CCL25 Mediates the Localization of Recently Activated Cd8αβ+ Lymphocytes to the Small-Intestinal Mucosa

CCL25 mediates the localization of recently activated CD8αβ+ lymphocytes to the small-intestinal mucosa

Marcus Svensson, … , Gabriel Márquez, William W. Agace

J Clin Invest. 2002;110(8):1113-1121. https://doi.org/10.1172/JCI15988.

Article Aging

The recruitment of antigen-specific T lymphocytes to the intestinal mucosa is central to the development of an effective mucosal immune response, yet the mechanism by which this process occurs remains to be fully defined. Here we show + + that the CC chemokine receptor 9 (CCR9) is selectively and functionally expressed on murine αEβ7 naive CD8αβ lymphocytes and a subset of recently activated CD69+ CD8αβ+ lymphocytes. Using a T cell receptor transgenic transfer model, we demonstrate that CCR9 expression is functionally maintained on CD8αβ+ lymphocytes following activation in mesenteric lymph nodes but rapidly downregulated on CD8αβ+ lymphocytes activated in peripheral lymph nodes. These recently activated CCR9+ CD8αβ+ lymphocytes selectively localized to the small-intestinal mucosa, and in vivo neutralization of the CCR9 ligand, CCL25, reduced the ability of these cells to populate the small-intestinal epithelium. Together these results demonstrate an important role for chemokines in the localization of T lymphocytes to the small- intestinal mucosa and suggest that targeting CCL25 and/or CCR9 may provide a means to selectively modulate small- intestinal immune responses.

Find the latest version: https://jci.me/15988/pdf CCL25 mediates the localization Rapid Publication of recently activated See the related Commentary beginning on page 1079. CD8αβ+ lymphocytes to the small-intestinal mucosa

Marcus Svensson,1 Jan Marsal,1 Anna Ericsson,1 Laura Carramolino,2 Therese Brodén,1 Gabriel Márquez,2 and William W. Agace1

1Immunology Section, Department of Cell and Molecular Biology, Lund University, Lund, Sweden 2Departmento de Inmunología, Centro Nacional de Biotecnología, Universidad Autónoma de Madrid, Madrid, Spain

The recruitment of antigen-specific T lymphocytes to the intestinal mucosa is for an involvement of chemokines central to the development of an effective mucosal immune response, yet the and their receptors in the recruit- mechanism by which this process occurs remains to be fully defined. Here we ment of activated lymphocyte sub- show that the CC chemokine receptor 9 (CCR9) is selectively and functionally sets to effector tissues (5). For exam- + + expressed on murine αEβ7 naive CD8αβ lymphocytes and a subset of recent- ple, the CC chemokine receptor 4 ly activated CD69+ CD8αβ+ lymphocytes. Using a T cell receptor transgenic (CCR4) and the CCR10 ligand, transfer model, we demonstrate that CCR9 expression is functionally main- CCL27, were recently shown to con- tained on CD8αβ+ lymphocytes following activation in mesenteric lymph tribute to lymphocyte recruitment to nodes but rapidly downregulated on CD8αβ+ lymphocytes activated in periph- inflamed skin (6, 7). The chemokine eral lymph nodes. These recently activated CCR9+ CD8αβ+ lymphocytes selec- receptor CCR9 is selectively and tively localized to the small-intestinal mucosa, and in vivo neutralization of the functionally expressed on human CCR9 ligand, CCL25, reduced the ability of these cells to populate the small- small-intestinal lymphocytes (8), and intestinal epithelium. Together these results demonstrate an important role its ligand, CCL25, is constitutively for chemokines in the localization of T lymphocytes to the small-intestinal expressed by murine and human mucosa and suggest that targeting CCL25 and/or CCR9 may provide a means small-intestinal epithelial cells (9, to selectively modulate small-intestinal immune responses. 10), indicating a potential role for This article was published online in advance of the print edition. this chemokine receptor/chemokine The date of publication is available from the JCI website, http://www.jci.org. pair in lymphocyte localization to the J. Clin. Invest. 110:1113–1121 (2002). doi:10.1172/JCI200215988. small intestine. However, examination of CD8αβ+ lymphocyte numbers in the small-intestinal epithelium of Introduction recruitment of T lymphocytes to intes- CCR9–/– mice has yielded conflicting The intestinal mucosa is continually tinal effector sites is thought to play a results (11, 12). Furthermore, small- exposed to a large array of foreign anti- critical role in this process. intestinal epithelial cells constitu- gens and must respond appropriately to Following activation in secondary tively express a number of maintain mucosal integrity while at the lymphoid organs, T lymphocytes gain chemokines with activity for T lym- same time mounting effective immune the ability to migrate from the blood phocytes, including CXCL12, CCL28, responses to potential pathogens. The to tertiary effector tissues such as the and CX3CL1, indicating a potential intestine and skin (1). Subsets of pre- role for other chemokines in this Received for publication May 23, 2002, and viously activated T lymphocytes dis- process (13–15). Thus the in vivo role accepted in revised form August 20, 2002. play selective tissue tropism for these of CCL25 and CCR9 in T lymphocyte Address correspondence to: William W. sites, a process that is controlled by recruitment to the small intestine Agace, Immunology Section, Department of specific combinations of cell adhesion remains unclear. Cell and Molecular Biology, Lund University, molecules (1, 2). Previously activated In the current study we have exam- BMC I-13, S-22184 Lund, Sweden. T lymphocytes homing to the intes- ined expression and regulation of Phone: 46-46-2220416; Fax: 46-46-2224218; + E-mail: [email protected]. tine express high levels of α4β7 inte- CCR9 on murine CD8αβ lympho- Conflict of interest: No conflict of interest grin, whose ligand, MAdCAM-1, is cytes in vivo and determined the role has been declared. expressed on postcapillary venules in of CCL25 in the recruitment of recent- Nonstandard abbreviations used: CC the intestinal lamina propria. Indeed, ly activated CD8αβ+ lymphocytes to chemokine receptor (CCR); ovalbumin (OVA); 5- and 6-carboxy-fluorescein diacetate β7 integrin appears to be critical for the small-intestinal mucosa. succinimidyl ester (CFSE); intraepithelial the entry of previously activated T lymphocyte (IEL); lamina propria lymphocyte lymphocytes into the intestinal lami- Methods (LPL); lymph node (LN); mesenteric lymph na propria and epithelium (3, 4). Mice. C57BL/6J-Ly5.1 mice were node (MLN); peripheral lymph node (PLN); T cell receptor (TCR); gut-associated In addition to cell adhesion mole- obtained from Charles River Labora- lymphoid tissue (GALT). cules, accumulating evidence exists tories (Wilmington, Massachusetts,

