| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
,
* The Center for Blood Research and
Department of Pathology, Harvard Medical School, Boston, MA 02115; and
Institute of Pathology and
Department of Surgery, University of Oslo, Rikshospitalet, Norway
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
2 integrin LFA-1 (CD11a/CD18), which interact with peripheral node addressin (PNAd), the lymphoid chemokine CC chemokine ligand (CCL)21 (TCA-4/SLC/6C-kine/exodus 2) and ICAM-1/-2, respectively (2). HEV in PP express CCL21 and ICAM-1/-2, but not PNAd. Instead, they express mucosal addressin cell adhesion molecule (MAdCAM)-1, a member of the Ig superfamily that interacts with the integrin 
47 and, via its mucin domain, with L-selectin (3).
In contrast to their migration through secondary lymphoid organs, naive T cells are normally excluded from extralymphoid effector sites. However, if inflammation exists for a prolonged time, such as in certain autoimmune diseases, large numbers of naive lymphocytes are frequently found in the affected tissue. These cells become organized in LN-like aggregates, which contain HEV that express PNAd and/or MAdCAM-1 (reviewed in Ref. 4). It has been suggested that the induction of these vascular addressins in newly formed HEV is important for the formation of such tertiary lymphoid tissue (a process called lymphoid neogenesis) by facilitating naive T cell entry (4). However, thus far, the adhesive behavior of naive T cells has not been specifically examined in nonlymphoid microcirculation, and it is unclear whether HEV are a prerequisite for, or consequence of, lymphoid neogenesis. Indeed, endothelial cells in acutely inflamed tissues express functional L-selectin ligands, which are distinct from PNAd and contribute to the recruitment of inflammatory cells (5, 6). Moreover, naive T cells express a second chemokine receptor, CXCR4, whose ligand CXC chemokine ligand (CXCL)12 (stromal cell-derived factor-1) is constitutively expressed by endothelial cells in many normal tissues and even up-regulated upon inflammation (7, 8, 9). Thus, naive T cells should be capable of interacting with L-selectin ligands and CXCL12 on inflamed microvessels. However, naive T cells are not usually part of the leukocyte infiltrate in acute inflammation. Thus, the critical step(s) of the adhesion cascade that either prohibit(s) or enable(s) naive T cell entry into acutely and chronically inflamed nonlymphoid sites, respectively, remain(s) to be identified.
Recent studies have implicated chemokines, such as CCR7 ligands, in lymphoid neogenesis; transgenic expression of CCL21 or, to a lesser degree CCL19, led to the formation of LN-like structures in pancreatic islets (10, 11, 12). Interestingly, lymphoid neogenesis was absent in CCL21-transgenic mice that were lymphocyte-deficient, but was still evident in animals that were deficient in B but not T cells, indicating that the effect of ectopically expressed CCL21 is T cell-dependent (10, 11). These studies elegantly demonstrate that certain transgenically expressed chemokines are sufficient to mediate naive T cell accumulation within nonlymphoid tissues in vivo. However, a drawback of these reports is that the transgenic chemokines were continuously present during development and adult life, and therefore could have had indirect effects on lymphocyte migration. Thus, conclusions about their effects on naive T cell recruitment during homeostatic and acute and chronic inflammatory conditions in adult animals cannot be drawn. These questions are of importance, as recent evidence suggests that Ag can be presented to naive T cells in tertiary lymphoid tissues (13). Such priming of T cells in nonlymphoid sites might be of benefit for tumor therapy (13), but could also be involved in the pathogenesis or progression of autoimmune diseases, because locally produced autoantigens may be presented to naive T cells directly at the site of inflammation. Thus, a better understanding of the mechanisms of lymphoid neogenesis/naive T cell trafficking to nonlymphoid tissues could facilitate the development of new therapies for human disease.
We have determined the minimal requirements for naive T cell entry into nonlymphoid tissues in adult animals. We found that these cells underwent selectin-mediated rolling, but were incapable of firm adherence in acutely inflamed peripheral tissue. Exogenous application of CCL21, but not CXCL12, rapidly induced accumulation of naive T cells, even in the absence of HEV or vascular addressins. Consistent with this finding, we found that CCL21 is expressed on PNAd+ and PNAd- blood vessels in human diseases accompanied by lymphoid neogenesis. Importantly, the presence of naive T cells in perivascular areas correlated with CCL21 expression in blood vessels. These results suggest that de novo expression of CCL21 on endothelial cells could represent a key step for naive T cell recruitment in the periphery. In addition, these findings implicate CCL21 in the pathogenesis of those autoimmune diseases associated with lymphoid neogenesis, and suggest the CCL21/CCR7 pathway as a potential therapeutic target in these diseases.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
T-GFP mice, in which green fluorescent protein (GFP) is expressed under the control of the murine CD4 proximal enhancer and promoter without the intronic silencer, have been described previously (14). Mice were housed and bred in a specific pathogen-free/virus Ab-free animal facility. All experiments were in accordance with National Institutes of Health guidelines and approved by the Committees on Animals of both Harvard Medical School and The Center for Blood Research (Boston, MA).
