Differential homing mechanisms regulate regionalized effector CD8αβ⁺ T cell accumulation within the small intestine

Abstract

The CC chemokine receptor (CCR)9 is expressed on the majority of small intestinal, but few colonic, T cells, whereas its ligand CCL25 is constitutively expressed by small intestinal epithelial cells. As such, CCR9/CCL25 have been proposed to play a central role in regulating small intestinal but not colonic immune responses and thus to organize regionalized immunity within the intestinal mucosa. Here, we demonstrate that CCL25 is expressed at reduced levels by epithelial cells in the distal compared with proximal small intestine, which correlated with less efficient CCR9-dependent effector CD8αβ⁺ T cell entry into the ileal epithelium. In vitro-generated α₄β₇⁺ effector CD8αβ⁺ T cell entry into the lamina propria was less dependent on CCR9 than entry into the epithelium along the entire length of the small intestine and in particular in the ileum. CCR9-independent α₄β₇⁺ effector CD8αβ⁺ T cell entry was pertussis toxin-sensitive, suggesting a role for additional Gα_I-linked G protein-coupled receptors. Finally, in vivo-primed effector CD8αβ⁺ T cells displayed regionalized differences in their entry to the small intestinal epithelium with enhanced CCR9-independent entry to the ileum. These results highlight a hitherto underappreciated compartmentalization of immune responses within the small intestine and have direct implications for targeting strategies aimed at regulating T cell localization to the small intestinal mucosa.

The intestinal mucosa, including the intestinal epithelium and underlying LP, contains a large number of antigen-experienced T cells that play a central role in regulating local adaptive immune responses and contribute to disease pathology in inflammatory bowel diseases, such as Crohn's disease and ulcerative colitis.

Effector T cell entry into the intestinal mucosa is a complex event regulated by selective expression of intestinal homing receptors on the T cell surface and corresponding ligands within the intestinal mucosa. The intestinal homing receptors, α₄β₇ and CCR9, are selectively induced on T cells primed in mesenteric lymph nodes (MLN) and Peyer's patches (1–4). Interactions between the integrin α₄β₇ on circulating effector T cells and its ligand mucosal addressin cell adhesion molecule-1 on intestinal microvascular endothelial cells are important for efficient CD8αβ⁺ effector T cell entry to both the small intestine and colon (5). In contrast, the CCR9 ligand, CCL25, is selectively expressed in the small intestine but not the colon (6–8), indicating a selective role in regulating small intestinal immunity. Consistent with this hypothesis, TCR transgenic T cell adoptive transfer studies using neutralizing antibodies to CCL25 or CCR9^−/− T cells have demonstrated a role for CCL25/CCR9 in mediating CD4⁺ and CD8⁺ effector T cell localization to the small intestinal mucosa (1, 4, 9). Selective expression of CCL25 in the small intestine is maintained in patients with small bowel Crohn's disease (10), suggesting a potential role for this chemokine in inflammatory bowel disease, and a small molecular weight CCR9 antagonist, Traficet-ENTM, has shown evidence of clinical benefit in a preliminary phase II trial in Crohn's disease patients.^† Despite these findings, CCR9^−/− mice have relatively normal numbers of small intestinal CD8αβ⁺ and CD4⁺ T cells (12–14). Furthermore, antibodies to CCR9 and CCL25 only partially ameliorate early but note late chronic murine ileitis (15). Thus, a clearer understanding of the relative roles of CCR9-dependent and independent T cell entry to and localization within the intestinal mucosa is warranted.

In the current study, we have determined the role of CCR9-dependent and independent effector CD8αβ⁺ T cell localization to the intestinal mucosa. Our findings demonstrate significant differences in T cell localization mechanisms both within and along the entire length of the small intestine highlighting for the first time compartmentalized recruitment mechanisms within the small intestine.

Results and Discussion

Intestinal Epithelial CCL25 Expression Decreases from the Proximal to Distal Portion of the Small Intestine.

