Access of T effector cells to sites of inflammation is a prerequisite for an efficient action in immune defense and is mediated by different, partly tissue-specific sets of adhesion molecules. To what extent lymphocytes memorize the site of initial priming and develop organ-specific homing properties is still a matter of debate. Notably, data on the stability of homing receptor expression on T cells in vivo are largely lacking. We approached this question by the adoptive transfer of CD4+ T cells sorted for the expression of P-selectin ligands, which contribute to migration into inflamed sites in skin and other tissues. We observed long-term expression of P-selectin ligands on roughly one-third of effector cells. On those cells that had lost P-selectin ligands, re-expression upon Ag challenge was observed but only within pLNs, similar to the organ-selective induction upon the primary activation of naive T cells. The frequency of cells stably expressing P-selectin ligands was higher when cells were repeatedly stimulated under permissive conditions in the presence of IL-12, indicating a gradual fixation of this phenotype. In line with that finding, isolated P-selectin ligand positive memory T cells showed the highest frequency of long-term expressing cells. A tissue-specific environment was not required for the long-term maintenance of P-selectin ligand expression on the subfraction of effector cells. These data indicate that the expression of selectin ligands can become clonally imprinted under certain conditions, but also that a major fraction of the cells remains flexible and subject to environmental modulation upon restimulation.
Inflammation-seeking homing properties are induced on T cells upon differentiation, in parallel to the acquisition of effector functions and commitment to a Th1 or Th2 phenotype (1, 2, 3). Whereas naive T cells express L-selectin and CCR7, effector T cells up-regulate P-selectin ligands (P-ligs)3, E-selectin ligands (E-ligs), and chemokine receptors for inflammatory chemokines like CCR5 and CXCR3 (4, 5). In particular, E-ligs and P-ligs play a major role in the recruitment of T cells into inflamed regions, notably inflamed skin (6). Thus, a blockade of selectin-dependent adhesion abrogates the accumulation of effector cells within delayed-type hypersensitivity and contact-dependent hypersensitivity reactions (7, 8, 9). Although many authors see E-selectin associated with skin homing, P-selectin-dependent interactions have also been described for other sites of inflammation (10, 11, 12, 13, 14).
Selectin ligands are composed of a carrier protein decorated with carbohydrate structures that serve as a binding epitope (15). The main carrier protein of selectin ligands is P-selectin glycoprotein-1 (PSGL-1), which is constitutively expressed and, upon posttranslational modification, functions as both a P-lig and an E-lig. Critical for the generation of the carbohydrate epitopes are glycosyltransferases, namely fucosyltransferase VII (FucT-VII) and core-2β1,6-glucosaminyltransferase I (C2GlcNAT-I) (16). Neither of these enzymes is expressed in naive T cells but they become induced after Ag-specific activation, leading to the induction of functional P-ligs and to a lower extent E-ligs on the cell surface (17). A deficiency in either of these enzymes results in a loss of the P-selectin-binding capacity of lymphocytes (8, 18). Induction on CD4+ T cells requires TCR signals and costimulatory signals such as IL-12 (19). In vitro, IL-12 has been shown to strongly induce P-lig binding involving T-bet-dependent and STAT4-dependent signals (20, 21, 22). In humans, but not in mice (23), TGFβ was also described as up-regulating P-lig expression on T cells (24). In contrast to in vitro data, which suggested an association between the induction of the Th1 effector phenotype and selectin ligand expression in vivo, both IFN-γ- and IL-4-producing effector cells express P-ligs and E-ligs (25, 26). Further studies investigating the induction of selectin ligands in vivo showed that E-ligs and P-ligs are induced in a tissue-specific manner, i.e., within peripheral lymph nodes (pLNs). In contrast, α4β7 as a mucosal homing receptor is up-regulated when T cells are activated within mesenteric lymph nodes (mLNs) (27). Dendritic cells isolated from the respective tissue were shown to mediate the induction of skin or gut tropism (28, 29, 30). Further studies suggested a major effect of the local environment on the capability of dendritic cells to induce the respective homing receptors (31, 32). Although retinoic acid, a metabolite of vitamin A, was found to be instrumental for the induction of α4β7, the tissue-specific mediators of selectin ligand induction within pLNs are largely unknown (33). 1,25(OH)2D3, the active metabolite of vitamin D, induces skin-tropic CCR 10 expression but had no effect on selectin ligand expression (34). Our own recent data suggest that IL-12-dependent signaling is only partially involved in these environment-dependent induction pathways in vivo (23).
