Abstract
CCR9 mediates chemotaxis in response to CCL25/thymus-expressed chemokine and is selectively expressed on T cells in the thymus and small intestine. To investigate the role of CCR9 in T cell development, the CCR9 gene was disrupted by homologous recombination. B cell development, thymic αβ-T cell development, and thymocyte selection appeared unimpaired in adult CCR9-deficient (CCR9−/−) mice. However, competitive transplantation experiments revealed that bone marrow from CCR9−/− mice was less efficient at repopulating the thymus of lethally irradiated Rag-1−/− mice than bone marrow from littermate CCR9+/+ mice. CCR9−/− mice had increased numbers of peripheral γδ-T cells but reduced numbers of γδTCR+ and CD8αβ+αβTCR+ intraepithelial lymphocytes in the small intestine. Thus, CCR9 plays an important, although not indispensable, role in regulating the development and/or migration of both αβ− and γδ− T lymphocytes.
Development of T cells continues into adulthood and requires the periodic migration of progenitor cells from the bone marrow to the thymus (1, 2). The ordered progression of thymocytes through distinct stages of development is also associated with the migration of cells into and between different thymic microenvironments (3). Chemokines are a group of small (8–14 kDa), mostly basic, structurally related molecules that regulate trafficking of leukocytes through interactions with a subset of seven-transmembrane, G protein-coupled receptors (4, 5, 6, 7, 8). Several different chemokines are expressed in the thymus, including XCL1/lymphotactin, CCL3/macrophage-inflammatory protein (MIP)3-1α, CCL4/MIP-1β, CCL17/thymus and activation-regulated chemokine, CCL19/EBI-1 ligand chemokine/MIP-3β, CCL21/6-cysteine chemokine/secondary lymphoid-tissue chemokine, CCL25/thymus-expressed chemokine, and CXCL12/stromal cell-derived factor-1, suggesting that chemokines may play an important role in thymopoiesis (4, 5, 6, 7, 8). Although a great deal has been learned about the function of chemokines and chemokine receptors in regulating the migration of mature lymphocytes, little information is available concerning their potential role in lymphocyte development.
Expression of CCL25/thymus-expressed chemokine was initially detected in medullary dendritic cells (9), but recent experiments indicate that it is also expressed by thymic epithelial cells in both cortex and medulla (10, 11). Interestingly, CCL25 is also expressed in fetal thymus, raising the possibility that it may participate in recruiting T progenitor cells to the thymus (12). The only other known site of CCL25 production is the epithelial layer of the small intestine (11, 13, 14, 15). Thus, CCL25 may also be important for the development, homeostasis, and/or function of mucosal T lymphocytes.
CCR9 mediates chemotaxis in response to CCL25 (16, 17, 18, 19). CCR9 is expressed on the majority of immature CD4+CD8+ (double-positive (DP)) thymocytes, is down-regulated during their transition to the CD4+ or CD8+ (single-positive (SP)) stage, and is expressed on a minor subset of CD8+ lymph node T cells (20, 21). CD69+ thymocytes demonstrate enhanced CCL25-induced migration compared with CD69− thymocytes (21, 22), and thymocyte migration in response to CCL25 is augmented by TCR signaling (21). These findings suggest that CCR9 may be involved in regulating T cell migration within the thymus, particularly during thymocyte selection. Approximately half of all γδTCR+ thymocytes and peripheral γδ-T cells express CCR9, and these cells migrate to CCL25 (21). The expression of CCR9 on specific γδ-T cell subsets (e.g., Vγ2+, but not Vγ3+) indicates that CCR9 may also function in the development and/or trafficking of γδ-T cells (21). Finally, pre-pro-B cells in the bone marrow respond to CCL25, raising the possibility that CCR9 may regulate the early stages of B cell development (23).
To investigate the role of CCR9 during lymphocyte development, we generated CCR9-deficient (CCR9−/−) mice by homologous recombination. Surprisingly, both αβ-T cell and B cell development appeared normal in adult CCR9−/− mice. In addition, thymocyte selection of αβ-lineage T cells was unaffected in CCR9−/− mice, even through most DP thymocytes express CCR9 at high levels and respond to CCL25. However, the results of competitive bone marrow transplantation experiments demonstrated that CCR9−/− bone marrow cells had a reduced capacity to repopulate the thymus compared with bone marrow cells from CCR9+/+ mice. These results suggest that CCR9 may be involved in regulating the migration of progenitor cells to the thymus or the retention of T progenitor cells in the thymus. CCR9−/− mice also contained increased numbers of peripheral γδ-T cells but reduced numbers of γδTCR+ intraepithelial lymphocytes (IEL) in the small intestine. Thus, CCR9 also plays an important role in regulating the development and/or migration of γδ-T lymphocytes.
Materials and Methods
Generation of CCR9-deficient mice
An ∼15-kb fragment containing the mouse CCR9 gene was cloned from a 129 SvJ λ genomic library (Stratagene, La Jolla, CA). A 7-kb EcoRV fragment and a 1.6-kb HindIII-EcoRV fragment were then cloned into the XpPNT (NEO/TK) vector (24). The targeting construct was linearized with NotI and electroporated into 2 × 107 embryonic stem cells (ES cells). After transfection, ES cells were selected in the presence of G418 and gancyclovir and screened for homologous recombination. Chimeric mice were generated from CCR9+/− ES cell clones by injection into B6 blastocysts, and germline transmission of the mutant allele was confirmed by Southern blot analysis of DNA obtained from tail biopsies.