The Journal of Clinical Investigation | October 2002 | Volume 110 | Number 8 1113 USA). OT-1 mice were kindly provided 5- and 6-carboxy-fluorescein diac- incubated with fluorochrome-conju- by A. Mowat (Glasgow, United King- etate succinimidyl ester (CFSE) stain- gated or biotin-conjugated antibody dom). All mice were maintained at the ing, cells (107 cells/ml) were incu- at saturating concentrations for 20 animal facility at the Department of bated with 1 µM CFSE (Molecu- minutes at 4°C. For biotinylated Microbiology, Immunology and Gly- lar Probes Inc., Eugene, Oregon, antibodies, stained cells were washed cobiology, Lund University, and were USA) for 8 minutes at room temp- and fluorochrome-conjugated strep- used between 8 and 14 weeks of age. erature in PBS. tavidin was added to the cells for 20 Antibodies and reagents. Anti-CD8β Lymphocyte isolation. Small-intestin- minutes at 4°C. For CCR9 staining, (53-5.8), anti-Ly5.2 (104), anti- al and colonic intraepithelial lym- 10% normal goat serum was included CD62L (Mel-14), anti-CD44 (IM7), phocytes (IELs) were isolated as prein the initial blocking buffer. Cells anti-CD4 (RM4-5), and anti-CD69 viously described (17) or from the cell were then incubated with polyclonal (H1.2F3) antibodies and strepta- suspensions obtained following rabbit anti–mouse CCR9 antibody vidin-allophycocyanin were from EDTA treatment of intestinal tissue for 45 minutes at 4°C, washed, and Pharmingen (San Diego, California, (see below). Small-intestinal lamina incubated with biotinylated goat USA). Anti-αE (M290) and anti-β7 propria lymphocytes (LPLs) were iso- anti-rabbit antibody for 20 minutes (FIB-504) antibodies were kindly pro- lated essentially as previously at 4°C. Finally, cells were incubated vided by C. Parker (Dana-Farber Can- described (18). Briefly, intestinal with fluorochrome-conjugated strep- cer Institute, Boston, Massachusetts, pieces were incubated for 4 × 20 to tavidin for 20 minutes at 4°C. To USA). Hybridomas producing anti- 6 × 20 minutes in calcium- and mag- check staining specificity, anti-CCR9 CD8α (YTS 169-4), anti-CD4 nesium-free (CMF) HBSS supple- antibody was preincubated with the (GK1.5), anti-B220 (RA3.6B2), anti– mented with 10% FCS and 10 mM immunizing peptide used for genera- FcRII/III (2.4G2), and anti–MHC-II HEPES (GIBCO BRL; Life Technolo- tion of the CCR9 antibody (PTELT- (M5/114) antibodies were from gies, Paisley, Scotland, United King- SLIPGMFDDFSYDST; 5 µg/ml) for American Type Culture Collection dom) (HBSS-10), containing 5 mM 15 minutes at room temperature (Rockville, Maryland, USA). Goat EDTA at 37°C on an orbital shaker prior to use. Samples were analyzed anti-rabbit Ig was from Jackson (125 rpm). For isolation of colonic on a flow cytometer (FACSCalibur; ImmunoResearch Laboratories Inc. IEL, 1mM DTT was included with BD Biosciences Europe, Erem- (West Grove, Pennsylvania, USA), the EDTA. The remaining tissue bodegem, Belgium) using CellQuest polyclonal rabbit anti–mouse CCR9 pieces were washed free of EDTA and software (BD Biosciences). antibody was from G. Márquez digested with collagenase type VIII Chemotaxis. Chemotaxis assays and (K629; ref. 16), 7-amino-actinomycin (100 U/ml; Sigma-Aldrich) in RPMI analysis of lymphocyte subset migra- D (7AAD) was from Sigma-Aldrich supplemented with 10% FCS and 10 tion were performed as previously (Steinheim, Germany), and recombi- mM HEPES (RPMI-10) at 37°C for 1 described (19). The analysis of lym- nant murine CCL25 was from R&D hour, twice. For isolation of lympho- phocyte subset migration was per- Systems Europe (Abingdon, United cytes from inflamed skin, ears were formed as previously described (8). To Kingdom). Murine stromal cell- split into dorsal and ventral sheets, determine OT-1 cell chemotaxis, we derived–factor 1 was a kind gift from treated for 20 minutes with EDTA depleted CD4+, B220+, and MHC-II+ I. Clark-Lewis (University of British twice, and then digested with colla- cells from LN cells by magnetic cell Columbia, Vancouver, British Co- genase for 1 hour three times as sorting beads (Miltenyi Biotec lumbia, Canada). above. For isolation of liver and lung GmbH, Bergisch Gladbach, Ger- Adoptive transfers. We injected 2 × 106 lymphocytes, organs were perfused many) according to the manufactur- to 3 × 106 CD8+ OT-1 cells intra- with 5 ml of cold PBS containing 75 er’s instructions prior to adding the venously into C57BL/6J-Ly5.1 mice. U/ml heparin prior to removal. Tis- cells to chemotaxic wells. Two to three days later, we injected sues were then minced, and the Real-time PCR. RNA was extracted mice intraperitoneally with 5 mg resulting cell suspension was filtered from tissues using the Absolutely ovalbumin (OVA, grade VI; Sigma- through a 70-µm cell strainer RNA kit (Stratagene, La Jolla, Cali- Aldrich) with or without 100 µg LPS (Costar, Corning Inc., Corning, New fornia, USA) according to the manu- (Escherichia coli, serotype 055:B5; York, USA). All tissue lymphocytes facturer’s instructions, and cDNA Sigma-Aldrich). Mice were sacrificed were enriched on 40/70 Percoll gradi- was prepared from RNA using Super- 2–3 days after OVA/LPS injection, ents (Amersham Pharmacis Biotech, script RT II (GIBCO BRL; Life Tech- and tissues were collected for further Uppsala, Sweden). Isolation of lym- nologies). Quantitative detection of analysis. For antibody treatment phocytes from lymph nodes (LNs) CCL25 was performed using a Light- experiments, mice received anti- and spleen was performed according Cycler (Roche Diagnostics, Basel, CCL25 antibody (89818; R&D Sys- to standard techniques. Switzerland). Briefly, using LightCy- tems Inc.) or rat IgG2b isotype con- Flow cytometry. After incubation cler DNA Master Hybridization trol antibody (A95-1; Pharmingen) with blocking buffer (PBS, Probes kit (Roche Diagnostics), 500 intraperitoneally, 6 hours prior to anti–FcRII/III antibody [10 µg/ml], nM forward and reverse primers, 200 (100 µg) and 1 day (75 µg) and 2 days 2% FCS, and 0.05% azide) for 20 nM fluorescently labeled probes, 2 µl (50 µg) after OVA administration. For minutes at 4°C, lymphocytes were cDNA product, and RNase- and