Abs and chemokines
Hybridoma cells for mAb 9A9 (rat IgG2b) and mAb 5H1 (rat IgG1), which neutralize mouse E-selectin and P-selectin, respectively, and the nonbinding control mAb 2-4A1 (rat IgG2b) were kindly provided by Dr. B. Wolitzky (Hoffman LaRoche, Nutley, NJ). Hybridoma cells for anti-L-selectin mAb Mel-14 (rat IgG2a), anti-PNAd mAb MECA-79 (rat IgM), and anti-LFA-1 mAb Tib 213 (rat IgG2b) were a kind gift of Dr. E. Butcher (Stanford University, Stanford, CA). The following Abs were used: anti-CD3, anti-CD44, and anti-CD62L (BD PharMingen, San Diego, CA); anti-human CD45RA (clone L48), anti-human HLA-DR (L243), anti-human CD11c (SHCL-3), and anti-human CD45RA (HI100; BD Biosciences, Mountain View, CA); anti-human CD3, anti-human von Willebrand factor (vWF), and anti-human CD8 (DK25; DAKO, Carpinteria, CA); anti-human CD4 (MAT-4; Medscand Diagnostics, Malmoe, Sweden); anti-human CCL21 (R&D Systems, Minneapolis, MN); and anti-human mAb PAL-E (Research Diagnostics, Flanders, NJ). Mouse mAb RIV9 against human CD3 (IgG3) and FITC-conjugated goat anti-human vWF were gifts from J. Hilgers (Bioprobe, Amsterdam, The Netherlands) and Dr. E. Butcher, respectively. Recombinant murine TNF, CXCL12, CCL19, and CCL21 were purchased from R&D Systems. For some experiments, human recombinant CCL21 (kindly provided by T. Springer, Center for Blood Research, Boston, MA) was used. No difference in chemotactic response of naive murine T cells to murine or human CCL21 was observed (data not shown). The use of the CCL19 (EBI1 ligand chemokine/macrophage-inflammatory protein-3)-Ig chimera to detect CCR7 on murine T cells was described previously (15).
Intravital microscopy and image analysis
Cremaster muscles were prepared as described previously (16). Briefly, adult male T-GFP mice were anesthetized by i.p. injection of 0.25 ml of a combination of ketamine HCl (5 mg/ml) and xylazine (1 mg/ml). The right cremaster muscle was prepared and covered with sterile, bicarbonate-buffered Ringer’s injection solution (pH 7.4). The surgical procedure as described in this study takes 
30 min. GFP+ leukocytes were visualized through a x40 water-immersion objective (Zeiss Achroplan NA 0.75 
; Oberkochen, Germany) by video-triggered stroboscopic epi-illumination on an intravital microscope (IV-500; Mikron Instruments, San Marcos, CA). One to three venular trees were chosen and 1-min recordings were made of individual segments of several postcapillary and small collecting venules at 5-min intervals to assess baseline rolling. Subsequently, 50 µg of mAb (9A9, 5H1, Mel-14, Tib-213, or 2-4A1) were injected i.v. Fifteen minutes later, the same venules were repeatedly recorded for 1–3 min intervals until 45 min after mAb injection. To study the effects of chemokines on T cell adhesion, one to two vessel trees in a preparation were chosen. To identify sticking cells, individual venular branches were recorded for 1 min each throughout an entire vessel tree. The superfusion buffer was then replaced with prewarmed buffer alone or buffer containing chemokines (100 nM). The same vessel segments were recorded again 5–25 min and 45–60 min after buffer replacement. For some experiments, TNF- (500 ng) was injected intrascrotally 2–3 h before the preparation.
The rolling fraction for each individual venule was determined as the percentage of leukocytes that interacted detectably with the vascular wall within the total number of fluorescent cells that passed a vessel during the observation period. Interactions were defined as any transient or continuous slowing of a GFP+ cell below the velocity of noninteracting blood cells in the same vessel. Firm adherence (sticking) was defined as a cell that remained stationary for 
30 s. Vessel cross-sectional diameters (D) and velocities of individual rolling (Vroll, at least 10 consecutive cells per venule) and noninteracting (Vfree, at least 20 consecutive cells per venule) leukocytes, as well as wall shear rate (WSR) and wall shear stress, were determined off-line as described previously using a PC-based interactive image analysis system (17, 18).
Induction of air pouches
Air pouches were generated by s.c. injection of sterile air (3–5 ml initially, 3 ml after 72 h) in the back of a T-GFP mouse as described (19). After 6 days, PBS alone or PBS containing TNF- (500 ng), CXCL12, CCL19 (1 or 2 µg), or CCL21 (2 µg) were injected into the air pouch at time points 0 and 4 h. Accumulated leukocytes were harvested from the pouch 8 h after the first injection by PBS lavage.
Immunofluorescence staining and quantification of naive T cells and CCL21+ blood vessels
Serial cryosections (8 µm) of inflamed human tissue from various inflammatory lesions including rheumatoid arthritis (RA) (n = 4), ulcerative colitis (UC) (n = 5), and psoriasis (n = 2) were used for analysis. Control material consisted of palatine tonsils (n = 3), normal colon (n = 3), and normal skin (n = 2) obtained from individuals undergoing tonsillectomy, colectomy (due to long-lasting chronic obstipation), or mastectomy, respectively. All procedures at Rikshospitalet involving patient material were performed in agreement with the Helsinki Declaration and were approved by the Regional Committee for Medical Research Ethics (Health Region South, Oslo, Norway). Patient data, pathohistological grading, and treatment schedules are provided in Table I. Three-color immunostaining with primary Abs and appropriate second- and third-step reagents was performed as described previously (20).
|
In situ hybridization for CCL21
A 345-bp digoxigenin (DIG)-labeled riboprobe was generated from the coding region of cDNA for human CCL21 with the DIG RNA labeling kit according to the manufacturer’s directions (Boehringer Mannheim, Mannheim, Germany). Hybridization and detection of the hybridized probe were performed as previously described (20).
Statistical analysis
Where appropriate, data are presented as mean ± SEM. Rolling fractions after Ab treatment vs baseline were compared using the Wilcoxon signed rank test. Sticking fractions before vs after superfusion with chemokines were compared using ANOVA. The effect of TNF on rolling velocities was analyzed using the Mann-Whitney U test. Accumulation of naive and central memory T cells in air pouches in response to chemokines and TNF- was analyzed using ANOVA. Significance was assumed at p < 0.05.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In the T-GFP strain, transgenic GFP is strongly and selectively expressed by T cells, making them easily detectable by flow cytometry and intravital fluorescence microscopy (Fig. 1A; Refs. 14, 15, 21). To specifically assess the adhesive behavior of naive T cells in vivo, it is necessary to distinguish them from Ag-experienced T cells, which migrate readily to nonlymphoid sites (22, 23). T-GFP mice allow this distinction to some extent, because prolonged antigenic stimulation (for 5 days) induces complete loss of GFP in >90% of CD8+ and 50% of CD4+ effector cells (14, 15). However, we also found that in vitro-generated CD8+ T cells with phenotypic and functional properties of central memory cells (24) maintain GFP expression, even after adoptive transfer in vivo (15, 25). Because it is not possible during intravital microscopy experiments to determine whether a GFP+ cell is naive or Ag-experienced, it was important to fully characterize GFP+ cells in the peripheral blood of T-GFP mice.