It might be reasoned that a rapid and efficient secondary immune response toward infection is greatly facilitated by the appropriate homing receptor expression on effector T cells. These were the teleological thoughts behind the old concept of organ-specific homing, which was based on numerous studies showing that memory lymphocytes display specialized patterns of homing receptors related to the site of initial Ag contact (35, 36, 37). However, recent studies on the plasticity of in vitro and in vivo primed skin- or gut-homing CD8+ effector cells showed a dynamic and reciprocal reprogramming of skin- and gut-tropic homing properties (38, 39). In fact, few studies have analyzed directly the clonal stability of homing receptor expression in vivo.
To address this issue we monitored P-lig expression over time on sorted P-lig+CD4+ effector and memory cells after transfer into recipient mice in the absence of a specific Ag. A fraction of 30–50% of transferred cells preserved the expression of P-lig. The maintenance of selectin ligand expression was independent of the organ environment, yet the rechallenge of adoptively transferred transgenic T cells by Ag led to reinduction only within cells residing in pLNs, similar to what has been observed for naive T cells.
Materials and Methods
DO11.10 mice (a gift from D. Y. Loh, Washington University School of Medicine, St. Louis, MO) carrying a transgenic TCR specific for the OVA323–339 peptide and BALB/c mice were bred under specific pathogen-free conditions in the Bundesinstitut für Risikobewertung, Berlin, Germany. All animal experiments were performed in accordance with institutional, state, and federal guidelines.
Isolation and cell purification
Naive CD62L+CD4+ T cells were purified from pooled spleens and peripheral and mLNs of DO11.10 mice by direct isolation of CD4+ cells using anti-CD4-FITC (clone GK1.5) mAb and anti-FITC multisort MACS beads (Miltenyi Biotec) to a purity of at least 98%. Naive CD4+CD62L+ cells were positively selected with anti-CD62L microbeads (Miltenyi Biotec) to a purity of at least 98%. APCs were prepared by the depletion of CD90+ cells from spleen cells or a purified CD4− fraction using anti-CD90 MACS microbeads (Miltenyi Biotec). APCs were irradiated (30 gray) before culture.
P-lig+ and P-lig− memory CD4+ T cells were isolated from the lymph nodes and spleens of ∼6-mo-old Do11.10 or BALB/c mice. Memory T cells were enriched by the depletion of naive cells using anti-CD62L MACS microbeads or CD45RB-FITC and anti-FITC microbeads followed by sorting with the autoMACS system (Miltenyi Biotec). In case of CD62L depletion, the negative fraction was stained with anti-CD4-FITC Ab (clone L3T4) and anti-FITC microbeads. After positive magnetic selection, CD4+ T cells were stained for CD45RB (CD45RbBio; streptavidin-allophycocyanin) and for P-lig expression (P-lig-IgG; anti-human IgG-PE) and sorted by FACSDiVa (BD Biosciences) for CD4+CD45RblowP-lig+ and CD4+CD45RblowP-lig− fractions. In case of CD45Rb depletion, CD4+ cells were enriched by the depletion of CD8 cells, B cells, and macrophages using rat-anti-mouse CD8 (clone 53-6.7), B220 (clone RA3.6B2), and MAC-1 (M1/70.15.11) Abs and goat anti-rat IgG MACS microbeads and autoMACS. The negative fraction was stained for CD45Rb-FITC, α4β7-PE, CD4-PerCP, and P-lig-Cy5 and sorted using FACSAria (BD Biosciences) for CD45Rblowα4β7+P-lig− and CD45Rblowα4β7−P-lig+ fractions.