H-Y (25), P14 (26), and AND (27) αβTCR transgenes were bred into the CCR9−/− background. The phenotype of thymocytes and peripheral T cells was analyzed by staining with PE-labeled anti-CD4, CyChrome-labeled anti-CD8α, and FITC-labeled anti-H-Y clonotypic receptor (T3.70), anti-TCR-Vα 2 (P14), or anti-TCR-Vα 11 (AND) mAb as previously described (28). Mice were maintained in the H-2Db background by mating with C57BL/6 mice.
Cell preparation and flow cytometry
Thymus, spleen, lymph node, and bone marrow were excised from mice, and single-cell suspensions were prepared. IEL were prepared from small and large intestines according to conventional methods (29). Standard flow cytometry was performed as previously described using a FACSCalibur and CellQuest software (BD Biosciences, Mountain View, CA) (21). Anti-CD3ε, anti-CD4, anti-CD8α, anti-CD8β, anti-CD25, anti-CD19, anti-CD44, anti-B220, anti-CD45.2, anti-βTCR, anti-δTCR, anti-TCR-Vγ2, anti-TCR-Vδ4, anti-TCR-Vα2, anti-TCR-Vα11, and anti-DX5 mAb were purchased from BD PharMingen (San Diego, CA). The polyclonal anti-CCR9 Ab has been described previously (21). Anti-H-Y clonotypic receptor mAb (T3.70) was purified from cell culture supernatants in our laboratory. Anti-TCR-Vγ1 (2.11) (30) and anti-TCR-Vγ5 (GL1) (31) mAbs were provided by Dr. L. Lefrancois (University of Connecticut Health Center, Farmington, CT). Unconjugated anti-FcγRII (2.4G2) was used to block nonspecific binding of the labeled Ab. PE-conjugated streptavidin and CyChrome-conjugated streptavidin were also purchased from BD PharMingen.
Chemotaxis assays
Chemotaxis assays were performed as previously described (18, 21) with modifications, using 6.5-mm Transwell tissue culture inserts with a 5-μm pore size (Corning, Cambridge, MA). Thymocytes were suspended at 1 × 107
Northern blotting
Total RNA was isolated from various organs from B6 and Rag-1−/− mice using TRIzol (Life Technologies, Gaithersburg, MD). Twenty micrograms of total RNA was fractionated on a 1% agarose/formaldehyde gel and transferred onto a GeneScreen Plus nylon membrane (New England Nuclear, Boston, MA). The membrane was hybridized with 32P-labeled cDNA fragments encoding mouse CCR9, mouse CCL25, or human EF-1α.
Bone marrow chimeras
Bone marrow cells were isolated from femurs of CCR9+/+ (CD45.2), CCR9−/− (CD45.2), and B6.SJL-Ptprca/BoAiTac mice (B6.CD45.1; Taconic Farms, Lexington, KY). A total of 2 × 106 cells consisting of various ratios of CCR9+/+ (CD45.2) and B6.CD45.1 bone marrow cells or CCR9−/− (CD45.2) and B6.CD45.1 bone marrow cells were injected i.v. into lethally irradiated (9.5 Gy) Rag-1−/− B6 mice (CD45.2). One to 2 mo after the transplantation thymocytes and lymph node B cells were analyzed by FACS for the presence of CD45.2+ cells.
Statistical analysis
Data from mice of the different phenotypes were analyzed using Student’s t test.
Results
Generation of CCR9-deficient mice
To investigate the biological significance of CCR9 in vivo, CCR9−/− mice were generated by homologous recombination. The gene-targeting vector was designed to delete exon 4, which contains most of the CCR9-coding sequence (Fig. 1⇓A). The expected structure of the targeted CCR9 locus was confirmed by Southern blotting with the indicated 3′ probe (Fig. 1⇓B). CCR9 mRNA was undetectable in the thymus, spleen, lymph node, and small intestine from CCR9−/− mice (data not shown). FACS analysis confirmed that thymocytes from CCR9−/− mice lacked CCR9 surface expression (Fig. 1⇓C). CCR9−/− mice did not exhibit any developmental abnormalities, were produced in Mendelian ratios, and were indistinguishable from CCR9+/+ or CCR9+/− littermates on the basis of size, activity, or fertility (data not shown).
Targeted disruption of the murine CCR9 gene. A, Targeting vector and homologous recombination at the CCR9 locus. Top, Restriction map of the murine CCR9 locus. Middle, The CCR9 gene targeting construct. An EcoRI-HindIII fragment containing most of the exon 4 coding sequence was replaced with a phosphoglyceraldehyde kinase-directed neomycin cassette (Neo). Bottom, Predicted structure of the disrupted CCR9 gene. B, Southern blot analysis of genomic DNA. DNA from mouse tails was digested with HindIII and probed with the 3′ probe depicted in A. C, FACS analysis of CCR9 surface expression on thymocytes. Total thymocytes from CCR9+/+, CCR9+/−, and CCR9−/− mice were incubated with biotinylated anti-CCR9 Ab (open tracing) or biotinylated control rabbit Ig (shaded tracing), then labeled with PE-conjugated streptavidin and analyzed by FACS.