|
8 wk) and aged (1.5 years) mice, respectively. This difference in frequency notwithstanding, the surface phenotype of each memory population remained remarkably constant irrespective of the animals’ age (not shown). As mice used in this study for intravital microscopy were young (6–10 wk), only 5% of all GFP+ cells were CD44high memory cells. Naive T cell behavior in cremaster muscle venules
Having established that the vast majority (95%) of detectable (i.e., GFPhigh) cells in the bloodstream of healthy young T-GFP mice are naive T cells, we set out to dissect the molecular mechanisms by which these cells interact with microvessels in nonlymphoid tissues. We used intravital microscopy of the cremaster muscle microvasculature, a model of trauma-induced, moderate inflammation (26, 27). Rolling was defined as any transient or continuous interaction of GFP+ cells below the velocity of noninteracting GFP+ cells in the same vessel. The baseline rolling fraction of GFP+ cells in cremaster muscle venules (n = 46 venules/12 animals) was 34 ± 2% and did not correlate with the venular diameter (Fig. 2A). To determine which adhesion molecules mediated naive T cell rolling in this setting, we compared the frequency of rolling GFP+ cells before and after i.v. injection of neutralizing mAbs (Fig. 2B). Hemodynamic parameters before and after Ab treatment (50 µg/mouse) revealed no significant differences (Table II). Treatment with nonbinding isotype control mAb 2-4A1 (n = 11/3), with anti-LFA-1 (n = 16/3), or with anti-E-selectin (n = 11/3) did not alter rolling fractions (p > 0.05). A moderate reduction in rolling was induced by anti-P-selectin mAb (41% inhibition; p < 0.01; n = 12/3). In contrast, rolling was nearly abolished by anti-L-selectin mAb (97% inhibition; p < 0.001; n = 10/3).
|
|
2–3 h before the experiment did not change rolling fractions of GFP+ cells (32 ± 2%; n = 11/4) compared with control conditions (Fig. 2C). However, while the median rolling velocity (Vroll) of GFP+ cells was 170.8 µm/s (n = 359 cells/24 venules/12 animals) in the absence of cytokine (Fig. 2D), TNF- pretreatment decreased this parameter to 97.5 µm/s (p < 0.001; n = 90/6/3). CCL21 induces sticking of GFP+ cells in cremaster muscle microvasculature
Having established that naive T cells roll in cremaster muscle venules, we next determined whether they also undergo sticking (defined as stationary arrest for 30s) when exposed to inflammatory stimuli (Fig. 3). Under baseline conditions, firmly adherent GFP+ cells were rare (3.5 ± 1.4 stickers/mm2; n = six venular trees/three animals). Pretreatment of cremaster muscles with TNF- did not increase sticking of GFP+ cells (2.4 ± 0.8 stickers/mm2; n = 8/4; p > 0.05), even though intravital microscopy under transmitted light revealed that TNF- induced a marked accumulation of GFP- (mostly myeloid) leukocytes.
|
35%, corresponds to a sticking fraction (i.e., sticking cells in percent of rolling cells) of 1%. CCL19 superfusion of the cremaster muscle induced firm adherence of GFP+ cells similarly to CCL21 (n = 3/2; 70.0 ± 17.1 GFP+ cells/mm2 at 45 min; Fig. 3). In contrast, superfusion with CXCL12 did not significantly induce sticking of GFP+ cells as compared with baseline values in the same vascular bed (n = 6/3; 12.8 ± 3.7 GFP+ cells/mm2; Fig. 3). However, when fluorescently labeled in vitro-generated central memory CD8+ T cells were injected into cremaster muscles, CXCL12 superfusion at equimolar concentration (100 nM) induced firm adherence of these cells, suggesting that the chemokine reached a concentration in the venular lumen that was sufficient to induce arrest of more responsive leukocyte subsets (L. Scimone and U. H. von Andrian, unpublished observation). We next set out to determine which adhesion receptor mediated CCL21-induced firm adherence of GFP+ cells. As the integrin LFA-1 (CD11a) is the major molecule mediating sticking of naive T cells in PLN HEV (28), we injected anti LFA-1 blocking mAb before cremaster muscle superfusion with CCL21. As shown in Fig. 3B, the number of firmly adherent GFP+ cells observed under these conditions did not increase over baseline levels (p > 0.05 as compared with control), indicating that LFA-1 is the main adhesion molecule in CCL21-induced T cell sticking. CCL21 elicits accumulation of naive T cells in dorsal skin air pouches
Although our intravital microscopy analysis showed that the presence of CCL21 in a nonlymphoid tissue is sufficient to induce intravascular sticking of GFP+ T cells, it remained to be determined whether this would lead to subsequent diapedesis and accumulation of naive T cell in the extravascular space. This question could not be answered by intravital microscopy alone, because the maximal observation period during which animals can be kept in anesthesia was too short to assess the magnitude of T cell emigration, which is a relatively slow and inefficient process. Moreover, intravital microscopy does not allow for phenotypic analysis of sticking and emigrated GFP+ cells.