1114 The Journal of Clinical Investigation | October 2002 | Volume 110 | Number 8 Figure 1 CCR9 expression on murine lymphocytes from normal mice. (a) CD8+ and CD4+ lymphocytes from MLNs were stained with anti-CCR9 antibody (open) or anti-CCR9 antibody that had been preincubated with the CCR9-immunizing peptide (shaded). Numbers represent percentage of CCR9+ cells with background staining removed. Background staining was consistently less than 2%. (b) CCR9+ CD8αβ+ lym- + + + – + + phocytes coexpress αEβ7 integrin. (c) CD8αβ αEβ7 , but not CD8αβ αEβ7 , lymphocytes migrate to CCL25. The ability of CD8αβ αEβ7 + – and CD8αβ αEβ7 lymphocytes to migrate to CCL25 (250 nM) or CXCL12 (100 nM) was determined in chemotaxis assays (see Methods). Results are mean ± SEM of quadruplicate wells from one experiment of two performed. *P < 0.0001. (d) CCR9 expression on CD8αβ+ lymphocyte subsets in MLNs. Numbers represent percentage of CCR9+ cells with background staining removed.

DNase-free water were added to the Results CCR9 was specific, since preincubation master mix to a final volume of 20 µl. Expression of CCR9 by murine T lymphocyte with the immunizing peptide reduced The following primers and probes subsets. Using an affinity-purified poly- staining to levels observed when no pri- were used: murine CCL25 primers: clonal antibody (K629) raised against a mary antibody was used (Figure 1a). + forward, 5′-CCGGCATGCTAGGAAT- peptide encoding amino acids 3–22 of Since CCR9 is expressed on αEβ7 TATCA-3′, and reverse, 5′-GGC- the N-terminal portion of murine CD8αβ+ human peripheral blood lym- ACTCCTCACGC-TTGTA-CT-3′; murine CCR9, Carramolino and colleagues phocytes (8), we examined expression of + + GAPDH primers: forward, 5′-CCTG- recently demonstrated low-level expres- CCR9 on murine αEβ7 CD8αβ lym- CACCACCAACTGCTTA-3′, and reverse sion of CCR9 on CD8+ but not on CD4+ phocytes. Almost all CCR9+ CD8αβ+ 5′-ATGACCTTGCCCACAGCCT-3′; lymphocytes isolated from murine LNs lymphocytes in the spleen, MLNs, and murine CCL25 probes: 3′ fluores- (16). In order to gain further insights PLNs coexpressed αEβ7 (Figure 1b, and cein-labeled, 5′-GGCGGAAGTAGAATC- into the expression of CCR9 on lym- data not shown). To determine whether TCACAGCACGTA-3′, and 5′ LC Red phocyte subsets, we used K629 in com- expression of CCR9 correlated with 640–labeled, 5′-TTGCAGCTTCCACT- bination with a more sensitive staining responsiveness to CCL25, the ability of + – + CACTTCCTGC-3′; GAPDH probes: 3′ procedure (see Methods). Under these splenic αEβ7 and αEβ7 CD8αβ lym- fluorescein-labeled, 5′-ATGACCACA- conditions, about 60% of mesenteric phocytes to migrate to CCL25 was GTCCATGCCATCACTGCC-3′, and 5′ lymph node (MLN) CD8αβ+ lympho- determined in chemotaxis assays (Fig- LC Red 640–labeled, 5′-CCCAGAAGA- cytes (mean 62.0%, SD 9.5, n = 19) and ure 1c). Results from this analysis + – CTGTCGATGGCCCCT-3′. All primers about 55% of peripheral lymph node demonstrated that αEβ7 but not αEβ7 and probes were designed in cooper- (PLN) (mean 53.0%, SD 9.7, n = 14) and CD8αβ+ lymphocytes migrated effi- – ation with O. Landt (TIB MOLBIOL, splenic (mean 55.3%, SD 10.5, n =8) ciently to CCL25. In contrast, αEβ7 Berlin, Germany). The amount of CD8αβ+ lymphocytes expressed CCR9 CD8αβ+ lymphocytes migrated more CCL25 and GAPDH was calculated (Figure 1a, and data not shown). In con- efficiently to CXCL12 (Figure 1c). Thus, from a standard curve of dilutions of trast, 3% of CD4+ lymphocytes in MLNs CCR9 is preferentially and functionally + + linearized plasmid encoding CCL25 and PLNs expressed CCR9 (Figure 1a, expressed on murine αEβ7 CD8αβ and GAPDH cDNA. and data not shown). The staining for lymphocytes.