Cell culture was set up with 2 × 106 cells/ml in complete RPMI 1640 containing 10% FCS and 10 μM 2-ME (Invitrogen Life Technologies). For the generation of P-lig+ T cells, CD4+CD62L+ T cells from OVA TCR transgenic (OVAtg) mice and APCs were cultured at a ratio of 1:4 in the presence of the OVA323–339 peptide (Biochemistry Department, Charité Universitätsmedizin, Berlin, Germany) at 0.5 μM. Cell cultures were supplemented with recombinant murine IL-12 (R&D Systems) at 5 ng/ml, IFN-γ (R&D Systems) at 20 ng/ml, and neutralizing anti-IL-4 Ab (clone 11B11; Deutsches Rheumaforschungszentrum, Berlin, Germany) at 5.5 μg/ml.
For reactivation of ex vivo isolated memory T cells, sorted cells were activated on plate-bound anti-CD3/anti-CD28 in the presence or absence of 1 nM retinoic acid and in the presence of rIL-2 (10 ng/ml). After 24 h the cells were removed from stimulus and cultured for the indicated time.
For the labeling of cultured CD4+ cells with CFSE (Molecular Probes), washed cells were resuspended at 1 × 107/ml in PBS, CFSE (5 μM final concentration) was added, and the cells were incubated for 3 min at room temperature. The reaction was stopped by washing with RPMI 1640 containing 10% FCS.
For cytometric analysis dead cells were excluded by staining with 4,6 diamidino-2-phenylindole (Sigma-Aldrich). Anti-CD4 Abs, anti-CD8 Abs, and PE–conjugated streptavidin were purchased from BD Biosciences. OVAtg T cells were identified by the clonotype-specific mAb KJ1–26.1. P-selectin binding ligands were detected by a P-selectin-human IgG chimeric protein (provided by M. Wild and D. Vestweber, Max-Planck-Institut für Vaskuläre Biologie, Münster, Germany) and PE–conjugated anti-human IgG Ab F(ab′)2 (Dianova) as a secondary reagent as previously described (2). Cytometric analysis was performed using a LSR or FACSCalibur flow cytometer and CellQuest research software (BD Biosciences).
Adoptive transfer of in vitro generated P-lig+ T cells
To generate P-lig+ T effector cells, CD4+ CD62L+ T cells from OVAtg mice were activated under conditions supporting high P-lig induction, i.e., under Th1 conditions as described (2). After 5–6 days, the cells were harvested and stained by P-selectin-human IgG and anti-human IgG-PE. P-lig+ T effector cells were positively selected with anti-PE microbeads (Miltenyi Biotec) to a purity of at least 98%. To generate repetitively induced P-lig+ cells, naive transgenic T cells were activated under P-lig-inductive conditions, sorted for P-lig+ cells on day 5, and reactivated under P-lig-inductive conditions by the OVA peptide, APCs, and a cytokine supplement. After 5–6 days P-lig+ T cells were isolated as described, i.e., by staining with a P-selectin-IgG chimera, a secondary Ab, and MACS microbeads. Single or repetitively induced P-lig+ cells were labeled with CFSE and 1 × 107 cells resuspended in 200 μl of PBS were transferred into BALB/c recipients by i.v. injection.
Adoptive transfer of P-lig+ and P-lig− ex vivo isolated memory T cells
Memory cells, positive or negative for P-lig, were isolated as described above. Before transfer, P-lig+ and P-lig− T cells were stained with CFSE. P-lig+ cells (0.8 × 106 per mouse) were transferred into three BALB/c mice and P-lig− cells (1.4 × 106 per mouse) were transferred into three BALB/c mice.