To determine whether and to what extent naive T cells can leave the circulation in nonlymphoid tissues upon encounter of CCL21, we generated air pouches in the dorsal skin of T-GFP mice. Six days later, groups of six to eight animals were injected twice with either PBS alone or with PBS containing TNF- (500 ng), or chemokines (2 µg for CCL21, 1 or 2 µg for CCL19 and CXCL12) into air pouches. Eight hours later, air pouches were thoroughly lavaged, leukocytes in the lavage fluid were enumerated and their phenotype was analyzed by flow cytometry as described in Fig. 1 (Fig. 4).
|
(Fig. 4, A–D), even though TNF- caused a massive inflammatory response as evidenced by a 5-fold increase in the total leukocyte number recovered (p < 0.05 vs PBS; Fig. 4B). In contrast, injection of CCL21 did not change the overall number of recovered cells (which were mostly myeloid cells). However, CCL21 provoked the appearance of two distinct populations of naive and central memory T cells, which represented 4.8 ± 2.0% and 1.4 ± 0.6% of gated lymphocytes, respectively. On average, 751 ± 182 naive (R3) and 394 ± 114 central memory (R2) T cells migrated to an air pouch after injection of CCL21 (Fig. 4D). These numbers were significantly higher than those after injection of PBS alone or PBS plus TNF- (naive T cells: 28 ± 14 and 11 ± 11, respectively; central memory T cells: 53 ± 26 and 58 ± 46, respectively; p < 0.01; Fig. 4D). There was no significant difference in the frequency or total number of CD44highGFPlow (R1 in Fig. 4A) cells after PBS, TNF-, or CCL21 injection (Fig. 4C). We also investigated the accumulation of GFP+CD44low cells after the administration of CXCL12 and CCL19 into s.c. air pouches (n = 7; Fig. 4E). To adjust for differences in the molecular mass of these chemokines (CCL21: 16–17 kDa; CCL19: 9.4 kDa; CXCL12: 8 kDa, according to the manufacturer’s specifications), in five of the seven animals in each group, 1 µg of CCL19 or CXCL12 was injected, whereas recipients of CCL21 always received 2 µg. Two additional mice received 2 µg of CCL19 or CXCL12 twice. Because there was no difference in the response to these different doses of CCL19 and CXCL12, results are pooled in Fig. 4E. Somewhat surprisingly, neither CCL19 nor CXCL12 induced a statistically significant increase in the number of GFPhighCD44low cells in the air pouches (Fig. 4E; CXCL12: 40 ± 32; CCL19: 118 ± 51, p > 0.05 vs PBS). Taken together, the local generation of an appropriate chemokine, i.e., CCL21, appears to be required for naive T cell entry into tissues.
CCL21 is expressed in blood vessel endothelial cells in RA synovial tissue (RAST) and UC
Having determined that the administration of exogenous CCL21 induces naive T cell recruitment into normal mouse tissues, we asked whether the ectopic production of this chemokine might play a role in human disease. Tissue sections derived from RAST, UC, psoriasis and, as a control, normal skin and colon were immunostained for CCL21. Sections were also stained for vWF, which is highly expressed by endothelial cells (EC) in blood vessels, but only weakly expressed or absent in lymphatic vessels (29). Normal colon (Fig. 5A) and skin (not shown) contained abundant vWF+ blood vessels, but no staining for CCL21 was seen in these vessels. In contrast, vWF- lymphatic vessels showed a prominent cytoplasmic staining pattern for CCL21 consistent with previous reports that lymphatic EC in normal tissues express this chemokine (30). As a positive control for CCL21 expression in blood vessels we used human tonsils, in which CCL21 staining was consistently found on vWF+ vessels (Fig. 5B). These vessels were also PNAd+, and therefore represented HEV (not shown). Importantly, numerous PNAd- vWF+ vessels in both UC (Fig. 5, C and E) and RAST (Fig. 5D) expressed CCL21. No staining was seen with an isotype-matched control mAb (not shown). Some sections were also double-stained using anti-CCL21 Ab as well as mAb PAL-E, which specifically visualizes blood vessel, but not lymphatic, endothelium (31). Identical results were obtained as with anti-vWF Ab (data not shown).
|
In addition to endothelial cells, a strong positive signal for CCL21 mRNA and protein was also detected in many stromal cells in UC and, especially, RAST (Fig. 5). These cells were HLA-DR+, but most did not express the dendritic cell (DC) marker CD11c (not shown).
Naive T cells are present in RAST and UC
We next examined the relationship between CCL21 expression and the presence of naive T cells in these diseases. Anti-CD45RA in combination with anti-CD3 was used to detect naive T cells by immunostaining of frozen sections. Within ectopic lymphoid tissue in RAST and UC, densely packed CD3+ T cells were observed (data not shown). Strikingly, CD45RA+CD3+ cells were abundant in these tissues, even outside organized lymphoid follicles where they were frequently located in small clusters around a central blood vessel (Figs. 5, C–E, and 6, A and B). At these sites, CD45RA+ cells constituted 14.8 ± 6.1% and 6.0 ± 1.7% of the total CD3+ T cell population in RAST and UC, respectively. In contrast, naive T cells were rarely found in psoriasis (Figs. 6C and 7), where 0.5 ± 0.5% of CD3+ T cells expressed CD45RA (p < 0.05 vs RAST and UC). CCL21 expression in blood vessels correlated positively with the influx of naive T cells (Fig. 7C).
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Some common autoimmune diseases, including RA, multiple sclerosis, Hashimoto’s thyroiditis, diabetes mellitus, and chronic inflammatory bowel diseases, as well as some infectious diseases are associated with the accumulation of lymphocytes and DCs in target organs, resulting in the de novo formation of organized lymphoid tissue (reviewed in Ref. 35). Although it has been suggested that these LN-like structures causally contribute to the perpetuation of autoimmune diseases, relatively little is known about the molecular mechanisms that lead to their formation. In particular, homing pathways of lymphocytes into these tissues have yet to be characterized.