The Journal of Clinical Investigation | October 2002 | Volume 110 | Number 8 1115 Figure 2 Regulation of CCR9 expression on antigen- specific CD8αβ+ lymphocytes following activation in vivo. (a) OT-1 (Ly5.2+) cells were injected into C57BL/6J-Ly5.1 mice, and their expression of CCR9 was determined by flow cytometry 3 days after intraperitoneal OVA administration. Numbers represent the percentage of OT-1 cells that are CCR9+ with background staining removed. Results are from one representative experiment of nine (MLN), seven (PLN), and two (spleen) experiments performed. (b) OT-1 cells from the MLNs of OVA-treated mice respond to CCL25. MLN and PLN lymphocytes were isolated 3 days after intraperitoneal administration of OVA, and the ability of OT-1 cells and endogenous CD8αβ+ lymphocytes to migrate to CCL25 (250 nM) was determined. Results are the mean ± SEM of four experiments with three mice per experiment. *P < 0.005.

Further phenotypic analysis of CCR9 expression similar to that of endogenous CD8αβ+ lymphocytes CD8αβ+ lymphocyte subsets in MLNs CD8αβ+ lymphocytes from wild-type from both MLNs and PLNs migrated and PLNs demonstrated that CCR9 mice (data not shown). To examine the to CCL25 (Figure 2b). was expressed on the majority of naive effect of activation on CCR9 expres- To determine the kinetics of CCR9 (CD62Lhi CD44lo) CD8αβ+ lympho- sion, mice were injected with OVA expression during the initial stages of cytes (Figure 1d, and data not shown), intraperitoneally, and the expression antigen-driven CD8αβ+ lymphocyte while CD62Lhi CD44hi CD8αβ+ lym- of CCR9 on OT-1 cells in MLNs, PLNs, proliferation, OT-1 cells were labeled phocytes were CCR9– (Figure 1d). Fur- and spleen was determined. As expect- with CFSE and injected into recipi- thermore, a small population of ed, administration of OVA led to a ent mice, and expression of CCR9 on CD62Llo CD44hi (previously activated) large increase in the total number of CFSE+ cells was determined (Figure and about 50% of CD69+ (recently OT-1 cells in LNs and spleen (data not 3). Since adjuvants are commonly activated) CD8αβ+ lymphocytes were shown). Six to twenty percent (mean used to boost the immune responses CCR9+ (Figure 1d). The CD44hi 11.5%, SEM 1.7, n = 9) of OT-1 cells to soluble antigens, OVA was coin- CCR9+ CD8αβ+ lymphocytes also expressed CCR9 in the MLNs 3 days jected with LPS to increase the num- expressed αEβ7, although some after OVA administration (Figure 2a). ber of OT-1 cells available for analy- hi + + CD44 αEβ7 CD8αβ lymphocytes This staining was specific, since prein- sis. Furthermore, to exclude the were CCR9– (Figure 1d). Expression of cubation of the anti-CCR9 antibody possibility that the CCR9+ OT-1 cells CCR9 on a subset of recently activat- with the immunizing peptide against in the MLNs 3 days after stimulation ed CD8αβ+ lymphocytes indicated which it was raised reduced staining to derived from another site, CCR9 that CCR9 may be differentially regu- background levels (data not shown). In expression was examined on OT-1 lated on CD8αβ+ lymphocytes follow- marked contrast, OT-1 cells in PLNs cells 2 days after OVA injection, ing their activation in vivo and that and spleen were CCR9– (Figure 2a). before these cells had started to recir- CCR9 may function in the localiza- The differential regulation of CCR9 in culate. In the absence of activation, tion of a subset of recently activated MLNs and PLNs was selective, since the percentage of CCR9+ OT-1 cells CD8αβ+ lymphocytes. CXCR3, whose ligands are upregulat- within MLNs and PLNs did not dif- CCR9 is selectively maintained on anti- ed at sites of inflammation, was fer significantly (Figure 3, a and b). gen-specific CD8αβ+ lymphocytes following induced on OT-1 cells in both MLNs Following activation, CCR9 expres- activation in MLNs. To further under- and PLNs (data not shown). To deter- sion was selectively maintained on a stand the mechanisms regulating mine whether CCR9 expression on subset of OT-1 cells in MLNs for at CCR9 expression on CD8αβ+ lympho- recently activated OT-1 cells in MLNs least five cell divisions (Figure 3, a cytes following activation in vivo, we was functional, the ability of OT-1 and b). Furthermore, occasionally the used a T cell receptor (TCR) transgenic cells from MLNs and PLNs to respond percentage of CCR9+ OT-1 cells at transfer model in which OVA-specific to CCL25 was determined in chemo- early divisions was higher than seen CD8αβ+ TCR transgenic OT-1 cells taxis assays (Figure 2b). OT-1 cells in unstimulated OT-1 cells (Figure were transferred to C57BL/6J-Ly5.1 from MLNs but not from PLNs of 3a), and the levels of CCR9 on CCR9+ mice (20). Initial studies confirmed OVA-treated mice migrated to CCL25 OT-1 cells were higher after activa- that OT-1 cells showed a pattern of in chemotaxis assays. In contrast, tion (Figure 3a). Together these