In vivo challenge
Transferred OVA-specific cells were stimulated by i.p. injection of 500 μg of OVA protein (Sigma-Aldrich) in the presence of 80 μg of LPS (Escherichia coli serotype 055:B5; Sigma-Aldrich) in a total volume of 200 μl of PBS. Control animals received LPS alone.
FTY720 treatment of recipient BALB/c mice was started 2 days after the adoptive transfer of T cells from DO.11.10 mice. FTY720 at 100 μg/kg (provided by V. Brinkman, Novartis Pharma) was injected every other day i.p. in 150 μl of H2O. Control animals received H2O alone. The effect of FTY720 treatment was controlled by the staining of peripheral blood for CD4+ and CD8+ cells.
Data are presented as mean ± SD. Significance was determined by Mann-Whitney U test. Differences were considered statistically significant with p ≤ 0.05.
Partial loss of P-lig expression on CD4+ effector cells in vitro
To test the stability of P-lig expression on CD4+ T cells under defined conditions, P-lig was induced by the activation of naive OVAtg T cells under P-lig-permissive conditions in vitro. CD4+ T cells from DO11.10 mice were stimulated with the OVA323–339 peptide in the presence of APCs and IL-12, IFN-γ, and anti-IL-4. P-lig expression on transgenic T cells identified by the clonotype-specific Ab KJ1.26–1 was monitored over time. As shown in Fig. 1 A, starting from day 2 after activation P-lig expression rapidly rose and reached peak levels around day 4. P-lig expression remained high up to day 7 and than decreased to levels of ∼50%, similar to what has been reported earlier for CD8+ cells (40).
Long-term expression of P-lig on a subset of CD4+ T effector cells enlarged by repetitive induction
To clarify whether P-lig expression persists on a subset of CD4+ effector cells for extended observation periods, we moved to an in vivo adoptive transfer system. Naive OVAtg T cells were activated in vitro as shown in Fig. 1,A. Five days after culture, P-lig+ T cells were sorted to a purity of 98% (Fig. 1,B), stained with the intracellular dye CFSE, and transferred into BALB/c recipients. After 2, 4, 8, 15, 30, and 60 days we determined the frequency of P-lig-expressing OVAtg T cells within peripheral and mLNs and spleens. OVA-transgenic T cells were identified by KJ1–26.1 and CFSE stain as shown in Fig. 1,C. Among transgenic T cells we determined the frequency of P-lig-expressing cells. In correspondence to the in vitro data, we found a rapid decline of the frequency of P-lig+ T cells over the first 15 days after transfer (Fig. 1, C and D). However, from day 30 on a plateau was reached and a subset of ∼35% of transgenic T cells remained positive for P-lig over a period of at least 60 days. This applied to all tissues analyzed and suggests that a significant fraction of cells was committed to the long-term expression of P-ligs. Proliferation of P-lig+ and P-lig− cells as indicated by CFSE loss was not significantly different, excluding a bias due to the selective outgrowth of one of the populations (Fig. 1 C and data not shown).
To test whether repetitive induction reinforces the commitment to express P-ligs, naive OVAtg T cells were activated as described in Fig. 1,A. Five days after culture P-lig+ T cells were sorted and reactivated under similar conditions by the OVA323–339 peptide, APCs, and IL-12. After 5–6 days of secondary stimulation, P-lig+ T effector cells were again enriched to 98% P-lig+ cells and transferred into BALB/c recipients. Fifteen and 30 days after transfer, P-lig expression was determined within lymph nodes and spleens. Fig. 1 D shows that P-lig expression on cells after secondary stimulation was significantly enhanced compared with cells transferred after primary stimulation.