To examine potential differences in CCL25 expression in the proximal and distal small intestine, epithelial cells were removed from the duodenum, jejunum, ileum, and colon with EDTA and the levels of CCL25 mRNA assessed by real-time RT-PCR. CCL25 mRNA was detected in small intestinal epithelial cells, but not in colonic epithelial cells, or colonic or small intestinal lamina propria (LP) (Fig. 1A), consistent with previous results (16). Notably however, epithelial CCL25 mRNA levels were significantly reduced in the distal compared with the proximal small intestine (Fig. 1A). To rule out the possibility that nonepithelial cells in the EDTA preparations influenced CCL25 mRNA measurements, epithelial cells were dissected from intestinal sections by laser capture microscopy (Fig. 1B). Similar to epithelium from EDTA preparations, epithelial cells prepared from the distal small intestine by laser capture microscopy expressed lower levels of CCL25 mRNA than epithelial cells taken from the proximal small intestine (Fig. 1B). The reduction in CCL25 mRNA in the ileum compared with the jejunum correlated with reduced levels of CCL25 protein (Fig. 1C). CCL25 protein levels in the duodenum varied considerably between animals (n = 5; 6,270, 1,120, 8,610, 2,040, and 1,347 pg/mg protein), for reasons that we currently do not understand.

Fig. 1.

Intestinal epithelial CCL25 expression decreases from the proximal to the distal part of the small intestine. Total RNA was extracted from epithelial cells isolated by consecutive EDTA treatments (A) or laser capture microscopy (B), and CCL25 mRNA levels were quantified by real-time RT-PCR. (A) Mean (SEM) of three (LP) or five (epithelial cells) experiments. (B) Mean (SEM) shown from two separate experiments. Levels of CCL25 mRNA in the intestinal LP and colon were similar to that found in negative control tissues (lung and liver). (C) CCL25 protein levels were determined by sandwich ELISA. Mean (SEM) from five (intestine) and two (spleen) individual mice. ∗, P < 0.05, calculated by using two-tailed paired Student's t test.

Reduced CCR9-Dependent T Cell Accumulation in the Distal Small Intestine.

To determine whether T cells that enter the epithelium by a CCR9-dependent mechanism accumulate less in the distal compared with the proximal small intestinal epithelium, WT (Ly5.2⁺Ly5.1⁺) and CCR9^−/− (Ly5.2⁺) OT-I cells were stimulated in vitro with ovalbumin (OVA) peptide-pulsed splenic dendritic cells (DCs) in the presence of retinoic acid [that induces expression of CCR9 and α₄β₇ (17)]. These stimulatory conditions generated homogenous populations of effector CCR9⁺ α₄β₇⁺ WT and α₄β₇⁺ CCR9^−/− OT-I cells (Fig. 2A) that were subsequently coinjected into recipient mice. Analysis of the CCR9^−/−:WT OT-I ratio 12 and 24 h after injection demonstrated that α₄β₇⁺ CCR9^−/− OT-I cells were dramatically disadvantaged in their ability to enter the epithelium along the entire length of the small intestine compared with their WT counterparts (Fig. 2 B and C and data not shown). In contrast, CCR9^−/− and WT OT-I cells were detected at equal frequencies in secondary lymphoid organs and the liver (Fig. 2C). Endogenous CD8αβ⁺ intraepithelial lymphocyte (IEL) were evenly distributed along the length of the small intestine (Fig. 2D). To compare the efficiency of OT-I cell accumulation to the different epithelial compartments, we determined the percentage of WT OT-I cells within the CD8αβ⁺ IEL compartment. Results from this analysis demonstrated that the accumulation of WT OT-I cells in the ileal IEL compartment was lower than that of the jejunum and duodenum (Fig. 2E). Thus, under conditions of CCR9-dependent effector T cell entry, reduced expression of CCL25 in the ileal epithelium correlates with reduced effector T cell accumulation at this site.

Fig. 2.

Effector CD8αβ⁺ T cells displaying CCR9-dependent entry into the small intestinal epithelium accumulate less efficiently in the ileal epithelium. In vitro-generated CCR9^−/− and WT effector OT-I cells were injected i.v. into C57BL/6.Ly5.1 recipient mice and the ratio of CCR9^−/− to WT OT-I cells in different tissues determined 24 h later. (A) Representative flow cytometry plots showing CCR9 and α₄β₇ expression on in vitro-generated effector OT-I cells. (B) Representative flow cytometry plots showing CCR9 and WT OT-I cells in the MLN and duodenal, jejunal, and ileal epithelium. (C) CCR9^−/−:WT OT-I ratio in the different organs. Mean (SEM) from three experiments, using three individual mice per experiment for spleen, liver, PLN, and MLN and three pooled mice for duodenal (duo), jejunal (jej), and ileal epithelium. (D) Percentage of endogenous CD8αβ⁺ T cells among CD45⁺IEL in the duodenal (duo), jejunal (jej), and ileal epithelium. Numbers represent mean (SEM) from three experiments using three individual mice per experiment. (E) Percentage of WT OT-I cells within the CD8αβ⁺ IEL gate normalized to duodenal values. Mean (SEM) of three experiments, using three pooled mice per experiment.