Normal αβ-lineage T cell development in adult CCR9−/− mice
We initially focused our analysis on T cell development in CCR9−/− mice, because CCR9 is expressed on the surface of most thymocytes, especially CD4+CD8+ (DP) cells (20, 21). The numbers and distribution of the major thymic and peripheral T cell subsets, as defined by the expression of CD4 and CD8, were normal in CCR9−/− mice (Fig. 2⇓). In addition, surface expression of CD3, βTCR, CD5, and CD69 were indistinguishable on thymocytes and lymph node T cells from CCR9−/− and CCR9+/+ littermates (data not shown). Histological studies revealed normal thymus architecture in CCR9−/− mice (data not shown). Peripheral T cells from CCR9−/− mice demonstrated normal proliferative responses after CD3 cross-linking, and CCR9−/− mice had normal Ab responses to T cell-dependent Ags (data not shown). Finally, there were no differences in the number or phenotype of thymocytes from newborn CCR9−/− and CCR9+/+ mice (data not shown). Collectively, these data indicate that αβ-T cells can develop normally in the absence of CCR9.
Normal development of αβ-T cells in adult CCR9−/− mice. A, Representative two-color fluorescence plots showing the expression of CD4 and CD8α on thymocytes and lymph node cells from CCR9+/+ and CCR9−/− mice. B, Comparison of total cell numbers from thymus, lymph node, and spleen from CCR9+/+ and CCR9−/− mice. C, Frequency of CD4−CD8− (DN), CD4+CD8+ (DP), CD4+CD8− (CD4-SP), and CD4−CD8+ (CD8-SP) cells in thymus from CCR9+/+ and CCR9−/− mice. D, Frequency of CD3+, CD4+, CD8α+, and B220+ cells in lymph nodes from CCR9+/+ and CCR9−/− mice. Open and closed circles represent data from CCR9+/+ and CCR9−/− mice, respectively.
We next analyzed the responsiveness of thymocytes to CCL25. CCR9−/− thymocytes did not respond to CCL25, although they migrated to CXCL12 in a manner comparable to CCR9+/+ cells (Fig. 3⇓). CCR9+/− thymocytes expressed CCR9 at lower levels than CCR9+/+ littermates (Fig. 1⇑C) and exhibited a reduced chemotactic response to CCL25 (Fig. 3⇓). These data indicate that CCR9 expression is essential for chemotaxis of thymocytes to CCL25.
Lack of chemotactic response of thymocytes from CCR9−/− mice to CCL25. Total thymocytes from CCR9+/+, CCR9+/−, and CCR9−/− mice were used for chemotaxis assays with CCL25 (200 nM) and CXCL12 (50 nM). Results are expressed as cells migrating per 106 input cells. Determinations were performed in duplicate, and error bars represent the SD.
CCL25 preferentially induces the migration of CD3highCD69+ DP thymocytes compared with CD69− DP thymocytes (20, 21, 22), and TCR cross-linking of DP thymocytes enhances CCL25-mediated migration (21). These data suggest that CCL25 may be involved in thymocyte selection. To determine whether positive or negative selection was affected in CCR9−/− mice, we crossed the H-Y TCR transgene into the CCR9−/− background and analyzed the phenotype of thymocytes in the positively selecting (female) or negatively selecting (male) background. The absence of CCR9 did not alter the efficiency of either positive or negative thymocyte selection in H-Y TCR transgenic mice (data not shown). In addition, no differences in the efficiency of thymocyte selection were observed in CCR9−/− mice that expressed the P14-TCR transgene or the AND-TCR transgene (data not shown). These data suggest that CCR9 does not play a critical role in thymocyte selection.
CCR9−/− mice contain increased numbers of γδ-T lymphocytes in spleen and lymph node
Approximately half of all γδTCR+ cells in thymus, spleen, and lymph node express CCR9 and can respond to CCL25 (21). Therefore, we next compared the number and phenotype of γδ-T cells in different organs from CCR9−/− and CCR9+/+ mice. The number and percentage of γδTCR+ thymocytes were similar in CCR9−/− and CCR9+/+ mice (Fig. 4⇓A). However, both the absolute number and the percentage of γδTCR+ cells were increased ∼2-fold in spleen and lymph nodes of CCR9−/− mice compared with CCR9+/+ mice (Fig. 4⇓A). We previously observed that γδTCR+ cells that resemble recent thymic emigrants (i.e., CD44lowCD45RBlow) preferentially express CCR9 (21). However, no significant differences in CD44, CD45RB, CD62L, or αIEL integrin expression were observed on peripheral γδ-T cells from CCR9−/− and CCR9+/+ mice (data not shown). To ascertain whether γδ-T cell subsets were different in CCR9−/− and CCR9+/+ mice, we analyzed the percentages of TCR-Vγ1, -Vγ2, -Vγ5, or -Vδ4-expressing cells among total CD3+ T cells. TCR-Vγ1+, -Vγ2+ -Vγ5+, or -Vδ4+ cells appeared to be increased in the lymph node (Fig. 4⇓B) and spleen (data not shown) of CCR9−/− mice. On average, TCR-Vγ1+, TCR-Vγ5+, and TCR-Vδ4+ cells were increased more than TCR-Vγ2+ cells (TCR-Vγ1+, 3.1-fold; TCR-Vγ2+, 1.8-fold; TCR-Vγ5+, 2.9-fold; TCR-Vδ4+, 2.6-fold). In contrast, no significant difference in TCR-Vγ1, -Vγ2, -Vγ5, or -Vδ4 usage was detected in thymocytes from CCR9−/− and CCR9+/+ mice (data not shown).