Naive T cells draining from chronically inflamed skin have been observed in sheep, and PNAd+ venules have been demonstrated at the site of inflammation (36). Moreover, expression of PNAd and/or MAdCAM-1 on HEV-like vessels is commonly found in lymphoid neogenesis (reviewed in Ref. 4). These vessels represent a likely port of entry for naive T cells in chronically inflamed tissues. However, naive T cell traffic via this route could only occur at relatively late stages, i.e., after HEV have formed. An alternative scenario would be that, at least in the early course of chronic inflammation, naive T cells gain access to the periphery via regular blood vessels that do not (yet) express HEV-specific vascular addressins. This hypothesis is supported by the fact that inflamed venules in many vascular beds up-regulate L-selectin ligands, which are distinct from PNAd and MAdCAM (5, 6). Moreover, transgenic approaches have shown that the continuous presence of CCL21 during embryonal development, and postnatally, leads to the recruitment of naive T cells, which are indispensable for ectopic lymphoid tissue formation (10, 11, 12), indicating that their recruitment may precede, and, in fact, induce HEV formation. Thus, although HEV are likely to facilitate lymphocyte influx, their presence may be a consequence of, rather than a prerequisite for, lymphoid neogenesis.
Although the observations discussed above highlight the role of T lymphocytes, including naive T cells, in chronic inflammatory diseases, the recruitment mechanisms that act on different T cell subsets under pathologic conditions are incompletely understood. Similarly, although it is widely accepted that naive T cells are excluded from nonlymphoid tissues under normal and acute inflammatory conditions, the molecular mechanisms behind this exclusion have yet to be examined. To set a baseline for the adhesion behavior of naive T cells in nonlymphoid organs, we used our well-established T-GFP model (14). T-GFP mice are especially suited to address this question, because >95% of GFPhigh cells in the peripheral blood of young animals are of a naive phenotype, i.e., they are CD44low, express high levels of CCR7 (Fig. 1), and are readily detectable by intravital microscopy. We used the cremaster muscle as a model because the endothelial traffic molecules involved in leukocyte accumulation in that tissue have been defined in detail by intravital microscopy (26, 27). Thus, the unavoidable surgical trauma associated with tissue preparation induces sequential expression of endothelial adhesion molecules. Initially, P-selectin is up-regulated, which mediates leukocyte rolling during the first hour. Thereafter, endothelial L-selectin ligands also contribute to rolling (26). We found that 35% of GFPhigh cells interact with venules in the cremaster muscle in a largely L-selectin-dependent manner. The two major known endothelial L-selectin ligands, PNAd and MAdCAM-1 (3, 28), are not expressed in acutely inflamed cremaster muscle venules, and the molecular nature of L-selectin ligands in the periphery remains to be identified.
T cell-endothelial cell interactions in nonlymphoid organs in vivo were studied previously by intravital microscopy using CD2-EGFP transgenic mice, in which enhanced GFP (EGFP) was expressed under the control of the CD2 promoter (37). In these mice both naive and effector/memory T cells express EGFP. A small subset of CD8+ T cells (<1% of circulating leukocytes) expressed high levels of EGFP, permitting in vivo detection. These cells rolled in an 4 integrin-dependent manner in TNF- and IFN-
-treated cremaster muscle venules. Because all naive T cells and central memory cells in T-GFP mice are GFP+, whereas CD8+ EGFPhigh cells in CD2-EGFP mice were probably composed of Ag-experienced effector/memory cells which might have down-regulated L-selectin expression, the present findings and the observations reported in CD2-EGFP mice do not contradict each other.
Interestingly, we found a partial reduction in the rolling fraction of GFP+ cells after injection of blocking Abs against P-selectin. The protein backbone of the principal P-selectin counterreceptor, P-selectin glycoprotein ligand (PSGL)-1, is expressed by all leukocytes, but requires extensive posttranslational glycosylation and sulfation to be functional (38). Highly functional (i.e., soluble P-selectin binding) lymphocyte-expressed PSGL-1 is primarily observed on effector Thl cells and CTL, but is normally absent on naive T cells in vitro (15, 39). However, the present in vivo findings are consistent with earlier observations that activated platelets bind up to 50% of human and murine peripheral blood T cells in a P-selectin-dependent manner (40), indicating that functional, but presumably low affinity and/or density, P-selectin ligands are expressed on many naive T cells. Alternatively, the modest effect of anti-P-selectin could have been indirect, e.g. by blocking the adhesion of GFP- leukocytes, such as neutrophils, which express functional L-selectin ligands on their surface. Thus, anti-P-selectin and anti-L-selectin might have additionally interfered with so-called secondary tethering, where L-selectin on naive T cells interacts with PSGL-1 on leukocytes that are already bound to the endothelium in a P-selectin-dependent fashion (41, 42).
Numerous in vivo and in vitro studies have shown that rolling leukocytes, including naive T cells, only arrest when integrins are functionally activated by a chemoattractant (reviewed in Ref. 2). Although naive T cells express 4 integrins and LFA-1 which mediate effective sticking in HEV (3, 28), they failed to arrest in the cremaster muscle and to migrate into s.c. air pouches, even under severe, TNF--induced inflammatory conditions. However, accumulation of GFP+ cells could be provoked by exogenous application of CCL21 in both the cremaster muscle and air pouch models. It should be cautioned that 5% of GFPhigh cells in peripheral blood of young mice display a memory cell phenotype (CD44high). These cells also express CCR7 (Fig. 1) and, hence, resemble central memory T cells (24). Thus, we cannot formally exclude that a disproportionally high fraction of the GFP+ cells that arrested in response to CCR7 ligands in cremaster muscle venules were Ag-experienced. However, phenotypic analysis of cells isolated from air pouches in response to CCL21 revealed the presence of GFPhighCD44low cells. This finding argues strongly that the mere presence of CCL21 is indeed capable of attracting naive T cells to the periphery. These results are also consistent with our previous demonstration that the intracutaneous injection of CCL21 or CCL19 can transiently restore the defect in T cell homing in plt/plt mice (21, 32). They further suggest that the lack of the appropriate chemokine in the periphery is a rate-limiting factor for naive T cell arrest, and that the failure to adhere can be overcome by the application or induction of endothelial presentation of CCL21.