CD8αβ⁺ T Cell Recruitment to the Small Intestinal LP Depends Less on CCR9 Than Entry to the Epithelium.

Studies have demonstrated a role for CCR9 in effector CD4⁺ T cell and IgA B cell recruitment to the small intestinal LP (9, 18, 19), suggesting that CCR9 functions in the initial recruitment of lymphocytes from the vasculature. Consistent with this hypothesis, CCL25 protein is detected on microvascular endothelium in the small intestine (19–21) and neutralizing antibodies to CCL25 reduced adhesion of adoptively transferred intestinal lamina propria lymphocytes and IEL to small intestinal post capillary venules (21). To determine whether in vitro-generated effector OT-I cell entry to the small intestinal LP was similarly CCR9-dependent, the CCR9^−/−:WT OT-I ratio was compared in the intestinal LP and epithelium. OT-I effector cells displayed less CCR9-dependent entry to the intestinal LP compared with the intestinal epithelium (Fig. 3 A and B) along the entire length of the small intestine (compare LP ratio in Fig. 3C with the epithelial ratio in Fig. 2C). Furthermore, OT-I cell entry into the ileal LP was less dependent on CCR9 than entry into the jejunal and duodenal LP (Fig. 3C).

Fig. 3.

CD8αβ⁺ T cell recruitment to the small intestinal LP is less dependent on CCR9 than their recruitment to the small intestinal epithelium. In vitro-generated CCR9^−/− and WT effector OT-I cells were coinjected i.v. into C57BL/6.Ly5.1 mice and the CCR9^−/−:WT OT-I ratio determined 24 h later in the MLN and small intestinal LP and epithelium. (A) Representative flow cytometry plots of CCR9^−/− and WT OT-I cells in the MLN and intestine gating on live CD8β⁺ cells. (B) CCR9^−/−:WT OT-I ratio in MLN, the small intestinal LP, and epithelium. (C) MLN and intestinal LP segments. Mean (SEM) from four (B) and three (C) experiments, using three individual mice per experiment for MLN samples and three pooled mice per experiment for the intestinal LP and epithelial samples. ∗, P < 0.05, calculated by using paired two-tailed Student's t test.

CCR9 Independent Effector CD8αβ⁺ T Cell Entry to the Small Intestinal Mucosa Is Pertussis Toxin Sensitive.

To determine whether CCR9-independent effector CD8αβ⁺ T cell migration to the small intestinal LP involves alternative Gα_i protein-coupled receptors, in vitro-generated α₄β₇⁺ CCR9^−/− effector OT-I cells were pretreated with pertussis toxin (PTX) and injected into recipient animals (Fig. 4 A and B). PTX-treated and untreated CCR9^−/− OT-I cells were equally capable of entering the spleen (Fig. 4B), demonstrating that PTX-treatment had no detrimental effect on the in vivo survival of these cells. In marked contrast, PTX-treatment inhibited CCR9^−/− OT-I cell entry into MLN and the small intestinal LP (Fig. 4 A and B). Thus, CCR9-independent T cell entry into the small intestinal LP is an active mechanism involving Gα_i-coupled receptors.

Fig. 4.

CCR9-independent migration of effector CD8αβ⁺ cells to the small intestinal LP is pertussis toxin sensitive. PTX-treated or untreated in vitro-generated α₄β₇⁺ CCR9^−/− effector OT-I cells were injected into C57BL/6.Ly5.1 recipients and the proportion of cells within the indicated tissues determined 24 h later. (A) Representative flow cytometry plots showing donor cells in the recipient LP, gating on total live cells. (B) The percentage of injected PTX-treated or untreated CCR9^−/− OT-I cells among total CD8β⁺ cells in the spleen, MLN, and small intestinal LP. Mean (SEM) of seven and six mice receiving PTX-treated or untreated OT-I cells, respectively, pooled from three separate experiments. ∗∗∗, P < 0.001, calculated by using unpaired two-tailed Student's t test. n.s., not significant.