γδ-T cells are increased in lymph nodes and spleen in CCR9−/− mice. A, Thymocytes, lymph node cells, and spleen cells were isolated and stained with mAbs directed to CD3, δTCR, and B220. For thymocytes, the percentage of CD3+δTCR+B220− cells among total cells is shown. For lymph node and spleen, CD3+B220− cells were gated, and the percentage of δTCR+ cells among total CD3+ cells is shown. The percentage of γδTCR+ cells was increased in lymph node and spleen of CCR9−/− mice compared with CCR9+/+ mice (p < 0.001). B, Lymph node cells were stained with mAbs directed to CD3, B220, and TCR-Vγ1, TCR-Vγ2, TCR-Vγ5, or TCR-Vδ4. CD3+B220− cells were gated, and the percentages of TCR-Vγ1-, TCR-Vγ2-, TCR-Vγ5-, or TCR-Vδ4-expressing cells among total CD3+ cells are shown. Open and closed circles represent data from CCR9+/+ and CCR9−/− mice, respectively.
Abnormal distribution of γδTCR+ IEL in CCR9−/− mice
Mucosal lymphocytes are composed of IEL and lamina propria lymphocytes. IEL consist of αβTCR+ and γδTCR+ cells, with γδTCR+ cells making up ∼50% of the total population in the small intestine and ∼20% of the total population in the large intestine (32). CCL25 and CCR9 mRNAs are detectable in duodenum and small intestine, but not in esophagus, stomach, appendix, and large intestine (Fig. 5⇓A). In addition, CCR9 mRNA is expressed in both αβTCR+ and γδTCR+ small intestinal IEL (21). The number of small intestinal IEL or large intestinal IEL was not statistically different in CCR9−/− and CCR9+/+ mice (data not shown). However, in CCR9−/− mice the percentage of γδTCR+ IEL was decreased in small intestine but increased in large intestine (Fig. 5⇓, B and C). There was no difference in CD2 and CD8α expression on large intestinal IEL from CCR9−/− and CCR9+/+ mice (data not shown). In addition, small intestinal γδTCR+ IEL from CCR9−/− mice resembled those from CCR9+/+ mice in that they were uniformly CD8α+ and expressed high levels of αIEL integrin (data not shown).
Reduction in small intestinal IEL in CCR9−/− mice. A, Northern blot determination of CCL25 and CCR9 mRNA expression in gastrointestinal tract. Twenty micrograms of total RNA from various tissues was separated by gel electrophoresis, transferred to membranes, and hybridized with cDNA probes to mouse CCL25, mouse CCR9, or human EF-1α. Lane 1, thymus; lane 2, esophagus; lane 3, stomach; lane 4, duodenum; lane 5, small intestine; lane 6, appendix; lane 7, large intestine. B, Representative FACS analysis of purified IEL from small and large intestines. Small and large intestinal IEL were isolated from CCR9+/+ or CCR9−/− mice. The cells were stained with mAbs to CD3, βTCR, and δTCR and then analyzed by FACS. CD3+ cells were gated and analyzed for βTCR and δTCR expression. Values indicate the percentage of cells in each quadrant. C, γδTCR+ IEL are decreased in the small intestine of CCR9−/− mice relative to CCR9+/+ mice (p < 0.01). The percentages of γδTCR+ IEL among CD3+ small intestinal IEL were plotted. ○, Data from CCR9+/+ mice; •, data from CCR9−/− mice. D, Reduction in TCR-Vδ4+ small intestinal IEL in CCR9−/− mice relative to CCR9+/+ mice (p < 0.01). Small intestinal IEL from CCR9+/+ and CCR9−/− mice were stained with mAbs against CD3, δTCR, and TCR-Vγ1, TCR-Vγ2, or TCR-Vγ5. CD3+δTCR+ cells were gated, and the percentages of TCR-Vγ1-, TCR-Vγ2-, and TCR-Vγ5-expressing cells were analyzed. Because mAbs directed to δTCR and TCR-Vδ4 compete for binding, CD3+βTCR− IEL were gated to determine the percentage of TCR-Vδ4+ cells. E, TCR-Vγ1/Vδ4+ and TCR-Vγ5/Vδ4+ IEL are decreased in CCR9−/− mice. Small intestinal IEL from CCR9+/+ and CCR9−/− mice were stained with mAbs directed to CD3, βTCR, TCR-Vδ4, and TCR-Vγ1, or TCR-Vγ5. CD3+βTCR− cells were gated and analyzed for TCR-Vδ4 and TCR-Vγ1 or TCR-Vγ5 expression. Values indicate the percentage of cells in each quadrant. F, Reduction in αβTCR+CD8αβ+ small intestinal IEL in CCR9−/− mice relative to CCR9+/+ mice (p < 0.05). Small intestinal IEL from CCR9+/+ and CCR9−/− mice were stained with mAbs directed to βTCR, CD4, CD8α, and CD8β. βTCR+ cells were gated and analyzed for the percentages of CD8αα (CD8α+CD8β−), CD8αβ (CD8α+CD8β+), CD4−CD8α+, CD4+CD8α+, and CD4+CD8α− cells in αβTCR+ IEL. ○, Data from CCR9+/+ mice; •, data from CCR9−/− mice.