The influx of naive T cells into s.c. air pouches in response to CCL21 is in apparent contrast to a recent report showing absence of lymphoid tissue formation in KCCL21 mice with transgenic expression of CCL21 in the epidermis under the control of the keratin-14 promoter (11). A possible explanation for this phenomenon could be that transgenic expression of CCL21 in KCCL21 mice by a large number of keratinocytes may result in high systemic levels of CCL21, which might have desensitized circulating naive T cells. Such a phenomenon has been observed in mice with transgenic expression of CCL2 (monocyte chemoattractant protein-1/JE) under the control of the mouse mammary tumor long-terminal repeat (43). Given these considerations, the phenotype of KCCL21 mice is not necessarily in conflict with our current observations. Whether the same adhesion mechanisms as observed in cremaster muscle venules apply for homing of naive T cells to air pouches remains to be determined.
The differential effects of CCL21 and CCL19 on naive T cell accumulation in the air pouch model were surprising, given that both chemokines induced firm adherence of GFP+ cells in cremaster muscle venules. Moreover, we have previously shown that s.c. injected CCL19 and CCL21 are similarly transported to and across LN HEV, where both are capable of reconstituting naive T cell recruitment in plt/plt mice (21, 32). One has to consider that the LN represents a highly specialized microenvironment, where lymph-borne chemokines reach HEV via the fibroblastic reticular cell conduit (44). Such a mechanism may not exist for skin microcirculation, which could conceivably result in differential transport of the two chemokines to the vascular lumen. Alternatively, it is also possible that both CCL19 and CCL21 triggered integrin activation, but only CCL21 may be able to provide additional signals necessary for the subsequent transmigration of naive T cells into the pouch cavity. The molecular mechanisms behind this striking difference remain to be identified.
In contrast to CCL21, CXCL12 was incapable of inducing firm adherence and accumulation of GFP+ cells in the cremaster muscle and the air pouch model, respectively. This failure of CXCL12 to induce recruitment of naive T cells was unexpected, because immobilized CXCL12 is as efficient at inducing integrin activation on rolling T cells in vitro as CCL21 (45). Moreover, addition of CXCL12 to the luminal surface of endothelial monolayers in flow chamber experiments resulted in retention of this chemokine on the endothelial surface and efficient T cell recruitment (46). However, our data are consistent with the recent observation that transgenic expression of CXCL12 in pancreatic islets, in contrast to CCL21 and CCL19, results in little accumulation of T cells (12). Thus, CXCL12 apparently has different effects on naive T cell recruitment in vitro and in vivo. However, we cannot exclude that the concentration of CXCL12 on the luminal surface of the endothelium under the conditions used were insufficient to mediate the arrest of fast rolling naive T cells (despite the fact that they induced sticking of Ag-experienced T cell populations; L. Scimone and U. H. von Andrian, unpublished observation).
Previous studies in mouse models have demonstrated that lymphoid neogenesis is associated with CCL21 expression on HEV-like vessels (13, 47). In addition, the expression of CCL21 protein on endothelial cells in RAST, chronic inflammatory liver disease, and certain autoimmune skin diseases has been reported recently (48, 49, 50). However, in these studies the exact nature of endothelial cells, i.e., whether they represented HEV, was not determined. The present work extends these findings by showing expression of CCL21 protein not only in PNAd+ HEV-like vessels, but also in PNAd-vWF+/PAL-E+ blood vessels with flat endothelium, in two autoimmune diseases, RA and UC. In addition, we demonstrate by in situ hybridization that blood vessel endothelium expresses CCL21 mRNA. This finding is of importance, as endothelial cells can pick up chemokines from surrounding tissues, which can be further transcytosed to the luminal surface (21, 32). We also show that CCL21 expression on blood vessels positively correlated with the presence of CD45RA+ T cells in the inflammatory infiltrate.
What might be the pathophysiologic consequences of ectopic CCL21 expression in the course of autoimmune diseases? Is the acquisition of CCL21 production by endothelial cells a mere epiphenomenon of the inflammatory process, or does it represent a critical step in the process of lymphoid neogenesis? The guidance of T cells and APC to T cell areas in secondary lymphoid organs is tightly controlled by constitutively expressed chemokines (2, 51). In particular, the entry of naive T cells from the bloodstream into PLN is mediated by HEV-displayed CCL21 (and probably CCL19). Stromal cells express high amounts of CCL21 in the T cell zone in murine PLN (52). From this it has been suggested that these stromal cells may establish migrational "corridors" where both CCR7+ naive T cells and CCR7+ DC are brought together by the high local concentration of CCL21, and so can establish physical interactions which are prerequisite for T cell activation (44, 52). Consequently, CCR7 and its ligands are key elements in the spatial and temporal regulation of naive T cell priming. Hence, CCL21 expression by blood vessels might indeed attract CCR7+ naive and central memory T cells to target organs in certain autoimmune diseases. Our demonstration of CCL21 mRNA and protein expression by stromal cells in RAST and UC suggests that in analogy to secondary lymphoid organs, T cells and APC might come into physical contact and, consequently, autoreactive naive T cells may be primed directly at the site of inflammation. Interestingly, activation of naive CD4+ T cells induces LT12 expression on their surface (12, 53, 54). LT12 is a major factor in the formation of organized lymphoid tissues (55, 56). Thus, a plausible scenario for the development of lymphoid neogenesis could be that it is initiated by de novo presentation of CCL21 on endothelial cells, which, as we show in this study, triggers the influx of naive T cells. When naive T cells become stimulated, they provide a rich source of LT12, which would further enhance the expression of CCL21 and other lymphocyte trafficking molecules, thus establishing a positive feedback loop.