Together, these results demonstrate that in vitro-generated α₄β₇⁺ effector CD8αβ⁺ T cells can use CCR9-dependent and CCR9 independent Gα_i-coupled receptor-dependent mechanisms to gain entry into the small intestinal LP, and that the latter pathway appears particularly relevant in the distal small intestine. Moreover, they provide strong evidence that epithelial derived CCL25 plays an additional role in the recruitment of CCR9⁺α₄β₇⁺ CD8⁺ T cells from the LP into the epithelium along the entire length of the small intestine. Although the mechanism(s) regulating the regionalized requirements for in vitro-generated α₄β₇⁺ effector CD8⁺ T cell entry along the small intestine remain to be determined, one potential explanation for our findings is that microvascular endothelial cells in the distal small intestine express higher levels of Gα_i-coupled receptor ligands that promote CCR9-independent entry. We have demonstrated that LP preparations containing microvascular endothelial cells do not contain CCL25 mRNA (16). Thus, an additional nonexclusive possibility is that the lower levels of epithelial-derived CCL25 in the distal small intestine result in less efficient CCL25 presentation by vascular endothelial cells at this site. Of note, in contrast to previous reports (19–21), we were unable to detect CCL25 protein on small intestinal vascular endothelium, presumably because CCL25 levels were below the detection limit of our assays.

CCR9-Dependent and Independent Entry of in Vivo-Primed Effector CD8αβ⁺ T Cells to Small Intestinal Epithelium.

Seventy-five to 90% of endogenous duodenal, jejunal and ileal CD8⁺ IEL expressed CCR9 and a proportion of these isolated IEL responded to CCL25 in chemotaxis assays [supporting information (SI) Fig. 7 A and B]. Nevertheless, the relative importance of CCR9-dependent versus CCR9-independent effector CD8αβ⁺ T cell localization to these distinct intestinal epithelial compartments after in vivo priming is currently unclear. To address this question, CCR9^−/− and WT OT-I cells were injected i.v. at an equal ratio into recipient mice that were subsequently immunized i.p. or orally with OVA and adjuvant. Three days after immunization the ratio of CCR9^−/−:WT OT-I cells was assessed in the MLN and intestinal segments. Although both immunization regimes induce efficient generation of CCR9⁺α₄β₇⁺ effector OT-I cells in the MLN (1, 22), oral antigen administration is likely to preferentially target intestinal DC populations and OT-I cell priming in the MLN after oral immunization requires CCR7-dependent DC migration from the intestinal LP into the MLN (22). I.p. immunization, on the other hand, results in simultaneous OT-I priming in multiple secondary lymphoid organs, and priming in the MLN can occur independently of CCR7 (22). After i.p. immunization, the CCR9^−/−:WT OT-I ratio in the MLN was similar to that of the input population (data not shown), consistent with results in ref. 1. Furthermore, the CCR9^−/−:WT OT-I ratio in the cecal and colonic epithelium was similar to that observed in the MLN, demonstrating that CCR9 does not play a role in effector CD8αβ⁺ T cell migration to these sites (Fig. 5 A and B). As observed with in vitro-generated gut tropic OT-I cells, CCR9^−/− OT-I cells were heavily disadvantaged in their ability to enter the duodenal epithelium compared with WT OT-I cells (Fig. 5 A–C). However, the importance of CCR9 for OT-I cell localization to the small intestinal epithelium displayed marked regional differences. Thus, OT-I cell entry to the jejunum was less dependent on CCR9 than entry into the duodenum and entry to the ileum was less dependent on CCR9 compared with the jejunum (Fig. 5 A–C). In contrast to in vitro-generated effector WT OT-I cell migration to the small intestinal epithelium (Fig. 2E), effector WT OT-I cells generated after i.p immunization localized more efficiently to the ileal compared with the jejunal and duodenal epithelium (Fig. 5D). The majority of WT OT-I cells entering the ileal epithelium after i.p immunization expressed CCR9 and α₄β₇, whereas CCR9^−/− OT-I cells expressed α₄β₇ (SI Fig. 8) suggesting that these cells had been primed in intestinal LN. Furthermore, WT and CCR9^−/− OT-I cells that entered the intestinal epithelium expressed similar levels of CXCR3 [WT 92.8% (SEM = 3.4); CCR9^−/− 93.9% (SEM = 1.4), n = 6], CXCR6 [WT 85.6% (SEM = 8.7); CCR9^−/− 80.7% (SEM = 9.3), n = 4], CCR6 [WT 32% (SEM = 13.4); CCR9^−/− 20% (SEM = 9.2), n = 4], and CCR2 [WT 5.5% (SEM = 2.6); CCR9^−/− 3.9% (SEM = 3.9), n = 2], indicating that CCR9 deficiency did not result in compensatory induction of these receptors. Together, these results demonstrate that CCR9-independent entry mechanisms can play an important role in directing effector CD8⁺ T cells to the ileal epithelium and suggest that the recruitment mechanism mediating this CCR9-independent ileal epithelial localization may not be well recapitulated by our in vitro induction protocol. In contrast to the i.p immunization regime, oral immunization resulted in more pronounced CCR9-dependent OT-I cell migration to the jejunal and ileal epithelium (Fig. 6A). Furthermore, under these conditions, WT OT-I cells localized less efficiently to the ileal than the jejunal epithelium (Fig. 6B), similar to the in vitro-generated WT effector OT-I cells. Despite enhanced CCR9-dependent migration to the ileum, OT-I cell entry to the ileal epithelium remained significantly less dependent on CCR9 compared with their localization to the jejunal epithelium (Fig. 6C).