Although the TCR-Vγ/Vδ repertoire is diverse in γδTCR+ IEL, the γδTCRs expressed by small intestinal IEL consist predominantly of TCR-Vγ1 or -Vγ5 paired with TCR-Vδ4, -Vδ5, -Vδ6, or -Vδ7 (33, 34). To assess whether there were any differences in the TCR-Vγ/Vδ repertoire of γδTCR+ IEL in small intestine, we analyzed γδTCR+ IEL in CCR9−/− and CCR9+/+ mice for the expression of TCR-Vγ1, -Vγ2, -Vγ5, and -Vδ4 by flow cytometry. The percentages of TCR-Vγ1, -Vγ2, and -Vγ5-bearing cells among γδTCR+ IEL were similar in CCR9−/− and CCR9+/+ mice (Fig. 5⇑, D and E). However, the percentage of TCR-Vδ4-bearing IEL was markedly decreased in CCR9−/− mice (Fig. 5⇑, D and E). Most TCR-Vδ4+ IEL coexpress TCR-Vγ5 and, to a lesser extent, TCR-Vγ1. Consistent with this observation, both TCR-Vγ5/Vδ4+ and TCR-Vγ1/Vδ4+ IEL subsets were decreased in CCR9−/− mice (Fig. 5⇑E). Collectively, these data indicate that CCR9 is involved in the generation or maintenance of γδTCR+ IEL and, in particular, TCR-Vδ4+ IEL in small intestine.
We next evaluated αβTCR+ IEL in the small intestine of CCR9−/− and CCR9+/+ mice. αβTCR+ IEL subsets can be distinguished by the expression of CD8α, CD8β, and CD4. The percentage of αβTCR+ IEL that were CD4+CD8α− and CD4+CD8+ was not consistently different in CCR9−/− and CCR9+/+ mice (Fig. 5⇑F). However, the percentage of CD8αβ+ IEL was reduced in CCR9−/− mice (Fig. 5⇑F). No significant differences were observed in the number or subset distribution of small intestinal lamina propria lymphocytes in CCR9−/− and CCR9+/+ mice (data not shown).
Normal B lymphopoiesis in CCR9−/− mice
Northern blot analysis revealed that CCR9 mRNA is expressed in bone marrow from B6 mice (Fig. 6⇓A). Rag-1−/− bone marrow cells, which lack mature T and B cells, had an equivalent level of CCR9 mRNA expression, indicating that CCR9 is expressed on lymphoid progenitor cells and/or myeloid cells. In contrast to CCR9 expression, CCL25 mRNA was undetectable in the bone marrow (Fig. 6⇓A). Bowman et al. (23) described an immature population of bone marrow cells that migrates in response to CCL25. This population is phenotypically similar to DX5−CD19−B220+ bone marrow cells (35, 36). DX5−CD19−B220+ cells can be further subdivided on the basis of CD4 surface expression into B cell precursors (CD4−DX5−CD19−B220+) and cells of unknown lineage and potential (CD4+DX5−CD19−B220+) (37). Significantly, staining of bone marrow cells with anti-CCR9 Ab revealed high level expression of CCR9 on CD4+DX5−CD19−B220+ cells, but only very low levels of CCR9 on CD4−DX5−CD19−B220+ cells (Fig. 6⇓B). Both populations migrated in response to CCL25 (data not shown). No statistically significant differences in the number or the percentage of CD4+DX5−CD19−B220+ and CD4−DX5−CD19−B220+ bone marrow cells were observed in CCR9−/− and CCR9+/+ mice (Fig. 6⇓C). In addition, CCR9−/− mice contained normal numbers of bone marrow cells and peripheral B cells and contained normal proportions of pro-B (CD43+IgM−B220low), pre-B (CD43−IgM−B220low), immature B (CD43−IgM+B220low), and recirculating bone marrow (CD43−IgM+B220high) B cells and exhibited normal IgM and IgD surface profiles on splenic B cells (Fig. 6⇓C and data not shown). Thus, CCR9 is not essential for normal B cell development.
CCR9 expression in bone marrow and analysis of B cell development in CCR9−/− mice. A, CCR9 mRNA expression in bone marrow. Total RNA (20 μg) from bone marrow cells from B6 or Rag-1−/− mice was separated by gel electrophoresis, transferred to membranes, and hybridized with cDNA probes to mouse CCL25, mouse CCR9, or human EF-1α. B, Immature bone marrow cells express CCR9. Bone marrow cells were stained with Abs directed to CD4, DX5, CD19, B220, and CCR9. Gated CD4+DX5−CD19−B220+ and CD4−DX5−CD19−B220+ cells were analyzed for surface CCR9 expression. C, Normal development of early B progenitor cells in CCR9−/− mice. Bone marrow cells from CCR9+/+ or CCR9−/− mice were stained with mAbs directed to B220, CD43, DX5, CD19, and CD4. CD19−B220+ were gated and analyzed for CD4 and DX5 expression. The percentages of CD43+B220+ and CD19−B220+ cells in total bone marrow cells, and the percentages of CD4−DX5− and CD4−DX5− subsets among CD19−B220+ cells are indicated.
