Functionally naive CD8 T cells in peripheral blood from adult humans can be fully described by their CD45RAbrightCCR7+CD62L+ cell surface phenotype. Cord blood lymphocytes, from healthy newborns, are homogenously functionally naive. Accordingly, the majority of cord blood CD8 T cells express the same pattern of cell surface molecules. Unexpectedly, however, a significant fraction of cord blood CD8 T cells express neither CCR7 nor CD62L. Yet these cells remain functionally naive as they contain high levels of TCR excision circles, have long telomeres, display highly polyclonal TCRs, and do not exhibit immediate effector functions. In addition, these CD8 T cells already represent a significant fraction of the mature naive CD8 single-positive thymocyte repertoire and may selectively express the cutaneous lymphocyte Ag. We suggest that CD8 single-positive thymocytes comprise two pools of naive precursors that exhibit distinct homing properties. Once seeded in the periphery, naive CCR7+CD62L+ CD8 T cells patrol secondary lymphoid organs, whereas naive CCR7−CD62L− CD8 T cells selectively migrate to peripheral tissues such as skin.
CD8 T cells provide a complex means of defense against pathogens and cancer. Quiescent, naive CD8 T cells constantly travel from the blood to secondary lymphoid organs, where they can encounter mature dendritic cells and specifically recognize MHC class I/peptide complexes at their cell surface. The mechanism of naive T cell extravasation to gain access to the peripheral lymph nodes is known in detail. It involves the sequential interaction of three ligand receptor pairs, namely CD62L/LPNAd, CCR7/CCL21, and LFA-1/ICAM-1, at the high endothelial venules (1, 2). Typically, naive T cells from healthy adult blood donors express high levels of cell surface CCR7 and CD62L. Upon appropriate antigenic stimulation in the lymph node T cell zone, CD8 T cells get activated and undergo a program of proliferation and differentiation. Their progeny includes 1) effector cells that turn off lymphoid-specific addressins (CD62L− CCR7−) while turning on tissue-specific addressins, which allow them to enter peripheral tissues to display immediate effector function and control invading pathogens; and 2) memory cells that keep their capacity to home to lymphoid organs (CD62L+ CCR7+) and can generate a new wave of effector cells upon secondary challenge (3, 4, 5, 6).
In studies of human T cells, cord blood lymphocytes are frequently used as source of naive T cells (7, 8). However, we found that a subset of CD8 T lymphocytes present in the cord blood do not express CCR7 and CD62L and therefore cannot migrate to secondary lymphoid tissues. In this study we characterized the phenotypic and functional makeup of this subset of cord blood CD8 T lymphocytes. We provide extended evidence that these cells are indeed naive. Moreover, they can express the cutaneous lymphocyte Ag (CLA),5 suggesting that they may reach peripheral nonlymphoid tissues such as skin. Roles for the trafficking of naive T cells through newborn peripheral tissues are discussed.
Materials and Methods
Tissues and cells
Cord blood from 30 newborns and thymus tissue from eight children, 0–12 years of age (mean ± SD, 39 ± 58 mo), who had undergone corrective cardiac surgery were obtained at the University Hospital of Ghent, Belgium, following the guidelines of the local medical ethical commission. Lymphocytes were collected from 13 blood donors at the Blood Transfusion Center (Lausanne, Switzerland). Cells were prepared as previously described (9).
Monoclonal Abs were obtained from BD Biosciences (Mountain View, CA), except for anti-CD45RO-FITC (DAKO, Copenhagen, Denmark), anti-CD28-FITC (Immunotech, Marseilles, France), anti-perforin-FITC (Ancell, Bayport, MN), anti-granzyme B-FITC (Hoelzel, Diagnostika, Koln, Germany), goat anti-rat IgG-FITC (Southern Biotechnology Associates, Birmingham, AL), anti-CD4-ECD and anti-CD8-ECD (Coulter, Miami, FL), and goat anti-rat IgG-PE or -allophycocyanin (Caltag, Burlingame, CA). Anti-CCR7 rat IgG mAb 3D12 was provided by Dr. M. Lipp (Max Delbrück Institute, Berlin, Germany).
