Suematsu, Y. Terashima, A. Harada, 8. Kim, C. H., L. Rott, E. J. Kunkel, M. C. Genovese, D. P. Andrew, L. Wu, and H. Hamada, H. Asakura, H. Ishikawa, and K. Matsushima. 2003. Peyer’s patch E. C. Butcher. 2001. Rules of chemokine receptor association with T cell polar- is the essential site in initiating murine acute and lethal graft-versus-host reaction. ization in vivo. J. Clin. Invest. 108:1331. Nat. Immunol. 4:154. 32. Murai, M., H. Yoneyama, A. Harada, Z. Yi, C. Vestergaard, B. Guo, K. Suzuki, 9. Kim, C. H., E. J. Kunkel, J. Boisvert, B. Johnston, J. J. Campbell, ϩ ϩ M. C. Genovese, H. B. Greenberg, and E. C. Butcher. 2001. Bonzo/CXCR6 H. Asakura, and K. Matsushima. 1999. Active participation of CCR5 CD8 T expression defines type 1-polarized T-cell subsets with extralymphoid tissue lymphocytes in the pathogenesis of liver injury in graft-versus-host disease. homing potential. J. Clin. Invest. 107:595. J. Clin. Invest. 104:49. 10. Kim, C. H., K. Nagata, and E. C. Butcher. 2003. Dendritic cells support sequen- 33. Serody, J. S., S. E. Burkett, A. Panoskaltsis-Mortari, J. Ng-Cashin, E. McMahon, tial reprogramming of chemoattractant receptor profiles during naive to effector G. K. Matsushima, S. A. Lira, D. N. Cook, and B. R. Blazar. 2000. T-lymphocyte production of macrophage inflammatory protein-1␣ is critical to the recruitment T cell differentiation. J. Immunol. 171:152. ϩ 11. Langenkamp, A., K. Nagata, K. Murphy, L. Wu, A. Lanzavecchia, and of CD8 T cells to the liver, lung, and spleen during graft-versus-host disease. F. Sallusto. 2003. Kinetics and expression patterns of chemokine receptors in Blood 96:2973. human CD4ϩ T lymphocytes primed by myeloid or plasmacytoid dendritic cells. 34. Shields, P. L., C. M. Morland, M. Salmon, S. Qin, S. G. Hubscher, and Eur. J. Immunol. 33:474. D. H. Adams. 1999. Chemokine and chemokine receptor interactions provide a 12. Calabresi, P. A., S. H. Yun, R. Allie, and K. A. Whartenby. 2002. Chemokine mechanism for selective T cell recruitment to specific liver compartments within receptor expression on MBP-reactive T cells: CXCR6 is a marker of IFN␥-pro- hepatitis C-infected liver. J. Immunol. 163:6236. ducing effector cells. J. Neuroimmunol. 127:96. 35. Duffner, U., B. Lu, G. C. Hildebrandt, T. Teshima, D. L. Williams, P. Reddy, 13. Wilbanks, A., S. C. Zondlo, K. Murphy, S. Mak, D. Soler, P. Langdon, R. Ordemann, S. G. Clouthier, K. Lowler, C. Liu, et al. 2003. Role of CXCR3- D. P. Andrew, L. Wu, and M. Briskin. 2001. Expression cloning of the STRL33/ induced donor T-cell migration in acute GVHD. Exp. Hematol. 31:897. BONZO/TYMSTRligand reveals elements of CC, CXC, and CX3C chemokines. 36. Gerard, C., and B. J. Rollins. 2001. Chemokines and disease. Nat. Immunol. J. Immunol. 166:5145. 2:108.