1116 The Journal of Clinical Investigation | October 2002 | Volume 110 | Number 8 results indicate that some de novo results demonstrate that CCR9 expressed on previously activated induction of CCR9 expression may expression is maintained on CD8αβ+ CD8αβ+ lymphocytes in the murine also occur on these cells during their lymphocytes following activation in small intestine. To determine whether activation in MLNs. In marked con- MLNs but downregulated on recently activated CCR9+ OT-1 cells trast, CCR9 expression was rapidly CD8αβ+ lymphocytes in PLNs. selectively localize to the small intes- lost on OT-1 cells following activa- CCR9 is selectively expressed on OT-1 tine, OT-1 cells were transferred to tion in PLNs (Figure 3, a and b). lymphocytes localizing to the small-intes- recipient mice, and the expression of Analysis of CCR9 expression on tinal epithelium. In humans, CCR9 is CCR9 on OT-1 cells localizing to the OT-1 cells 3 days after stimulation, at selectively expressed on all small- small intestine, liver, and lung follow- a time point when the large majority intestinal lymphocytes and about ing OVA stimulation was examined. of cells had undergone more than 30% of colonic lymphocytes, but not In unimmunized mice, OT-1 cells in five divisions (data not shown), on lymphocytes isolated from other the small intestine were barely demonstrated that CCR9 continued mucosal effector sites or sites of detectable (<0.2% of CD8β+ cells; see to be expressed on a large subset of inflammation (10). We therefore Figure 5e), consistent with the OT-1 cells activated in MLNs but was determined expression of CCR9 on requirement for activation for entry absent on cells activated in PLNs CD8αβ+ lymphocytes present in of these cells to effector sites. After (Figure 3c). Of note, the percentage murine effector sites (Figure 4a). administration of OVA, 57% (SD 6.6, of OT-1 cells expressing CCR9 in About 40% of CD8αβ+ small-intestin- n = 5) of OT-1 cells isolated from the MLNs was greater following stimula- al IELs (mean 39.6%, SD 14.7, n = 9) small-intestinal epithelium and 36% tion with OVA plus LPS than with and about 50% of CD8αβ+ small- (n = 1) of OT-1 cells in the small-intes- OVA alone (for comparison see Fig- intestinal LPLs (mean 52.3%, SD 2.6, tinal lamina propria expressed CCR9 ures 2a and 3c), indicating that a n = 3) expressed CCR9 (Figure 4a). In (Figure 4b). In contrast, less than 2% large population of activated CCR9+ contrast, CCR9 was not expressed on of OT-1 cells in the lung and liver CD8αβ+ lymphocytes may be gener- colonic CD8αβ+ IELs or CD8αβ+ lym- expressed CCR9 (data not shown). ated in gut-associated lymphoid tis- phocytes isolated from inflamed Given that CCR9 is maintained on sue (GALT) during a productive delayed-type hypersensitivity skin a subset of CD8αβ+ lymphocytes dur- immune response. Together these (Figure 4a). Thus, CCR9 is selectively ing activation in MLNs, a possible

Figure 3 CCR9 is maintained on a subset of OT-1 cells following activation in MLNs. (a) CCR9 expression on CFSE+ CD8αβ+ OT-1 (Ly5.2+) lymphocytes. OT-1 cells were labeled with CFSE and injected into C57BL/6J-Ly5.1 mice. MLNs and PLNs were removed 2 days after intraperitoneal administration with OVA plus LPS, and the expression of CCR9 on CFSE+ OT-1 cells was determined by flow cytometry. Results are from one representative experiment of four performed with pooled cells from three to five mice. (b) Combined results of CCR9 expression on OT-1 cells from MLNs (filled symbols) and PLNs (open symbols) 2 days after stimulation with OVA plus LPS. Each symbol represents the mean ± SEM of three to four experiments with cells pooled from three to five mice per experiment. The data from cells that had undergone no divisions and one division are from two experiments, since there were too few cells for analysis in the remaining two experiments. *P < 0.02, **P < 0.002. (c) CCR9 expression on OT-1 (Ly5.2+, CD8αβ+) cells 3 days after stimulation with OVA plus LPS. Results are from the same experiment as in a and are representative of three performed.

The Journal of Clinical Investigation | October 2002 | Volume 110 | Number 8 1117 not shown). Thus, CCL25 does not play a critical role in the activation and subsequent proliferation of CD8αβ+ lymphocytes in secondary lymphoid organs. In addition, the percentage of OT-1 cells in the liver of mice receiving anti-CCL25 or control antibody did not differ (Figure 5c). In marked contrast, anti-CCL25 antibody treatment led to a signifi- cant reduction in the number of OT-1 cells localizing to the small- intestinal epithelium (Figure 5, d and e). Together these results demonstrate an important role for CCL25 in the selective localization of antigen-specific CD8αβ+ lymphocytes to the small-intestinal epithelium.

Figure 4 Discussion + Activated CCR9 OT-1 cells selectively localize to the small intestine. (a) CCR9 expression on Previous studies, based largely on + CD8αβ lymphocytes isolated from tissue effector sites. Cells were stained with anti-CCR9 expression data in humans (8, 10, 22), antibody (open) or anti-CCR9 antibody that had been preincubated with the CCR9-immunizing peptide (shaded), and analyzed by flow cytometry. Results are representative stainings have provided indirect evidence for a from six mice for small-intestinal and colonic IELs and three mice for small-intestinal LPLs. role of CCR9 and CCL25 in small- Skin CD8αβ+ lymphocytes were obtained from the ears of mice with 2,4-dinitro-1-fluo- intestinal immunity. In the present robenzene–induced DTH and are from two stainings using ears pooled from a total of 40 study we demonstrate that CCR9 is mice. Numbers represent percentage of CD8αβ+ lymphocytes that are CCR9+ with background preferentially and functionally ex- + + + staining removed. (b) OT-1 cells (Ly5.2 ) were injected into C57BL/6J-Ly5.1 mice. Three days pressed on murine αEβ7 CD8αβ after intraperitoneal administration of OVA, lymphocytes were isolated from the small intes- lymphocytes and is selectively main- tine, and the expression of CCR9 on OT-1 lymphocytes was determined by flow cytometry. tained on a subset of these cells fol- Results are from five (IEL) and one (LPL) separate stainings using a total of eleven and three mice respectively. Numbers represent the percentage of OT-1 cells expressing CCR9 with back- lowing activation in MLNs. These + + ground staining removed. (c) CCL25 is selectively expressed in the murine small intestine (SI). recently activated CCR9 CD8αβ Quantitative analysis of CCL25 mRNA expression was determined by real-time RT-PCR. lymphocytes specifically localized to the small-intestinal mucosa, and in vivo neutralization of the CCR9 lig- explanation for the expression of involved in the selective localization and, CCL25, selectively inhibited CCR9 on small-intestinal CD8αβ+ of CCR9+ CD8αβ+ lymphocytes to localization of these cells to the small- lymphocytes is that the CCR9 ligand, the small intestine following their intestinal epithelium. Thus, CCL25 CCL25, is involved in the selective activation in MLNs. plays an important and selective localization of recently activated CCL25 mediates the localization of role in the localization of CD8αβ+ CCR9+ CD8αβ+ lymphocytes to the recently activated CD8αβ+ lymphocytes to lymphocytes to the small-intest- small-intestinal mucosa. Indeed, pre- the small-intestinal mucosa. To provide inal mucosa. vious data have demonstrated that proof of concept, we determined the Analysis of CCR9 expression on CCL25 is selectively and constitutive- role of CCL25 in CD8αβ+ lympho- murine lymphocytes demonstrated ly expressed in the murine small cyte localization, using neutralizing that the majority of CD8αβ+ lympho- intestine (9, 12, 21). Since quantita- anti-CCL25 antibody to block CCL25 cytes expressed CCR9 and that, as in tive analysis of CCL25 expression in function in vivo. Following OT-1 cell humans, these cells coexpressed the murine PLNs and MLNs had not transfer, recipient mice were injected integrin αEβ7 (8). In contrast to human been previously performed, we exam- intraperitoneally with neutralizing CD8αβ+ lymphocytes (8, 23), CCR9 ined expression of CCL25 in second- anti-CCL25 or control antibody and αEβ7 were expressed on the major- ary lymphoid organs and intestinal before and after receiving OVA ity of naive murine CD8αβ+ lympho- and extra-intestinal tissues by real- intraperitoneally, and the percentage cytes, indicating differential expres- time RT-PCR. Results from this of OT-1 cells in individual organs sion of these proteins on naive analysis demonstrated that CCL25 is was determined 3 days later. The per- CD8αβ+ lymphocytes in humans and constitutively and selectively ex- centage of OT-1 cells isolated from mice. Further phenotypic analysis of pressed in the small intestine but not the PLNs, spleen, and MLNs after murine CD8αβ+ lymphocytes demon- in the spleen, MLN, PLN, lung, liver, OVA administration was similar in strated that CD62Lhi CD44hi CD8αβ+ or colon (Figure 4c). Together these mice receiving anti-CCL25 or control lymphocytes were CCR9–. This pop- results indicate that CCL25 is antibody (Figure 5, a and b, and data ulation of cells has recently been