Reduced thymus repopulating activity of CCR9−/− bone marrow cells
As shown in Fig. 2⇑, no obvious defects in αβ-T cell development were observed in adult CCR9−/− mice. Bleul et al. (12) reported previously that CCL25 is expressed in the early thymic anlage of the mouse fetus, and that fetal blood prothymocytes (Thy1+c-kitlow) respond to CCL25. These findings suggest that CCL25/CCR9 may be involved in the migration of prothymocytes into the thymus. To explore this possibility further, we examined CCR9 expression on immature thymocyte populations from adult mice. CCR9 expression was undetectable on CD3−CD4−CD8− (triple-negative (TN)) thymocytes from adult mice, including the most immature (CD44+CD25− TN) subset, and adult TN thymocytes cells failed to migrate in response to CCL25 (21) (data not shown). In addition, the number and distribution of TN thymocyte subsets in CCR9−/− and CCR9+/+ mice, as defined by the expression of CD44 and CD25, were similar (data not shown). Thymus size and cellularity can be normal even if the number of immature T progenitor cells in the thymus is reduced, presumably because these cells are capable of expanding (1). Consequently, examination of thymocytes in the adult steady state condition may not reveal a potential defect in the progenitor cell population. To determine whether the loss of CCR9 affects the migration of bone marrow progenitor cells into the thymus or the establishment or retention of T progenitor cells in the thymus, we performed a competitive transplantation experiment. Total bone marrow cells from CCR9−/− mice (CD45.2) and B6.CD45.1 (CCR9+/+) mice were mixed in different ratios and injected into lethally irradiated Rag-1−/− (CD45.2) mice. As a control, identical mixtures of bone marrow cells from littermate CCR9+/+ (CD45.2) and B6.CD45.1 mice were injected into irradiated Rag-1−/− (CD45.2) mice. One to 2 mo after the bone marrow transplantation, the number and percentage of CD45.2+ thymocytes and peripheral (lymph node) B cells were determined. Significantly, when mixtures of CCR9−/− and B6.CD45.1 bone marrow were injected, the percentage of thymocytes derived from CCR9−/− bone marrow cells was consistently lower than the expected ratio (Fig. 7⇓, A and C). In contrast, the percentage of lymph node B cells derived from CCR9−/− bone marrow cells was consistently close to the expected ratio (Fig. 7⇓D). The reduction in CCR9−/− bone marrow-derived thymocytes was evident at both the CD44+/−CD25+ TN and the CD44−CD25− TN stage (Fig. 7⇓B), indicating that cells from CCR9−/− mice were competitively disadvantaged at or before the CD44+/−CD25+ TN stage. Taken together, these data indicate that under competitive conditions, bone marrow cells from CCR9−/− mice are impaired in their ability to reconstitute T cell, but not B cell, development in irradiated Rag-1−/− mice.
Competitive bone marrow transplantation into irradiated Rag−/− mice. Bone marrow cells were isolated from femurs of CCR9+/+ (CD45.2), CCR9−/− (CD45.2), and B6.SJL-Ptprca/BoAiTac mice (B6.CD45.1; Taconic Farms). A total of 2 × 106 cells consisting of mixtures of CCR9+/+ or CCR9−/− bone marrow with B6.CD45.1 bone marrow was injected i.v. into lethally irradiated (9.5 Gy) Rag-1−/− B6 mice (CD45.2). One to 2 mo after transplantation, thymocytes and lymph node B cells were examined for the expression of CD45.2 by FACS. A, Representative FACS plots showing the expression of CD4, CD8α, and CD45.2 on thymocytes from irradiated Rag-1−/− mice injected with equal numbers of bone marrow cells from CCR9+/+ and B6.CD45.1 mice (upper left panel) or CCR9−/− and B6.CD45.1 mice (upper right panel). Histograms show the expression of CD45.2 on total thymocytes. B, Representative FACS plots showing the expression of CD44 and CD25 on gated (CD4−CD8−CD3− (TN)) thymocytes from irradiated Rag-1−/− mice injected with equal numbers of bone marrow cells from CCR9+/+ and B6.CD45.1 mice (upper left panel) or CCR9−/− and B6.CD45.1 mice (upper right panel). Histograms show CD45.2 expression on CD44+/−CD25+ or CD44−CD25− thymocytes. C, Plot of the percentage of CD45.2+ cells in total thymocytes from irradiated Rag−/− mice injected with the indicated ratios of bone marrow cells. The percentage of thymocytes derived from CCR9−/− bone marrow cells was significantly lower than that of CCR9+/+ bone marrow-derived thymocytes (p < 0.01). D, Plot of the percentage of CD45.2+ cells in lymph node B cells (B220+) from irradiated Rag-1−/− mice injected with the indicated ratios of bone marrow cells. ○, Data from mice injected with mixtures of bone marrow from CCR9+/+ (CD45.2) and B6.CD45.1 mice; •, data from mice injected with mixtures of bone marrow from CCR9−/− (CD45.2) and B6.CD45.1 mice.