Fig. 5.

CCR9-dependent and -independent OT-I cell localization to the small intestinal epithelium after i.p immunization. CCR9^−/− and WT OT-I cells were coinjected into C57BL/6.Ly5.1 mice, and their relative efficiency in entering epithelial sites along the intestine determined 3 days after immunization with OVA and LPS i.p (A–C). (A) Representative flow cytometry plots used to determine the ratio of CCR9^−/− to WT OT-I cells in individual tissues gating on live CD8β⁺ cells. Numbers represent CCR9^−/−:WT OT-I ratios for the respective plots. (B) CCR9^−/−:WT OT-I ratio in different intestinal segments. Mean (SD) of four separate experiments, using three or four individual mice per experiment and pooled colonic tissues from three individual mice per experiment. (C) CCR9^−/−:WT OT-I cell ratio in intestinal segments from individual mice. ∗∗, P < 0.01 and ∗∗∗, P < 0.001, calculated by using paired two-tailed Student's t test. (D) Percentage of in vivo-activated WT OT-I cells within the CD8β⁺ IEL gate normalized to duodenal values. Mean (SEM) of eight mice from three separate experiments. ∗, P < 0.05, calculated by using two-tailed paired Student's t test.

Fig. 6.

CCR9-dependent and independent OT-I cell localization to the small intestinal epithelium after oral immunization. CCR9^−/− and WT OT-I cells were coinjected into C57BL/6.Ly5.1 mice and their relative efficiency in entering epithelial sites along the intestine determined 3 days after oral immunization with OVA and cholera toxin. (A) CCR9^−/−:WT OT-I ratio in different intestinal segments. Mean (SEM) of two experiments, using six or seven individual mice. (B) Percentage of in vivo-activated WT OT-I cells within the CD8β⁺ IEL gate normalized to duodenal values. Mean (SEM) of six or seven individual mice from two separate experiments. (C) CCR9^−/−:WT OT-I cell ratio in intestinal segments from individual mice. ∗, P < 0.05 and ∗∗, P < 0.01, calculated by using two-tailed paired Student's t test. n.s. not significant.

In summary, our results highlight that effector CD8⁺ T cells may use both CCR9-dependent and independent mechanisms to localize within the small intestinal mucosa and that the degree of CCR9-independent entry depends in part on the site within (LP versus epithelium) or along (proximal versus distal) the small intestine, as well as the conditions under which T cell priming takes place. Further studies will be required to determine the molecular mechanisms regulating CCR9 dependent/independent localization to the small intestine and in particular the role of antigen administration route/antigen dose and the influence of inflammatory mediators in this process.

Conclusion

CCR9 and its ligand CCL25 have been demonstrated to play a role in effector T cell localization to the small intestinal mucosa (4), and are potential therapeutic targets for small bowel Crohn's disease (†, 23). Nevertheless, CCR9^−/− mice display relatively normal numbers of small intestinal CD4⁺ and CD8αβ⁺ T cells (13, 14, 18), indicating that effector T cells may use CCR9-independent mechanisms of entry to the small intestine. In the current study, we demonstrate important regionalized differences in the role of CCR9 in effector CD8αβ⁺ T cell entry to the small intestinal mucosa, both along the length of the intestine mucosa and between the LP and epithelium, and that CCR9-independent entry is an active mechanism involving alternative Gα_i-coupled receptors. These findings demonstrate a remarkable compartmentalization of intestinal adaptive immune responses within the small intestine and suggest in particular that CCR9-independent entry mechanisms should be taken into account when attempting to regulate adaptive immune responses, particularly in the distal small intestine.