1118 The Journal of Clinical Investigation | October 2002 | Volume 110 | Number 8 Figure 5 Anti-CCL25 antibody reduces OT-1 cell recruitment to the small-intestinal epithelium. OT-1 (Ly 5.2+) cells were injected into C57BL/6J-Ly5.1 mice, and the percentage of CD8β+ lymphocytes expressing Ly5.2 in the MLNs (a), PLNs (b), liver (c), and small intestine (d) of control or anti-CCL25 antibody–treated mice was determined by flow cytometry 3 days after intraperitoneal chal- lenge with OVA. Results are the mean ± SEM of six (PLN, MLN, liver) and four (small intestine) mice in each group and show one representative experiment of three performed. *P < 0.05. (e) Representative flow cytometry analysis of the small-intestinal IEL populations in each group of mice.

described as containing long-lived Campbell et al., using a similar trans- selectively expressed in the murine memory cells that share key features of fer model with OVA-specific CD4+ small intestine. These findings are in “central” memory cells and can lymphocytes, recently demonstrated agreement with previous studies in migrate to secondary lymphoid organs that transgenic cells isolated from humans (10) and indicate a selective and sites of inflammation (24–26). MLNs but not from PLNs of mice that role for this chemokine/receptor pair Interestingly, the majority of CD62Lhi had received OVA and LPS migrated to in murine and human small-intestin- CD44hi CD8αβ+ lymphocytes were CCL25 in chemotaxis assays (27). In al immune responses. Interestingly, lo also β7 (data not shown), indicating the context of the present study, this the levels of CCR9 expressed on that they were unable to enter intes- suggests that CCR9 may be selectively murine small-intestinal CD8αβ+ lym- tinal effector sites. Finally, a subset of induced on CD4+ lymphocytes follow- phocytes were consistently lower than previously activated CD8αβ+ lympho- ing activation in MLNs but not in those found on CCR9+ CD8αβ+ lym- cytes and recently activated CD8αβ+ PLNs. These authors also demonstrat- phocytes in MLNs, suggesting that lymphocytes expressed CCR9. These ed that acquisition of intestinal partial downregulation of this che- hi + cells also expressed αEβ7 and likely cor- (α4β7 ) and skin (P-selectin ligand ) mokine receptor may occur once + + + + respond to the CCR9 αEβ7 CD8αβ tropic phenotypes in CD4 lympho- these cells have taken up residence lymphocyte population, previously cytes occurs soon after their activation within the small-intestinal mucosa. described in the human peripheral within GALT and PLNs, respectively. In addition, results from the adoptive blood, that expresses a previously acti- Thus, the microenvironment of sec- transfer experiments demonstrated vated phenotype (8). ondary lymphoid organs plays a criti- the presence of recently activated The expression of CCR9 on a subset cal role in establishment of tissue trop- CCR9+ CD8αβ+ lymphocytes in the of recently activated murine CD8αβ+ ic T lymphocyte subsets. Our results small-intestinal mucosa, indicating a lymphocytes indicated that CCR9 may indicate that selective regulation of potential role for CCL25 in the initial be differentially regulated during acti- chemokine receptors may be part of localization of these cells to this site. vation of these cells in vivo. Indeed, in this process. Finally, the lack of CCR9 Direct evidence for a role of CCL25 in an OVA-specific TCR transgenic trans- expression on a subset of OT-1 cells the localization of CD8αβ+ lympho- fer model, CCR9 was found to be func- activated in MLNs may be of relevance, cytes in the small intestine was tionally maintained on a subset of since lymphocytes localizing to the obtained using anti-CCL25 antibody CD8αβ+ lymphocytes activated in human and the murine large intestine to neutralize CCL25 activity in vivo. MLNs but was rapidly downregulated are CCR9– (10, 22) (Figure 4). Thus, In this set of experiments, neutraliza- on these cells following activation in activated CCR9– CD8αβ+ lymphocytes tion of CCL25 activity led to a selec- PLNs. To our knowledge, together, in MLNs may be destined for extra- tive reduction in the number of these results provide the first direct in intestinal or intestinal effector sites CD8αβ+ TCR transgenic lymphocytes vivo demonstration that chemokine outside of the small intestine, such as localizing to the small-intestinal receptor expression on lymphocytes the colon or rectum. epithelium. The reduction of CD8αβ+ can be differentially regulated depend- Examination of CCR9 expression lymphocyte numbers in the small- ing on the site of initial antigen on CD8αβ+ lymphocytes resident intestinal epithelium was not encounter, and they suggest an impor- within tissue effector sites and CCL25 absolute, consistent with the observa- tant role for GALT in the generation of levels in different organs demon- tion that not all recently recruit- CCR9+ effector CD8αβ+ lymphocytes. strated that CCR9 and CCL25 are ed CD8αβ+ IELs expressed CCR9.