Discussion
In this study we examined the role of CCR9 in T and B cell development by generating CCR9-deficient mice by gene targeting. CCR9 is expressed on the surface of most αβ-lineage thymocytes and approximately half of all γδTCR+ thymocytes and T cells (20, 21). Mucosal T cells also express CCR9, and fetal blood prothymocytes and pre-pro-B cells migrate in response to CCL25 (11, 12, 13, 14, 15, 21, 23). Analysis of CCR9−/− mice revealed that 1) αβ-T cells develop normally in adult CCR9−/− mice, but CCR9−/− bone marrow cells exhibit a reduced capacity to repopulate the thymus of irradiated Rag-1−/− mice under competitive conditions compared with CCR9+/+ bone marrow cells; 2) CCR9−/− mice contain increased numbers of peripheral γδ-T cells, but reduced numbers of small intestinal γδTCR+ IEL; and 3) B cell development is unaffected in CCR9−/− mice.
The finding that αβ-T cell development and thymocyte selection appear unperturbed in adult CCR9−/− mice was unexpected given the fact that most thymocytes express high levels of CCR9 and migrate to CCL25. The inability of CCR9−/− thymocytes to respond to CCL25 demonstrates that these cells do not express another receptor for CCL25. Several different chemokines are expressed in the thymus, and it is possible that they may share overlapping targets and therefore compensate for the loss of CCR9. Histological examination of the thymus revealed no obvious abnormalities in CCR9−/− mice; however, more detailed localization studies of specific thymocyte subsets may be required to detect subtle intrathymic migration defects.
Previous results indicate that prothymocytes in fetal blood respond to CCL25 (12). This observation together with the finding that CCL25 is not expressed in bone marrow (Fig. 6⇑A) suggested that CCR9 may participate in the migration of T-progenitor cells from bone marrow to thymus. Although we were unable to detect CCR9 surface expression on immature TN thymocytes from adult mice, and these cells did not respond to CCL25, CCR9+ cells were present in adult bone marrow (Fig. 6⇑, A and B). CD4−DX5−CD19−B220+ bone marrow cells, which include early B progenitors (37, 38), expressed low levels of CCR9 and could respond to CCL25 (Fig. 6⇑B and data not shown). In addition, CD4+DX5−CD19−B220+ cells were found to express high levels of CCR9 and could respond to CCL25. The lineage affiliation and differentiation potential of CD4+DX5−CD19−B220+ cells remain unclear, although phenotypically similar cells from bone marrow that respond to CCL25 contain both B and T cell progenitors (36, 37, 39, 40). To investigate whether CCR9 is involved in regulating the migration of T-progenitor cells to thymus, we performed competitive bone marrow transplantation experiments. The results demonstrated that CCR9−/− bone marrow cells are competitively disadvantaged compared with CCR9+/+ bone marrow cells in their ability to repopulate the thymus of irradiated Rag-1−/− mice (Fig. 7⇑). These findings suggest three nonmutually exclusive possibilities: 1) that CCR9 regulates the generation of prothymocytes in the bone marrow, 2) that CCR9 regulates the migration of prothymocytes into the thymus or their migration or retention within the thymus, and 3) that CCR9 regulates the proliferation of early thymocytes in the thymus. Because immature TN thymocytes do not express surface CCR9 and do not respond to CCL25 (21), it appears unlikely that CCR9 is directly involved in the expansion of early thymocytes. Bleul et al. (12) reported that three chemokines (CCL25, CXCL12, and CCL21) are expressed in the thymic anlage, and both CCL25 and CXCL12 attract fetal blood prothymocytes. Similar to CCR9−/− mice, mice deficient in CXCL12 or its receptor, CXCR4, also showed no obvious abnormality in T cell development (41, 42, 43). On the basis of these findings, we speculate that CCR9 in addition to other chemokine receptors such as CXCR4 may play an important and partially redundant role in regulating the migration of prothymocytes into the thymus. Our inability to detect surface expression of CCR9 on immature TN thymocyte subsets is not necessarily in conflict with this idea, as the population of CCR9+ prothymocytes may be extremely small, or CCR9 may be down-regulated when prothymocytes enter the thymus.
γδ-T cell development and/or homeostasis were clearly altered in CCR9−/− mice. Although CCR9−/− mice contained normal numbers of γδTCR+ thymocytes, the number of γδ-T cells was increased in secondary lymphoid organs (spleen and lymph nodes; Fig. 4⇑A). All TCR-Vγ/Vδ pairs examined were increased in lymph nodes and spleen of CCR9−/− mice (Fig. 4⇑B and data not shown). In addition, we could not detect any significant difference in the expression of CD44, CD45RB, and other surface markers (e.g., CD62L and αIEL integrin) on peripheral γδ-T cells from CCR9−/− and CCR9+/+ mice (data not shown), suggesting that the increased number of peripheral γδ-T cells in CCR9−/− mice is not due to the accumulation of one particular subpopulation of γδ-T cells. The kinetics of thymocyte development are much more rapid for γδ- than αβ-T cells, and γδ-T cells appear to be dependent on the thymic environment for a relatively brief period during their development (44). Thus, in the absence of CCR9, γδ-T cells may be generated in higher numbers and immigrate more rapidly into the periphery.