Materials and Methods

Mice.

OT-I and C57BL/6.Ly5.1 mice were from The Jackson Laboratory (Bar Harbor, ME), C57BL/6 mice were from Taconic (Lille Skensved, Denmark), and CCR9^−/− OT-I and Ly5.1⁺Ly5.2⁺ WT OT-I mice were generated as described in ref. 1. All mice were bred and maintained at the BMC animal facility of Lund University. Animal experiments were approved by the local ethical review board at Lund University.

Isolation of Epithelial Cells and Intestinal T Cell Populations.

Intestinal tissues were divided into duodenum, jejunum, ileum, cecum, and colon. Duodenum was defined as the first 3-cm section after the pylorus, the jejunum as the following proximal part of the small intestine, and ileum as the distal 5 cm of the small intestine before the cecum (24). Lymphocytes were isolated from the lymph nodes, spleen, liver, intestinal epithelium, and LP of the small intestine, cecum, and colon as described in ref. 4. For real-time PCR analysis, intestinal tissue pieces were dissected and inverted before epithelial cell isolation with six treatments of HBSS supplemented with 10% FCS, 10 mM Hepes, 100 units/ml penicillin, 100 μg/ml streptamycin, 20 μg/ml gentamycin (all from Gibco, Paisely, U.K.), and 30 mM EDTA for 20 min each. Remaining LP was scraped with tweezers to remove remaining epithelial cells and washed in HBSS extensively. Purity of LP fractions was verified by running real-time PCR analysis for intestinal alkaline phosphatase as a marker of contaminating epithelial cells.

Laser Capture Microscopy and Real-Time PCR.

Laser capture of intestinal epithelium was performed exactly as described in ref. 16 on a Zeiss (Thornwood, NY) microscope equipped with a microcatapulting laser system (P.A.L.M. Microlaser Technologies, Bernried, Germany) according to manufacturers instructions. Briefly, Tissue-Tek O.C.T. embedded intestinal tissue sections (10-μm) were placed on PEN-membrane coated slides (P.A.L.M. Microlaser Technologies) and fixed in 70% ethanol for 30 s and acetone for 4 min. Fixed sections were stained with Harris haematoxylin for 10 s (Sigma–Aldrich, Steinheim, Germany), washed, overlaid with 10% DMSO, and placed on dry ice or kept at −80°C until use. All aqueous solutions were made from DEPC-treated water and supplemented with VRC RNase Inhibitor (Sigma–Aldrich). Cells with lymphocyte morphology were destroyed with the laser before capture. Total RNA was extracted from the catapulted samples, using Stratagene Microprep RNA kit (Stratagene, Stockholm, Sweden) and translated into cDNA using SuperScript II (Invitrogen, Stockholm, Sweden). A first PCR was performed with outer primers for 12 cycles (11) before quantification with real-time PCR, using SYBR-green (Stratagene) and a Roche (Pentzberg, Germany) LightCycler as described in ref. 16. The following primers were used: CCL25 outer forward 5′ATAGGCAATACACGCTACAAGC 3′, CCL25 outer reverse 5′GCGGAATTCTTTGATCCTGTGCTGGTAACCCAGG 3′, CCL25 inner forward 5′AGGCACCAGCTCTCAGGACC 3′, CCL25 inner reverse 5′GCGGAATTCGTCTTCAAAGGCACCTTGGGCATGG3′, β-actin outer forward 5′CCGGGACCTGACAGACTA 3′, β-actin outer reverse 5′ACGGATGTCAACGTCACACTTC 3′, β-actin inner forward 5′GAGAGGGAAATCGTGCGTGACA 3′, and β-actin inner reverse 5′GTTTCATGGATGCCACAGGAT 3′. Quantitative real-time PCR of EDTA isolated epithelial cells was performed as described above without the outer primer PCR step.

Intestinal Protein Extraction and ELISA for CCL25.