The Journal of Clinical Investigation | October 2002 | Volume 110 | Number 8 1119 Since all human IELs and LPLs with data obtained with CCR9–/– mice In conclusion, results from the cur- express CCR9 (8), it seems plausible (11, 12). In one report, CCR9–/– mice rent study demonstrate a critical role that this chemokine/chemokine were found to have normal numbers for GALT in the establishment of receptor pair may play an even more of small-intestinal CD8αβ+ TCRαβ+ effector CCR9+ CD8αβ+ lymphocytes dominant role in the localization of IELs (11), while a subsequent report and provide the first in vivo demon- lymphocytes to the human small- examining separately derived CCR9–/– stration (to our knowledge) of a role of intestinal mucosa. mice concluded that CD8αβ+ chemokines in lymphocyte trafficking Given the well-characterized func- TCRαβ+ IEL numbers were reduced in to the intestinal mucosa. Together tion of chemokines as chemoattrac- these animals (12). However, this con- these results suggest that CCR9 and tants, the most likely mechanism by clusion was based on the observation CCL25 may provide interesting targets which CCL25 promotes CD8αβ+ lym- that CCR9–/– mice showed a reduced for selective modification of small- phocyte localization within the small- percentage of CD8αβ+ TCRαβ+ IELs intestinal immune responses. intestinal epithelium is the initial among TCRαβ+ IELs. Since the per- recruitment of recently activated centage of TCRαβ+ cells among CD3+ Acknowledgments CCR9+ CD8αβ+ lymphocytes, either IELs was significantly increased due We would like to thank P. Kearney and directly from the circulation or from to a loss of TCRγδ+ IELs, the total A. Runström for kindly providing the small-intestinal lamina propria, number of CD8αβ+ TCRαβ+ IELs was inflamed ears from mice with 2,4-dini- to the epithelium. However, there are in fact unaltered. Thus, results from tro-1-fluorobenzene–induced delayed- a number of other possible mecha- CCR9–/– mice demonstrate that CCR9 type hypersensitivity reactions, and F. nisms, including a role for CCL25 in is not required for generation of the Ivars for constructive comments dur- the short-term retention, survival, or CD8αβ+ TCRαβ+ IEL compartment. ing the course of this work. This work proliferation of CD8αβ+ lymphocytes Several explanations for the differ- was supported by the Swedish Medical within the epithelium. Several lines of ences between our results and those Research Council (MFR 13131), the evidence argue against the involve- obtained with CCR9–/– mice may be Crafoordska, Österlund, Åke Wiberg, ment of CCL25 in the retention proposed. Firstly, CCL25 may be Richard and Ruth Julins, Nanna and/or survival of CD8αβ+ lympho- functioning through an alternative Svartz, and Kocks Foundations, a cytes in the epithelium. Firstly, receptor to CCR9, such as CCR11, Lund Family American Cancer Society endogenous CD8αβ+ IEL numbers which has been shown to bind to grant, the Royal Physiographic Society, would have been equally affected by CCL25 (29). However, the observation and a Crohn’s and Colitis Foundation anti-CCL25 antibody treatment, and that CD8αβ+ lymphocytes from of America project grant to W.W. as a consequence the percentage of CCR9–/– mice fail to migrate to Agace. W.W. Agace is an Assistant Pro- OT-1 cells in this compartment CCL25, indicating that CCR9 is the fessor with the Swedish Science would have remained unchanged. sole CCL25 receptor on these cells Research Council. Secondly, administration of anti- (11, 12), together with the strong cor- 1. Springer, T.A. 1994. Traffic signals for lympho- CCL25 antibody to wild-type mice, as relative data on CCR9 expression pre- cyte recirculation and leukocyte emigration: the performed in the present study, has sented in the current study, strongly multistep paradigm. Cell. 76:301–314. no effect on CD3+ IEL numbers, as argues against this possibility. Sec- 2. Butcher, E.C., and Picker, L.J. 1996. Lymphocyte homing and homeostasis. Science. 272:60–66. + + determined by quantitative immuno- ondly, since CD8αα TCRγδ IELs are 3. Wagner, N., et al. 1996. Critical role for beta7 inte- histochemical analysis (28). Finally, greatly reduced in number in CCR9–/– grins in formation of the gut-associated lym- + phoid tissue. Nature. 382:366–370. while IL-7 or IL-15 promotes the sur- mice (11, 12), CD8αβ IELs may 4.Lefrancois, L., et al. 1999. The role of beta7 vival of CD8αα+ and CD8αβ+ IELs in exhibit increased expansion in situ in integrins in CD8 T cell trafficking during an vitro, incubation of freshly isolated these mice to fill the IEL compart- antiviral immune response. J. Exp. Med. + + 189:1631–1638. CD8αα or CD8αβ IELs with CCL25 ment. Thirdly, other epithelial- 5. Kim, C.H., and Broxmeyer, H.E. 1999. alone or in combination with subop- derived chemokines such as CXCL12, Chemokines: signal lamps for trafficking of T and B cells for development and effector func- timal doses of these cytokines fails to CCL28, and CX3CL1 may compensate tion. J. Leukoc. Biol. 65:6–15. enhance their survival (ref. 28 and our over time for the CCR9 deficiency, 6. Reiss, Y., Proudfoot, A.E., Power, C.A., Campbell, unpublished observations). Thus, resulting in numbers of CD8αβ+ IELs J.J., and Butcher, E.C. 2001. CC chemokine receptor (CCR)4 and the CCR10 ligand cutaneous T CCL25 is unable to directly prevent similar to those in wild-type mice. cell-attracting chemokine (CTACK) in lympho- CD8αβ+ IELs from undergoing apop- Indeed, as described above, not all cyte trafficking to inflamed skin. J. Exp. Med. tosis. In conclusion, while the recently activated CD8 + IELs 194:1541–1547. αβ 7. Homey, B., et al. 2002. CCL27-CCR10 interac- reduced OT-1 numbers in the epithe- expressed CCR9, and anti-CCL25 tions regulate T cell–mediated skin inflamma- lium of anti-CCL25 antibody–treated antibody failed to completely block tion. Nat. Med. 8:157–165. 8. Zabel, B.A., et al. 1999. Human G protein-cou- mice suggest a deficiency in recruit- lymphocyte entry into the epithelium. pled receptor GPR-9-6/CC chemokine receptor 9 ment, only more direct short-term Such redundancy in chemokine usage is selectively expressed on intestinal homing T tracking studies will resolve this issue. was recently described for the lymphocytes, mucosal lymphocytes, and thymocytes and is required for thymus-expressed The ability of neutralizing anti- chemokine CCL27 and the chemokine-mediated chemotaxis. J. Exp. Med. CCL25 antibody to reduce CD8αβ+ chemokine receptor CCR4 in antigen- 190:1241–1256. 9. Wurbel, M.A., et al. 2000. The chemokine TECK localization within the small-intestin- specific T cell recruitment to the is expressed by thymic and intestinal epithelial al epithelium is in apparent conflict inflamed skin (6). cells and attracts double- and single-positive