Long-term expression of P-lig on memory CD4+ T cells in vivo
The previous data showed long-term P-lig expression for a subpopulation of T effector cells following a rapid decline in the total fraction of P-lig+ T cells after the initial stimulation. To analyze whether in vivo generated effector/memory T cells behave in a similar way, we isolated P-lig+ and P-lig− CD4+CD45Rblow T cells. These cells were enriched by MACS and FACS to 96% P-lig+ and 99% P-lig− memory T cells (Fig. 2 A). Before transfer, cells were labeled with CFSE. P-lig+ (0.8 × 106) and P-lig− (1.4 × 106) cells were transferred into individual recipients. After 15 days, a time point when 50–60% of the in vitro generated Th1 cells had lost P-lig expression, recipients were killed and their spleens and pLNs were collected. The complete cell suspension obtained from each organ was analyzed by FACS with a life gate on CD4+ T cells. Transferred memory/effector cells were identified by the CFSE label. Recovery was ∼1000 CFSE+ cells from spleen and 150–500 CFSE+ cells from the pLNs.
The majority of P-lig+ memory T cells remained positive for P-lig+ during 15 days; in spleens 90% and in lymph nodes 86% of the transferred T cells still expressed high levels of P-lig (Fig. 2, B and C). P-lig− T cells remained negative for P-lig-expression (Fig. 2, B and C). This indicates that the pool of P-lig+ memory cells generated in vivo consists to a large extent of cells with stable homing properties.
Reinduction in vivo depends on tissue-specific signals
Upon primary activation of naive T cells in vivo, preferential induction of P-lig is found predominantly within pLNs while mLNs do not efficiently support P-lig induction, suggesting that microenvironment-dependent factors regulate P-lig expression (25, 27). To determine whether reinduction on previously P-lig-positive effector cells follows the same pattern as that found for naive T cells, activated OVAtg P-lig+ T cells were transferred into recipients as in Fig. 1,B. On day 15 after transfer, recipient mice were challenged i.p. with LPS or LPS plus OVA or left untreated. Two days after immunization, OVAtg T cells within pLNs, mLNs, and spleens were analyzed for P-lig expression. In line with own previous experiments and other studies (41) in vitro generated effector cells had undergone about one further round of division in untreated and LPS-treated mice in the absence of Ag (Fig. 3,A). In contrast, treatment with LPS plus OVA resulted in vigorous proliferation within spleens and mLNs and pLNs as indicated by the loss of CFSE. However, only within pLNs was a significant reinduction of P-lig-expression observed (Fig. 3,B), suggesting that reinduction in vivo requires similar tissue-dependent inductive factors as primary induction on naive T cells does (Fig. 3 C). However, reinduction on effector cells appears to be more efficient than primary induction on naive cells, because 68.5% of the T effector cells that are P-lig− at the time point of challenge can be calculated to up-regulate P-lig expression compared with 29.7% of naive T cells after activation. In contrast, the activation of P-lig+ cells under conditions not supporting P-lig induction did not result in a loss of P-lig expression, suggesting that expression is resistant to environmental conditions favoring the expression of other (gut-specific) homing receptors.
Retinoic acid does not down-regulate P-lig expression on memory T cells but induces α4β7 expression on P-lig+ memory T cells
To analyze whether in vivo induced P-lig+ memory T cells can be modulated by activation in the presence of retinoic acid, which induces gut-homing molecules like α4β7, we enriched P-lig+α4β7− memory CD4+ T cells (CD45Rblow) and P-lig−α4β7+ memory CD4+ T cells (Fig. 4,A). The cells were activated on plate-bound anti-CD3/anti-CD28 for 24 h in the presence or absence of retinoic acid and then rested for 48 h. The presence of retinoic acid had no major effect on the frequency of P-lig+ T cells among enriched P-lig+ T cells (Fig. 4,B). The presence of retinoic acid, however, induced α4β7 expression on a significant fraction of P-lig+ cells (Fig. 4 B). This suggests that retinoic acid does not affect P-lig expression but can induce gut-homing properties on P-lig+ memory T cells without suppressing their capacity to migrate into inflamed sites.