Taken together, our data provide a number of strong arguments for the role of endothelium-expressed CCL21 in lymphoid neogenesis: 1) naive T cells can roll via L- and P-selectin, but are incapable of sticking on acutely inflamed endothelium in cremaster muscle; 2) the presence of CCL21 or CCL19, but not CXCL12, in cremaster muscle is sufficient to trigger integrin activation and sticking of naive T cells in this nonlymphoid site; 3) only the presence of CCL21, but not CCL19 or CXCL12, in air pouches leads to extravasation of naive T cells within 8 h, a period that is much shorter than the >6 days needed for the development of HEV in the pouch wall induced by chronic inflammatory stimuli (57); 4) lymphoid neogenesis in transgenic mice is inducible by chronic overexpression of CCL21 in the absence of other inflammatory stimuli, but is dependent upon (presumably CCL21-mediated) T cell recruitment (10); 5) in human RA, UC, and psoriasis, naive T cell accumulation correlates both quantitatively and spatially with the frequency and localization of CCL21+ microvessels; and 6) the presence of CD8+ T cells in rheumatoid synovitis in a human synovium-SCID chimera model was necessary for germinal center formation and maintenance (58). These findings suggest that CCL21 expression by blood vessels is an important early event in the formation of lymphoid tissue in tertiary organs, leading to the influx of naive T cells and possibly other CCR7+ lymphocytes.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Address correspondence and reprint requests to Drs. Wolfgang Weninger or Ulrich H. von Andrian, Center for Blood Research, Harvard Medical School, 200 Longwood Avenue, Boston, MA 02115. E-mail address: uva{at}cbr.med.harvard.edu or weninger{at}cbr.med.harvard.edu
3 Abbreviations used in this paper: PP, Peyer’s patch; LN, lymph node; PLN, peripheral LN; HEV, high endothelial venule; PNAd, peripheral node addressin; CCL, CC chemokine ligand; MAdCAM, mucosal addressin cell adhesion molecule; CXCL, CXC chemokine ligand; GFP, green fluorescent protein; vWF, von Willebrand factor; WSR, wall shear rate; RA, rheumatoid arthritis; UC, ulcerative colitis; DIG, digoxigenin; RAST, RA synovial tissue; EC, endothelial cell; PSGL, P-selectin glycoprotein ligand; EGFP, enhanced GFP; DC, dendritic cell.
Received for publication November 11, 2002. Accepted for publication February 24, 2003.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
47 and LFA-1 in lymphocyte homing to Peyer’s patch-HEV in situ: the multistep model confirmed and refined. Immunity 3:99.[Medline]
2 integrins in vivo. Proc. Natl. Acad. Sci. USA 88:7538.1 on synovial T cells contributes to their accumulation within the rheumatoid synovium. J. Immunol. 165:3423. to the tumor elicits an efficient immune response associated with induction of peripheral lymphoid-like tissue. Immunity 14:111.[Medline]
2 integrin CD11b/CD18 in neutrophil apoptosis: a homeostatic mechanism in inflammation. Immunity 5:653.[Medline]
(1, 3)-fucosyltransferase-IV and FucT-VII during leukocyte rolling in dermal microvessels. Immunity 12:665.[Medline]
. J. Immunol. 162:3790. is expressed on recently activated naive and Th1-like CD4 cells but is down-regulated by IL-4 during Th2 differentiation. J. Immunol. 162:1333. and revealed in lymphotoxin -deficient mice. Immunity 6:491.[Medline]
Related articles in The JI:
This article has been cited by other articles:
|
|
 |
|
  A. Kasorn, P. Alcaide, Y. Jia, K. K. Subramanian, B. Sarraj, Y. Li, F. Loison, H. Hattori, L. E. Silberstein, W. F. Luscinskas, et al. Focal Adhesion Kinase Regulates Pathogen-Killing Capability and Life Span of Neutrophils via Mediating Both Adhesion-Dependent and -Independent Cellular Signals J. Immunol., July 15, 2009; 183(2): 1032 - 1043. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. Sarraj, S. Massberg, Y. Li, A. Kasorn, K. Subramanian, F. Loison, L. E. Silberstein, U. von Andrian, and H. R. Luo Myeloid-Specific Deletion of Tumor Suppressor PTEN Augments Neutrophil Transendothelial Migration during Inflammation J. Immunol., June 1, 2009; 182(11): 7190 - 7200. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. D. Barish, A. R. Atkins, M. Downes, P. Olson, L.-W. Chong, M. Nelson, Y. Zou, H. Hwang, H. Kang, L. Curtiss, et al. PPAR{delta} regulates multiple proinflammatory pathways to suppress atherosclerosis PNAS, March 18, 2008; 105(11): 4271 - 4276. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Manzo, S. Bugatti, R. Caporali, R. Prevo, D. G. Jackson, M. Uguccioni, C. D. Buckley, C. Montecucco, and C. Pitzalis CCL21 Expression Pattern of Human Secondary Lymphoid Organ Stroma Is Conserved in Inflammatory Lesions with Lymphoid Neogenesis Am. J. Pathol., November 1, 2007; 171(5): 1549 - 1562. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Makita, T. Kanai, Y. Nemoto, T. Totsuka, R. Okamoto, K. Tsuchiya, M. Yamamoto, H. Kiyono, and M. Watanabe Intestinal Lamina Propria Retaining CD4+CD25+ Regulatory T Cells Is A Suppressive Site of Intestinal Inflammation J. Immunol., April 15, 2007; 178(8): 4937 - 4946. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G.-Z. Liu, L.-B. Fang, P. Hjelmstrom, and X.-G. Gao Increased CD8+ central memory T cells in patients with multiple sclerosis Multiple Sclerosis, March 1, 2007; 13(2): 149 - 155. [Abstract] [PDF]
|
|
|
|
 |
|
 J. K. Damas, C. Smith, E. Oie, B. Fevang, B. Halvorsen, T. Waehre, A. Boullier, U. Breland, A. Yndestad, O. Ovchinnikova, et al. Enhanced Expression of the Homeostatic Chemokines CCL19 and CCL21 in Clinical and Experimental Atherosclerosis: Possible Pathogenic Role in Plaque Destabilization Arterioscler Thromb Vasc Biol, March 1, 2007; 27(3): 614 - 620. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Gregorio, C. Gambini, V. Gerloni, A. Parafioriti, M. P. Sormani, S. Gregorio, G. De Marco, F. Rossi, A. Martini, and M. Gattorno Lymphoid neogenesis in juvenile idiopathic arthritis correlates with ANA positivity and plasma cells infiltration Rheumatology, February 1, 2007; 46(2): 308 - 313. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