In contrast to lymph nodes and spleen, we observed that the percentage of small intestinal γδTCR+ IEL was consistently decreased in CCR9−/− mice. In the gastrointestinal tract, CCL25 expression is restricted to the small intestinal epithelium (Fig. 5⇑A) (11, 13, 14, 15). Interestingly, although the percentage of γδTCR+ IEL in the small intestine of CCR9−/− mice was decreased, we found that the percentage of γδTCR+ IEL in the large intestine was increased (Fig. 5⇑, B and C). The large intestinal γδTCR+ IEL in CCR9−/− mice were phenotypically similar to those in CCR9+/+ mice (i.e., they did not resemble small intestinal IEL; data not shown), indicating that the increase in large intestinal γδTCR+ IEL is not due to migration of cells from the small intestine to the large intestine.
All IEL from human small intestine express CCR9 and respond to CCL25 (13, 14, 15), and in mice both γδTCR+ and αβTCR+ IEL have been shown to express CCR9 mRNA (21). Thus, CCR9 might be important for the recruitment of mature γδ-T cells to the intestinal mucosa. Indeed, the observation that large intestinal IEL were not decreased in CCR9−/− mice localizes the defect to the site of CCL25 production (Fig. 5⇑, A and B). Consistent with this idea is the finding that small intestinal αβTCR+CD8αβ+ IEL, which are thought to be thymically derived and would therefore migrate from the periphery to the small intestine, are also reduced in CCR9−/− mice (Fig. 5⇑F). Another possibility is that in the absence of CCR9, γδTCR+ IEL fail to be retained in the small intestine. However, the preferential loss of TCR-Vγ5/δ4+ IEL and our inability to find these cells in lymph node or spleen (Fig. 5⇑E and data not shown) suggest that there may be a direct role for CCR9 in Vγ5/δ4+ IEL development or that these cells may fail to survive if they are unable to home to the proper site. The accumulation of diverse subsets of γδ-T cells in the periphery of CCR9−/− mice could also reflect the pool of cells that initially migrate to the intestine and may differ significantly from the population that ultimately becomes established as IEL. Finally, the increase in peripheral γδ-T cells may reflect a shift in a dynamic equilibrium between the small intestine and the peripheral pool. The lower numbers of γδTCR+ IEL in CCR9−/− mice even in the face of what may be a compensatory increase in these cells in the periphery suggests an important role for CCR9 in establishing this equilibrium.
CCR9 and CCL25 mRNA are also detected in the small intestine of Rag-1−/− mice, which lack mature lymphocytes (21), raising the possibility that CCR9 plays a role in early mucosal T cell development and/or recruitment of IEL precursors to the small intestine. Cryptopatches are multiple clusters of c-kit+IL-7R+Thy1+ lymphocytes located in the crypt lamina propria of the murine intestine (45). CCR9 may be involved in cryptopatch formation and/or extrathymic γδTCR+ IEL development, perhaps by regulating the migration of progenitor cells from the fetal liver, fetal thymus, or adult bone marrow to the small intestine. γδTCR+ IEL are greatly reduced in nude mice and neonatal thymectomized mice (46). Transplantation of fetal or neonatal thymus, but not adult thymus, into nude mice results in the generation of a substantial number of thymically derived γδTCR+ IEL (46, 47). These data indicate that the thymus is involved in the development of γδTCR+ IEL, but the mechanism by which the thymus participates is still unclear. The thymus may provide IEL precursors or a thymus-derived factor that is required for the differentiation and/or expansion of IEL (48), and CCR9 could be important for these functions.
In summary, these results identify a role for CCR9 in αβ- and γδ-T cell development that may include regulating the migration of progenitor cells to specific sites of T lymphopoiesis. The selective expression of CCL25 in the small intestine and the deficiency of small, but not large, intestinal γδ-T cells also suggest the existence of distinctive mechanisms of lymphocyte recruitment that may permit functional specialization of immune responses in different segments of the gastrointestinal tract.
Acknowledgments
We gratefully acknowledge L. Lefrancois for providing anti-TCR-Vγ1 and anti-TCR-Vγ5 mAbs. We thank S. Hayes, C. Feng, and K. Laky for valuable advice and D. El-Khoury for excellent technical assistance.
Footnotes
-
↵1 This work was supported in part by an Intramural AIDS Targeted Antiviral Program grant. S.U. is supported by Japan Society for the Promotion of Science Research Fellowships for Japanese Biomedical and Behavioral Researchers at the National Institutes of Health.
-
↵2 Address correspondence and reprint requests to Dr. Paul E. Love, Laboratory of Mammalian Genes and Development, National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, MD 20892. E-mail address: pel{at}helix.nih.gov
-
↵3 Abbreviations used in this paper: MIP, macrophage-inflammatory protein; DN, double negative; DP, double positive; IEL, intraepithelial lymphocyte; TN, triple negative; ES cell, embryonic stem cell.
- Received November 27, 2001.
- Accepted January 18, 2002.
- Copyright © 2002 by The American Association of Immunologists