Maintenance of P-lig expression in vivo does not depend on tissue-specific signals
It could be assumed that those T cells maintaining P-lig expression require repeated encounters with the inductive environment of pLNs, which could occur upon recirculation. To test this possibility, we “locked” lymphocytes in their respective environments using FTY720, which inhibits the exit of T cells from lymph nodes and, hence, recirculation. P-lig+ OVAtg T cells generated and sorted as described in Fig. 2 were transferred into recipient mice. Beginning at day 2 after transfer, recipient mice were treated i.p. with 1 mg/kg FTY720 every other day or, as control, with H2O. Treatment with FTY720 resulted in the disappearance of CD4+ T cells from the circulation as reported (Fig. 5,A and Refs. 42, 43, 44). Twenty-one days after T cell transfer, P-lig expression on transgenic T cells recovered from pLNs and mLNs was compared. A blockade of recirculation did not significantly reduce P-lig expression on cells sequestered within mLNs as compared with pLNs (Fig. 5 B). Thus, recirculation through pLNs is not required for the maintenance of P-lig expression on CD4+ T cells.
Effector/memory T cells, once polarized under specific conditions, exhibit a rather stable phenotype in vivo with respect to their commitment to produce distinct effector cytokines (45, 46). In this way, memory is established upon primary antigenic contact that imparts long-term immunity not only by expansion of the Ag-specific cells but also by focusing the response toward distinct reaction patterns.
That lymphocytes can memorize the site of initial priming and develop organ-specific homing by means of specific sets of adhesion molecules has been recognized decades ago and provides apparently another example of long-term imprinting of functional properties. However, recent data have provided evidence that the expression of homing receptors might display a higher degree of plasticity than previously thought (38, 39). As a matter of fact, no studies are available that investigate the clonal stability of homing receptor expression in vivo and address alternative hypotheses, e.g., that a continual restimulation by tissue-localized Ag is required for specialized homing receptor expression on memory cells.
In this work we studied the stability of P-lig expression in vivo in the absence of the cognate Ag. P-ligs act as inflammation-specific homing receptors on CD4+ effector cells together with E-ligs and are of major importance, especially in homing into the inflamed skin, but are also involved in other organ systems (9, 12, 13). Previously, we reported that P-ligs and E-ligs become only transiently expressed on effector/memory T cells upon the induction of skin inflammation and that P-lig+ cells declined after several weeks to basal levels (12); it remained unclear, however, whether the small number of surviving Ag-specific memory cells permanently express selectin ligands. Strong evidence for a long-term imprinting came from studies in which preferential expression of the E-selectin binding epitope cutaneous lymphocyte-associated Ag (CLA) was found on Ag-specific memory cells from patients with allergic dermatitis (47); however, a role of Ag stored locally for extended periods cannot be formally excluded in these studies.
Using an adoptive transfer system we here analyzed the stability of P-lig on preformed T effector cells in vivo in the absence of Ag for periods up to 60 days. Our data show that a fraction of effector/memory cells displays long-term expression of selectin ligands. This fraction apparently is permanently committed to inflammation- and skin-specific migration, whereas the other cells lose P-lig expression over a time period of a few weeks. Cells losing P-lig expression and those retaining expression do not differ in their expression of CD62L (data not shown), which might suggest that negative cells do not merely convert to a central memory type of cell.
Our previous data suggested that epigenetic mechanisms known to stabilize the cytokine memory in CD4+ effector cells might also be involved in maintaining homing receptor expression. In that study, cell division, which provides an open window for chromatin modification, appeared to be a prerequisite for the induction of selectin ligands in naive T cells and artificial demethylation by 5-aza-2-deoxycytidine enhanced the induction of the selectin ligands, suggesting that DNA methylation regulates the expression of selectin ligands (48). Although direct proof is still lacking, epigenetic imprinting is an attractive hypothesis for explaining a persistent commitment to P-lig expression. We also show in the present study that the fraction of cells stably expressing P-lig was enhanced by repetitive in vitro induction before transfer, suggesting that imprinting of the homing phenotype requires a number of instructive inductions similar to what has been found for specific cytokine effector phenotypes (49). In addition, we found that in vivo generated P-lig+ memory T cells show an even stronger fixation of P-lig expression than the repetitively in vitro induced T cells.