CD8 T cells were enriched by MiniMACS (Miltenyi Biotec, Bergisch Gladbach, Germany) and incubated with anti-CCR7 mAb for 20 min at 4°C, with goat anti-rat mAbs for 20 min at 4°C, and then with appropriate mAbs. Cells were either immediately analyzed or sorted into defined populations on a FACSCalibur or FACSVantage SE, respectively, using CellQuest software (all from BD Biosciences). Immediate reanalysis of the sorted populations revealed >98% purity. For simultaneous analysis of CCR7 and CD27 or CD28 expression on mature CD8 single-positive (SP) thymocytes, CD4−-depleted cells were labeled with anti-CD1a-FITC, CD3-PE, CD4-ECD, CD8APC-Cy7, and CCR7-allophycocyanin mAbs. CD1a− thymocytes were sorted on a FACSVantage SE, labeled with anti-CD27-FITC or CD28-FITC, and analyzed again by flow cytometry. CD1a−CD3+CD4−CD8+ mature CD8 SP thymocytes were gated using CellQuest software.
Quantitative PCR of TCR excision circles (TRECs)
Quantification of signal joint TRECs was performed by real-time quantitative PCR with the 5′ nuclease (TaqMan) assay and an ABI7700 system (PerkinElmer, Foster City, CA) (9). The internal standard was provided by Dr. D. Douek (Human Immunology Section, Vaccine Research Center, National Institutes of Health, Bethesda, MD).
Telomere fluorescence in situ hybridization
The average length of telomere repeats was measured by quantitative fluorescence in situ hybridization and flow cytometry (10, 11). FITC-labeled fluorescent calibration beads (Quantum TM-24 Premixed; Bangs Laboratories, Fishers, IN) were used to convert telomere fluorescence data to molecules of equivalent soluble fluorescence units.
Complementary-determining region 3 (CDR3) size of TCR β-variable gene segment (TCR BV) transcripts
The CDR3 of the PCR-amplified TCR BV1–24 transcripts was analyzed using a run-off procedure (12, 13). The run-off products were run on an automated sequencer in the presence of fluorescent size markers. The lengths of DNA fragments and the fluorescence intensity of the bands were analyzed with Base ImagIR software (LI-COR Biotechnology Division, Bad Homburg, Germany).
IFN-γ cytospot assay
CD8 T cells were incubated for 4 h with 1 μg/ml PMA/0.25 μg/ml ionomycin. After 1 h, 10 μg/ml brefeldin A (Sigma-Aldrich, St. Louis, MO) was added. After 3 additional h, cells were stained with mAbs, fixed, permeabilized, and incubated with anti-IFN-γ-FITC in PBS/0.1% saponin for 20 min at 4°C.
Sorted CD8 T cell subsets were tested in anti-CD3 mAb-redirected 51Cr release assays at various lymphocyte/target cell ratios. FcγR-bearing P815 target cells were labeled with Na51CrO4 and used in the presence or the absence of anti-CD3 mAb (OKT3, 300 ng/ml). The percentage of target cell lysis was calculated as previously described (9).
CCR7− CD8 T cells in newborns
CD8 T cells (defined as CD3+CD8+) were retrieved from cord blood of healthy newborns and analyzed ex vivo by flow cytometry for the expression of CD45RA and CCR7. In line with previous observations (14, 15), the vast majority of CD8 T cells expressed a naive CD45RAhigh phenotype (Fig. 1,a). An intriguing finding, however, was the presence of a significant proportion of CCR7−CD8 T in all individuals analyzed (mean ± SD, 7 ± 4%), a phenotype currently attributed to T cells with effector function (16). The CCR7− cord blood CD8 T cells belonged to the TCRαβ T cell lineage (TCRαβ+, 92 ± 1%) and expressed CD8αβ receptors (CD8α+, 99 ± 1%; CD8β+, 99 ± 1%; data not shown). PBMCs from healthy adults were analyzed in parallel. As previously described (16), CD8 T cells were composed of four distinct subsets, including so-called naive (CCR7+CD45RAhigh, 37 ± 18%), central memory (CCR7+CD45RAlow, 9 ± 7%), effector memory (CCR7−CD45RAlow, 33 ± 11%), and terminally differentiated effector (CCR7−CD45RAhigh, 22 ± 9%) cells (Fig. 1 a).