N. Sakai, T. Wada, H. Yokoyama, M. Lipp, S. Ueha, K. Matsushima, and S. Kaneko Secondary lymphoid tissue chemokine (SLC/CCL21)/CCR7 signaling regulates fibrocytes in renal fibrosis PNAS, September 19, 2006; 103(38): 14098 - 14103. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Raine, P. Zaccone, P. Mastroeni, and A. Cooke Salmonella typhimurium Infection in Nonobese Diabetic Mice Generates Immunomodulatory Dendritic Cells Able to Prevent Type 1 Diabetes J. Immunol., August 15, 2006; 177(4): 2224 - 2233. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Marchal-Somme, Y. Uzunhan, S. Marchand-Adam, D. Valeyre, V. Soumelis, B. Crestani, and P. Soler Cutting Edge: Nonproliferating Mature Immune Cells Form a Novel Type of Organized Lymphoid Structure in Idiopathic Pulmonary Fibrosis J. Immunol., May 15, 2006; 176(10): 5735 - 5739. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. H. K. Lim, M. M. Burdick, S. A. Hudson, F. B. Mustafa, K. Konstantopoulos, and B. S. Bochner Stimulation of Human Endothelium with IL-3 Induces Selective Basophil Accumulation In Vitro J. Immunol., May 1, 2006; 176(9): 5346 - 5353. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P Middel, D Raddatz, B Gunawan, F Haller, and H-J Radzun Increased number of mature dendritic cells in Crohn's disease: evidence for a chemokine mediated retention mechanism Gut, February 1, 2006; 55(2): 220 - 227. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Kaiser, E. Donnadieu, J.-P. Abastado, A. Trautmann, and A. Nardin CC Chemokine Ligand 19 Secreted by Mature Dendritic Cells Increases Naive T Cell Scanning Behavior and Their Response to Rare Cognate Antigen J. Immunol., August 15, 2005; 175(4): 2349 - 2356. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. S. Carlsen, G. Haraldsen, P. Brandtzaeg, and E. S. Baekkevold Disparate lymphoid chemokine expression in mice and men: no evidence of CCL21 synthesis by human high endothelial venules Blood, July 15, 2005; 106(2): 444 - 446. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F.-E. Johansen, E. S. Baekkevold, H. S. Carlsen, I. N. Farstad, D. Soler, and P. Brandtzaeg Regional induction of adhesion molecules and chemokine receptors explains disparate homing of human B cells to systemic and mucosal effector sites: dispersion from tonsils Blood, July 15, 2005; 106(2): 593 - 600. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Burman, O. Haworth, D. L. Hardie, E. N. Amft, C. Siewert, D. G. Jackson, M. Salmon, and C. D. Buckley A Chemokine-Dependent Stromal Induction Mechanism for Aberrant Lymphocyte Accumulation and Compromised Lymphatic Return in Rheumatoid Arthritis J. Immunol., February 1, 2005; 174(3): 1693 - 1700. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. R. Mora, G. Cheng, D. Picarella, M. Briskin, N. Buchanan, and U. H. von Andrian Reciprocal and dynamic control of CD8 T cell homing by dendritic cells from skin- and gut-associated lymphoid tissues J. Exp. Med., January 18, 2005; 201(2): 303 - 316. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Cupedo and R. E. Mebius Cellular Interactions in Lymph Node Development J. Immunol., January 1, 2005; 174(1): 21 - 25. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. Liu, T. Ueno, S. Kuse, F. Saito, T. Nitta, L. Piali, H. Nakano, T. Kakiuchi, M. Lipp, G. A. Hollander, et al. The role of CCL21 in recruitment of T-precursor cells to fetal thymi Blood, January 1, 2005; 105(1): 31 - 39. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J.-M. Gauguet, S. D. Rosen, J. D. Marth, and U. H. von Andrian Core 2 branching {beta}1,6-N-acetylglucosaminyltransferase and high endothelial cell N-acetylglucosamine-6-sulfotransferase exert differential control over B- and T-lymphocyte homing to peripheral lymph nodes Blood, December 15, 2004; 104(13): 4104 - 4112. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. VALBUENA and D. H. WALKER EFFECT OF BLOCKING THE CXCL9/10-CXCR3 CHEMOKINE SYSTEM IN THE OUTCOME OF ENDOTHELIAL-TARGET RICKETTSIAL INFECTIONS Am J Trop Med Hyg, October 1, 2004; 71(4): 393 - 399. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. L. Scimone, T. W. Felbinger, I. B. Mazo, J. V. Stein, U. H. von Andrian, and W. Weninger CXCL12 Mediates CCR7-independent Homing of Central Memory Cells, But Not Naive T Cells, in Peripheral Lymph Nodes J. Exp. Med., April 19, 2004; 199(8): 1113 - 1120. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. Kerjaschki, H. M. Regele, I. Moosberger, K. Nagy-Bojarski, B. Watschinger, A. Soleiman, P. Birner, S. Krieger, A. Hovorka, G. Silberhumer, et al. Lymphatic Neoangiogenesis in Human Kidney Transplants Is Associated with Immunologically Active Lymphocytic Infiltrates J. Am. Soc. Nephrol., March 1, 2004; 15(3): 603 - 612. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Zippelius, G. Bioley, F.-A. Le Gal, N. Rufer, M. Brandes, P. Batard, M. De Smedt, J. Plum, D. E. Speiser, J.-C. Cerottini, et al. Human Thymus Exports Naive CD8 T Cells That Can Home to Nonlymphoid Tissues J. Immunol., March 1, 2004; 172(5): 2773 - 2777. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. N. Toka, M. Gierynska, and B. T. Rouse Codelivery of CCR7 Ligands as Molecular Adjuvants Enhances the Protective Immune Response against Herpes Simplex Virus Type 1 J. Virol., December 1, 2003; 77(23): 12742 - 12752. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
, 100 nM), CCL19 (
, 100 nM), CXCL12 (
, 100 nM), or buffer (•). B, Pretreatment with TNF-
).