Lineage decisions enforced by chromatin remodeling can result either in continuous transcription of genes, as seen for CD4 and CD8 expression in T cells, or can enable rapid recall expression as is found for effector cytokines (50, 51). Data from this study suggest that P-ligs are stably expressed on a significant subset of CD4+ effector cells. We previously showed that the induction of P-lig in vitro on memory T cells is facilitated and independent of cell division, in contrast to naive cells, suggesting the presence of “unlocked” genes in committed memory cells (48). Whether reinduction has also reduced requirements with respect to cosignals remained unclear. The present data indicate that the reinduction of P-lig in vivo on effector cells, having returned into a P-lig negative state, still requires tissue-specific signals that are provided only by the environment of pLNs. In quantitative terms, reinduction within effector/memory cells turned out to be more efficient compared with naive T cells.
Mora et al. (39) described the reprogramming of gut- or skin-specific CD8+ T cells by tissue-specific dendritic cells, suggesting that a majority of cells with polarized homing receptor expression still can be modulated by appropriate environmental signals (although this was less pronounced for selectin ligand as compared with α4β7 expression). Indeed, we also found that the presence of retinoic acid during the activation of P-lig+ CD4+ memory T cells results in the induction of α4β7 in T cells in addition to P-lig. In vivo under homeostatic conditions double positive cells are rarely found but can be detected under conditions of inflammation, namely intestinal inflammation (our unpublished observations). However, we did not observe that a restimulation of P-lig+ cells in mucosal sites (mLNs) led to a suppression of P-lig expression, and those cells that retained P-lig during in vivo passage also retained it upon restimulation in the presence of retinoic acid. These data allowed us to conclude that P-lig expression on CD4+ T cells, once imprinted, is rather stable and resistant to modulation by environmental signals.
To address the question of whether the recirculation of effector cells and exposure to the inductive environment-specific factors in pLNs contribute to a persistent P-lig expression, we blocked the recirculation of T cells using FTY720. This treatment leads to a trapping of T cells within lymph nodes and, hence, strongly reduced recirculation. However, we did not observe a loss of P-lig expression on transferred T cells isolated from mLNs as compared with pLNs under these conditions. This suggests that the maintenance of P-lig expression, in contrast to reinduction, is independent of the inductive signals provided by pLNs.
The expression of E-selectin ligands obeys in many aspects similar rules as those of P-ligs (52), but the frequencies of E-lig+ cells, especially on in vitro generated effector cells, are usually much lower (2, 12, 25). This precluded the consideration of E-ligs in the present study, where high numbers of positive cells had to be purified and the methods for costaining both ligands simultaneously on transferred cells were lacking. However, our expectation is that similar features will apply to E-lig expression because the underlining mechanisms of synthesis are overlapping (52, 53).
In conclusion, we here provide direct evidence that the development of effector/memory T cells under permissive conditions can imprint a phenotype encompassing the persistent expression of P-ligs on a fraction of cells, depending on the developmental conditions. Stable expression is maintained in the absence of antigenic activation and does not require re-exposure to tissue-specific inductive factors. However, a fraction of cells remains adaptive and can react to new stimulatory signals. This indicates that both imprinted and flexible phenotypes coexist among effector/memory cells and that the immune system provides specialization while retaining its plasticity.
We thank Oliver Pabst for advice in using FTY720, Volker Brinkmann at Novartis Pharma AG for supply of FTY720, and Dietmar Vestweber and Martin Wild for supplying the P-selectin-IgG chimera.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by Deutsche Forschungsgemeinschaft Grants SFB 366 C7 and SY31/2-1.
Abbreviations used in this paper: P-lig, P-selectin ligand; E-lig, E-selectin ligand; mLN, mesenteric lymph node; OVAtg, OVA TCR transgenic; pLN, peripheral lymph node.