Post-thymic replicative history
To directly define the contribution of thymic production to the pool of CCR7− CD8 T cells, we measured their post-thymic replicative history by quantification of signal joint TRECs (9). TREC levels were comparable in cord blood CCR7− and CCR7+CD8 T cells, indicating that the average number of cell divisions was similar in the two subsets (CCR7−, 24 ± 5%; CCR7+, 25 ± 3%; Fig. 1,b). Compared with the TREC levels of CD8 SP thymocytes (57 ± 7%; data not shown), it was estimated that CCR7− and CCR7+ cord blood CD8 T cells underwent, on the average, one population doubling. In marked contrast, TREC levels of CCR7− CD8 T cells in PBMCs from adults were 1 order of magnitude lower than those of autologous naive CD8 T cells (CCR7−, <1%; CCR7+, 8 ± 3%; Fig. 1,b), compatible with a significantly higher number of cell divisions. The reduced TREC content of naive CD8 T cells in adults compared with the ones in newborns probably reflects slow division rates driven by homeostatic mechanisms (17). As an independent approach to assess replicative history, the average length of telomere repeats of the populations described above was measured by fluorescence in situ hybridization and flow cytometry (fluorescence in situ hybridization) (10, 11). In agreement with the TREC data, telomere length was similar in CCR7− and CCR7+ cord blood CD8 T cells (13 ± 2 and 13 ± 2 kb, respectively), whereas CCR7− CD8 T cells in PBMCs from adults displayed significantly reduced telomere fluorescence corresponding to a telomere loss of ∼2–3 kb compared with autologous naive CD8 T cells (6 ± 1 and 9 ± 1 kb, respectively; Fig. 1 c). Both the lack of TREC dilution and the telomere shortening observed in cord blood CCR7− CD8 T cells allow exclusion of detectable peripheral expansion and strongly indicate that they are naive.
TCR BV repertoire
The CDR3 domain size distribution was assessed in amplified TCR BV gene segment transcripts from sorted CD8 T cell subsets. Cord blood CCR7− and CCR7+ fractions displayed a bell-shaped CDR3 size profile for all BV gene segments similar to the pattern observed for naive CD8 T cells in adults (Fig. 1 d). In contrast, the CDR3 size profiles for the majority of BV subfamilies from CCR7− CD8 T cells in adults displayed prominent peaks, indicating the accumulation of recurrent size transcripts as a result of clonal expansion. Thus, CCR7− CD8 T cells in cord blood did not express a canonical TCR, but comprised a large and diverse TCR repertoire, which is compatible with their naive state.
Function of CCR7− CD8 T cells
Effector CD8 T cells act by mediating apoptosis of targeted cells and/or generating local inflammation through cytokine release. However, CCR7+ and CCR7− cord blood CD8 T cells neither contained intracellular perforin and granzyme B (Fig. 2,a) nor expressed Fas ligand (data not shown) and hence lacked major effector molecules involved in cytolytic activity. Accordingly, both fractions were not cytolytic in anti-CD3 mAb-redirected assays (Fig. 2,b). Furthermore, neither CCR7+ nor CCR7− cord blood CD8 T cells produced IFN-γ upon mitogenic stimulation (Fig. 2,a). In clear contrast, circulating CCR7− CD8 T cells in adults contained intracellular perforin and granzyme B (Fig. 2,a), killed P815 target cells in redirected assays (Fig. 2,b), and produced IFN-γ upon mitogenic challenge (Fig. 2 a). Thus, CCR7− cord blood CD8 T cells homogeneously displayed the functional quiescence characteristic of naive CD8 T cells.