1120 The Journal of Clinical Investigation | October 2002 | Volume 110 | Number 8 thymocytes expressing the TECK receptor CCR9. 16. Carramolino, L., et al. 2001. Expression of CCR9 23. Cerf-Bensussan, N., et al. 1987. A monoclonal Eur. J. Immunol. 30:262–271. beta-chemokine receptor is modulated in thymo- antibody (HML-1) defining a novel membrane 10. Kunkel, E.J., et al. 2000. Lymphocyte CC cyte differentiation and is selectively maintained molecule present on human intestinal lympho- chemokine receptor 9 and epithelial thymus- in CD8(+) T cells from secondary lymphoid cytes. Eur. J. Immunol. 17:1279–1285. expressed chemokine (TECK) expression distin- organs. Blood. 97:850–857. 24. Sallusto, F., Lenig, D., Forster, R., Lipp, M., and guish the small intestinal immune compartment: 17. Schon, M.P., et al. 1999. Mucosal T lympho- Lanzavecchia, A. 1999. Two subsets of memory T epithelial expression of tissue-specific cyte numbers are selectively reduced in integrin lymphocytes with distinct homing potentials and chemokines as an organizing principle in region- alpha E (CD103)-deficient mice. J. Immunol. effector functions. Nature. 401:708–712. al immunity. J. Exp. Med. 192:761–768. 162:6641–6649. 25. Weninger, W., Crowley, M.A., Manjunath, N., and 11. Wurbel, M.A., et al. 2001. Mice lacking the CCR9 18. Harriman, G.R., Hornqvist, E., and Lycke, N.Y. von Andrian, U.H. 2001. Migratory properties of CC-chemokine receptor show a mild impairment 1992. Antigen-specific and polyclonal CD4+ lam- naive, effector, and memory CD8(+) T cells. J. Exp. of early T- and B-cell development and a reduc- ina propria T-cell lines: phenotypic and func- Med. 194:953–966. tion in T-cell receptor gammadelta(+) gut tional characterization. Immunology. 75:66–73. 26. Manjunath, N., et al. 2001. Effector differentia- intraepithelial lymphocytes. Blood. 98:2626–2632. 19. Agace, W.W., et al. 2000. Human intestinal lami- tion is not prerequisite for generation of memo- 12. Uehara, S., Grinberg, A., Farber, J.M., and na propria and intraepithelial lymphocytes ry cytotoxic T lymphocytes. J. Clin. Invest. Love, P.E. 2002. A role for CCR9 in T lymph- express receptors specific for chemokines 108:871–878. doi:10.1172/JCI200113296. ocyte development and migration. J. Immunol. induced by inflammation. Eur. J. Immunol. 27. Campbell, D.J., and Butcher, E.C. 2002. Rapid 168:2811–2819. 30:819–826. acquisition of tissue-specific homing phenotypes 13. Agace, W.W., et al. 2000. Constitutive expression 20. Kim, S.K., et al. 1997. Activation and migration of by CD4(+) T cells activated in cutaneous or of stromal derived factor-1 by mucosal epithelia CD8 T cells in the intestinal mucosa. J. Immunol. mucosal lymphoid tissues. J. Exp. Med. and its role in HIV transmission and propaga- 159:4295–4306. 195:135–141. tion. Curr. Biol. 10:325–328. 21. Vicari, A.P., et al. 1997. TECK: a novel CC 28. Marsal, J., et al. 2002. Involvement of CCL25 14. Pan, J., et al. 2000. A novel chemokine ligand for chemokine specifically expressed by thymic den- (TECK) in the generation of murine small intes- CCR10 and CCR3 expressed by epithelial cells in dritic cells and potentially involved in T cell devel- tinal CD8αα+CD3+ intraepithelial lymphocyte mucosal tissues. J. Immunol. 165:2943–2949. opment. Immunity. 7:291–301. compartment. Eur. J. Immunol. In press. 15. Muehlhoefer, A., et al. 2000. Fractalkine 22. Papadakis, K.A., et al. 2000. The role of thymus- 29. Gosling, J., et al. 2000. Cutting edge: identifica- is an epithelial and endothelial cell-derived expressed chemokine and its receptor CCR9 tion of a novel chemokine receptor that binds chemoattractant for intraepithelial lympho- on lymphocytes in the regional specializa- dendritic cell- and T cell-active chemokines cytes in the small intestinal mucosa. J. Immunol. tion of the mucosal immune system. J. Immunol. including ELC, SLC, and TECK. J. Immunol. 164:3368–3376. 165:5069–5076. 164:2851–2856.

The Journal of Clinical Investigation | October 2002 | Volume 110 | Number 8 1121