Intestinal tissues were washed in PBS, snap frozen in liquid nitrogen and homogenized in T-PER tissue protein extraction buffer (Pierce Biotechnology, Rockford, IL) supplemented with complete protease inhibitor mixture (Roche Diagnostics). Lysates were removed after centrifugation for 10 min at 10,000 rpm [Sorvall Instruments (Newton, CT) RC5C centrifuge; GSA-rotor], and total protein levels and CCL25 levels determined by using a Quick Start Bradford kit (Bio-Rad, Hercules, CA) and a CCL25 sandwich ELISA (R&D Systems, Minneapolis, MN), respectively.

In Vitro-Generation of CCR9⁺α₄β₇⁺ OT-I Cells.

CD8β⁺ T cells and DCs were purified from the spleens of CCR9^−/− and WT OT-I mice and C57BL/6 mice, respectively, by magnetic cell sorting (Miltenyi Biotech, Bergisch Gladbach, Germany), using biotinylated anti-CD8β antibody followed by streptavidin-conjugated magnetic beads (purity >98%) or anti-CD11c magnetic beads for DC isolation. Splenic DCs were pulsed with OVA_257–264 SIINFEKL peptide (200 pM; Innovagen, Lund, Sweden), washed, and incubated with OT-I cells in RPMI-c (RPMI medium 1640 supplemented with 10% FCS, 10 mM Hepes, 1 mM Na-puryvate, 50 μM 2-mercaptoethanol, 100 units/ml penicillin, 100 μg/ml streptamycin, 20 μg/ml gentamycin) and 10 nM retinoic acid (Sigma–Aldrich). After 3.5 days of coculture, cells were transferred into 24-well tissue culture plates (Corning, Corning, NY) and allowed to expand in RPMI-c supplemented with 10 ng/ml IL-7 and 10 ng/ml IL-15 (both from R&D Systems) for an additional 2.5 days. CCR9^−/− and WT OT-I cells (8–10 × 10⁶ cells of each) were injected i.v. into C57BL/6.Ly5.1 recipient mice and the CCR9^−/−:WT OT-I cell ratio in intestinal tissues determined 24 h later by flow cytometry. For PTX-treatment experiments, in vitro cultured CCR9^−/− OT-I cells were incubated with 200 ng/ml PTX (Sigma–Aldrich) at 37°C for 2 h in RPMI medium 1640 containing 5% FCS and 25 mM Hepes (untreated cells were kept under the same, but PTX-free, conditions). Cells were washed with PBS, and 8–10 × 10⁶ PTX-treated or untreated cells were injected into C57BL/6.Ly5.1 recipient mice. Twenty-four hours later, the mice were killed, and the ratio of injected cells to endogenous CD8β⁺ cells was determined in the spleen, MLN, and intestinal LP by flow cytometry. To increase the lymphocyte yield, intestinal LP lymphocytes were here isolated without percoll density centrifugation.

In Vivo Priming Model.

Splenic CD8b⁺ T cells were purified from CCR9^−/− (Ly5.2⁺) and WT (Ly5.1⁺Ly5.2⁺) OT-I mice by magnetic cell sorting and coinjected (3–5 × 10⁶ of CCR9^−/− and WT cells) i.v. into C57BL/6.Ly5.1 recipients. Recipient mice were immunized i.p with OVA (5 mg, grade VI, Sigma–Aldrich) and LPS (100 μg, Escherichia coli, serotype 055:B5; Sigma–Aldrich) or orally with OVA (50 mg, grade VI, Sigma–Aldrich) and cholera toxin (10 μg, Sigma–Aldrich), killed 3 days later and the ratio of CCR9^−/−:WT OT-I cells in the MLN and intestinal tissue determined by flow cytometry.

Antibodies and Flow Cytometry.

Flow cytometry analysis was performed as described in ref. 9, using the following antibodies and reagents: anti-Ly5.1 (A20), anti-Ly5.2 (104), anti-CD8β (53-5.8), anti-CD4 (L3T4), anti-α₄β₇ (DATK32), biotinylated mouse-anti-rat IgG2b (G15-337) (BD PharMingen, San Diego, CA); anti-CCR9 [7E71 (12)]; streptavidin-APC (Nordic BioSite, Täby, Sweden); 7-amino-actinomycin D (7AAD) (Sigma–Aldrich), and anti-FcRII/III (2.4G2) (American Type Culture Collection, Rockville, MD).

Abbreviations

CCR

CC chemokine receptor

IEL

intraepithelial lymphocyte

LP

lamina propria

MLN

mesenteric lymph node

OVA

ovalbumin

PTX

pertussis toxin.