CCR7− cord blood CD8 T cells may derive from CD8 single-positive thymocytes
Although CCR7− cord blood CD8 T cells were homogeneously CD27+ (99 ± 2%), it contained a fraction of CD28− cells (23 ± 10%; Fig. 3,a). To address the possibility that these cells originated directly from thymus, we investigated the cell surface expression of CCR7, CD27, and CD28 by five-color flow cytometry analysis of fresh human thymocyte single-cell suspensions. CCR7, CD27, and CD28 expression was gradually acquired along with thymocyte maturation, i.e., from immature CD1a+ CD3−CD4−CD8− to mature CD1a−CD3+CD4−CD8+ thymocytes, albeit with different kinetics (Fig. 3,b). Although mature CD8 SP thymocytes were homogeneously CD27+ (99 ± 1%), a fraction remained CCR7− (30 ± 12%) and CD28− (10 ± 11%). In fact, CD28− cells were selectively found within CCR7− CD8 SP thymocytes (28 ± 14%). Hence, CCR7+ and CCR7− CD8 SP thymocytes (Fig. 3,c) resembled CCR7+ and CCR7− cord blood CD8 T cells (Fig. 3 a), respectively. This suggests that human CCR7− CD8 SP thymocytes are exported from the thymus through a CCR7-independent pathway. T cells of extrathymic origin might express lower levels of CD3 and TCR (18). However, levels of CD3 and TCR expression were similar for cord blood CD8 T cells and CD8 SP thymocytes regardless of their CCR7 phenotype (n = 10 pairs of thymic and cord blood lymphocyte samples; data not shown), which further suggests that naive CCR7− CD8 T cells originate from the thymus. Comparable to their CCR7+ counterpart, CCR7− cord blood CD8 T cells exhibited an annexin V− phenotype (data not shown) and hence were comparably able to survive in the periphery. We also found that purified CCR7− as well as CCR7+ cord blood CD8 T cells fully preserved their phenotype over at least 10 days of in vitro culture in the presence or the absence of low dose IL-2/IL-7 (data not shown). These results suggest that human thymus exports two types of long-lived precursors, namely CCR7+ and CCR7− cells.
Homing potential of CCR7− cord blood CD8 T cells
The lack of CCR7 expression by CD8 T cells is associated with impaired migration to secondary lymphoid organs (1, 3, 4, 5, 19, 20). Accordingly, the majority of CCR7− cord blood CD8 T cells did not express CD62L (79 ± 6%), a selectin involved in lymph node homing (Fig. 4). In contrast, CD11b, a β2 integrin mediating adhesion to endothelial cells and extravasation (21), was selectively expressed by CCR7− CD8 T cells (41 ± 18%). As well, the cutaneous lymphocyte Ag, CLA, a carbohydrate determinant present on P-selectin glycoprotein ligand 1 and involved in the recruitment to the skin (22), was expressed by one-third of CCR7− cord blood CD8 T cells (33 ± 16%). Hence, such cells expressed molecular keys for homing to peripheral tissues such as skin. Chemokine receptors CXCR1/2/3 and CCR5/6/9 were not expressed on either CCR7+ or CCR7− cord blood CD8 T cells, in accordance with the absence of inflamed environment in healthy newborns (data not shown). Furthermore, both CCR7+ and CCR7− cord blood CD8 T cells were negative for αEβ7 and α4β7 integrins and hence were not directed to the mucosa (data not shown).
We identified naive CD8 T cells in newborns that not only lack the expression of CCR7 and CD62L, two major molecules involved in LN homing, but also express CLA, which allow them to selectively migrate to nonlymphoid tissues such as skin. By measuring their post-thymic replicative history, TCR repertoire, and immune function, we provide evidence that CCR7− CD62L− cord blood CD8 T cells are naive and hence are not components of an ongoing immune response.
A significant fraction of fully mature CD8 SP thymocytes also exhibit a CCR7− phenotype. Furthermore, CCR7−CD8 SP thymocytes resemble CCR7− cord blood CD8 T cells based on their CD27/CD28 phenotype. These findings suggest that CCR7− cord blood CD8 T cells derive from CD8 SP thymocytes. Although CCR7/CCL19 represents a major pathway of neonatal thymocyte export, it has been reported that some mature thymocytes successfully leave the thymus in CCR7-deficient newborn mice (23). Likewise, we suggest that human CCR7−CD8 SP thymocytes may be exported from the thymus through a CCR7-independent pathway. Alternatively, it cannot be excluded that a subpopulation of CCR7+ cord blood CD8 T cells might lose/mask/internalize the CCR7 determinant after thymic export and generate a CCR7− CD8 T cell pool in the periphery. However, when analyzed in vitro over a period of 10 days, CCR7+ cord blood CD8 T cells fully preserved their CCR7+ phenotype (data not shown). In a like manner, CCR7− cord blood CD8 T cells remained CCR7−. These findings further suggest that CCR7+ and CCR7− cord blood CD8 T cells characterize distinct mature, naive T cell pools. It is also formally possible that CCR7− cord blood CD8 T cells might not originate from the thymus, as it has been reported for T cells expressing low levels of CD3 and TCRαβ (18) or for TCRγδ T cells (24). Because CCR7− cord blood CD8 T cells homogeneously express high levels of CD3 and TCRαβ, it is unlikely that they underwent extrathymic differentiation.
The paradigm that naive T cells are excluded from nonlymphoid organs (19, 20) is currently revisited. Naive T cells in mice subjected to chronic inflammation were recently found to traffic to nonlymphoid tissues, and the CCR7/CCR21 pathway has been proposed for naive T cell recruitment (25). In this study we suggest that naive T cell trafficking to nonlymphoid tissues would also take place at birth in the absence of inflammatory conditions and in a CCR7-independent manner. In accordance with our findings, it has been proposed that a fraction of naive T cells could specifically circulate through nonlymphoid tissues during early fetal life in sheep (26) and in neonatal mice (27). Using a transgenic mouse model, Alferink et al. (27) documented that T cell trafficking through nonlymphoid tissues in the neonate is essential for the establishment of tolerance to self, skin-expressed Ags. It is tempting to speculate that the human naive T cells described in this study may also migrate to the skin during the neonatal period and acquire tolerance to peripheral Ags. The lack of expression of the costimulatory receptor CD28 by this subset of neonatal CD8 T lymphocytes is compatible with this hypothesis. In addition, it is conceivable that the trafficking at birth of naive T cells through nonlymphoid tissues could play a role in initiating a T cell niche in this compartment. It is also formally possible that the presence in nonlymphoid newborn tissues of naive T cells with a diverse TCR repertoire could be advantageous in protecting the host against exposure to pathogen-derived neoantigens. Future studies may identify additional homing patterns and receptors of naive T cells, opening the possibility to investigate further subsets of T cells specialized for various peripheral tissues.
The anti-CCR7 mAb and signal joint internal standard were generous gifts from Dr. M. Lipp (Max Delbrück Institute, Berlin, Germany) and Dr. D. Douek (Human Immunology Section, Vaccine Research Center, National Institutes of Health, Bethesda, MD), respectively. We thank Dr. B. Moser (Theodor-Kocher Institute, University of Bern, Bern, Switzerland) for his advice on chemokine receptor mAb staining, Dr. N. Lubenow (Blood Transfusion Center, University of Greifswald, Greifswald, Germany) for providing blood samples, and Céline Baroffio, Andrée Porret, and Martine van Overloop for excellent technical assistance.
This work was supported in part by the Emmy-Noether Program of the Deutsche Forschungsgemeinschaft (to A.Z.), and a grant from the National Center of Competence in Research Program on Molecular Oncology (to F.-A.L.G. and N.R.).
Abbreviations used in this paper: CLA, cutaneous lymphocyte Ag; CDR3, complementary-determining region 3; SP, single positive; TCR BV, TCR β-variable gene segment; TREC, TCR excision circle.