Common γ chain
We have compared four human subsets of Vγ9Vδ2 T cells, naive (Tnaive, CD45RA+CD27+), central memory (TCM, CD45RA–CD27+), effector memory (TEM, CD45RA–CD27–) and terminally differentiated (TEMRA, CD45RA+CD27–), for their capacity to proliferate and differentiate in response to antigen or homeostatic cytokines. Cytokine responsiveness and IL-15R expression were low in Tnaive cells and progressively increased from TCM to TEM and TEMRA cells. In contrast, the capacity to expand in response to antigen or cytokine stimulation showed a reciprocal pattern and was associated with resistance to cell death and Bcl-2 expression. Whereas antigen-stimulated cells acquired a TCM or TEM phenotype, IL-15-stimulated cells maintained their phenotype, with the exception of TCM cells, which expressed CD27 and CD45RA in various combinations. These results, together with ex vivo bromodeoxyuridine incorporation experiments, show that human Vγ9Vδ2 memory T cells have different proliferation and differentiation potentials in vitro and in vivo and that TEMRA cells are generated from the TCM subset upon homeostatic proliferation in the absence of antigen.
Vγ9Vδ2 cells represent a major peripheral blood T cell subset in humans (up to 1/20 of the peripheral blood lymphoid pool) and display a broad reactivity against microbial agents and tumors. They recognize both microbial metabolites (intermediates of the non-mevalonate pathway of isoprenoid biosynthesis) and endogenous metabolites of the mevalonate pathway, whose production is up-regulated upon cell stress 1.
Similar to CD4 and CD8 αβ T cells, Vγ9Vδ2 T lymphocytes are heterogeneous and comprise distinct populations that can be distinguished on the basis of surface marker expression and effector functions, such as cytokine secretion and cytotoxicity. Naive (Tnaive) CD45RA+CD27+ and central memory (TCM) CD45RA–CD27+ cells express lymph node homing receptors, abound in lymph nodes and lack immediate effector functions. Conversely, effector memory (TEM) CD45RA–CD27– and terminally differentiated (TEMRA) CD45RA+CD27– cells express receptors for migration to inflamed tissues, are poorly represented in the lymph nodes while abounding at sites of inflammation, where they display immediate effector functions (cytokine production and cytotoxicity, respectively) 2.
The differentiation pathways that lead to the generation of these cells, in particular TEM and TEMRA cells, are uncertain. A progressive and irreversible differentiation in vitro from Tnaive to TCM and TEM has been demonstrated for Vγ9Vδ2 T lymphocytes 2, but the relationship between the various subpopulations is still unclear. Moreover, little is known about the homeostasis of human Vγ9Vδ2 T lymphocytes and, in particular, how the functional heterogeneity of the memory pool is maintained.
Naive and memory T lymphocyte pools are maintained at a constant size to ensure that the organism can mount an immune response to a variety of new antigens, while keeping appropriate levels of memory cells to previously encountered pathogens 3–5.
Signaling from the TCR and cytokine receptors has been implicated in the maintenance and homeostatic expansion of T cell pools, but the relative contribution of these receptors may vary under different conditions. In particular, it has been shown that naive CD4 and CD8 T cells need to interact with self MHC molecules in order to survive and proliferate, while memory T cells show no such requirement 6–13.
There is growing evidence that cytokines that bind to receptors containing the common γ chain (γc), namely IL-2, IL-4, IL-7 and IL-15, are involved in T cell maintenance and homeostasis. IL-15, produced by several cell types including dendritic cells (DC), plays an essential role in T cell homeostasis: it expands CD8 memory T cells in vivo14, 15, and deletion of the IL-15R results in reduced numbers of CD8 memory cells 16. The effects of IL-15 can be counteracted by IL-2, which primes mouse T cells for death, suggesting that CD8 homeostasis might be regulated at the level of these two opposing cytokines 17. IL-7, produced by stromal cells, can enhance homeostatic proliferation of both naive and memory CD8 T cells 18.
Because Tnaive, TCM, TEM and TEMRA Vγ9Vδ2 T lymphocytes have different capacities to traffic in lymphoid and nonlymphoid tissues 2, it is possible that they have access to different environments and possibly to different survival stimuli, although the responsiveness of naive and memory Vγ9Vδ2 T cell subsets to cytokines has not been investigated. We therefore considered the possibility that the phenotype and function of distinct subsets of Vγ9Vδ2 T lymphocytes may be influenced by homeostatic mechanisms. Consequently, we investigated the proliferation and differentiation of human Vγ9Vδ2 T lymphocyte subsets in response to antigen or homeostatic cytokines. Results reported here indicate that Vγ9Vδ2 T lymphocyte subsets have different proliferative capacities in vitro and that TCM cells have the unique ability to differentiate in an antigen-independent fashion into TEM and TEMRA Vγ9Vδ2 T lymphocytes.
Expansion of different subsets of Vγ9Vδ2 T cells stimulated by antigen or homeostatic cytokines
Purified Tnaive, TCM, TEM and TEMRA subsets of Vγ9Vδ2 T lymphocytes were labeled with CFSE and compared for their capacity to proliferate in response to phosphoantigen or to homeostatic cytokines. Antigenic stimulation, provided by isopentenyl pyrophosphate (IPP) + IL-2, resulted in consistent expansion of Tnaive and TCM cells, but TEM cells and especially TEMRA cells performed few divisions (if any at all) and were recovered in lower numbers (Fig. 1, 2A). Of note, and in agreement with previously published data, no proliferation or cell division was obtained with IPP alone in the absence of exogenous IL-2 2 (and data not shown). Although TEM and particularly TEMRA Vγ9Vδ2 cells proliferate weakly, antigenic stimulation elicits effector responses such as IFN-γ production and cytotoxicity, respectively 2. Moreover, challenge with IPP determines the up-regulation of CD69, which is very marked in TCM and in TEM Vγ9Vδ2 cells and moderate in TEMRA cells. Vγ9Vδ2 Tnaive cells did not up-regulate CD69 expression in the first 12 h following antigen exposure (Fig. 3).
The low expansion potential of phosphoantigen-stimulated TEM and TEMRA cells was associated with a high rate of cell death as revealed by propidium iodide staining (Fig. 2B) and correlated with low expression of the anti-apoptotic Bcl-2 protein (Table 1).
Proliferation in response to IL-7 was comparable and low in all subsets, as shown by both the number of divisions performed and the relatively low numbers of cells recovered. In contrast, responsiveness to IL-15 was low in Tnaive cells, intermediate in TCM cells, and high in TEM and TEMRA cells (Fig. 1, 2A). Proliferation of different subsets in response to a combination of IL-7 and IL-15 did not differ from that seen with IL-15 alone (data not shown).
Cytokine responsiveness correlated with the expression of the relevant cytokine receptors (Fig. 2C, Table 1). Thus, whereas IL-7Rα chain expression was high on Tnaive and TCM cells and slightly decreased in TEM and TEMRA cells, IL-15Rα and IL-2/15Rβ chain expression were low on Tnaive cells, intermediate on TCM cells, and high on TEM and TEMRA cells. The γc and the IL-2Rα chain were expressed to a comparable level in all subsets.
Consistent with their high cytokine responsiveness, TCM and TEM cells were recovered in higher numbers after stimulation with IL-15, as compared to naive T cells (Fig. 2A). In contrast, TEMRA cells failed to accumulate, despite the fact that they expressed high levels of IL-15R, and most of them underwent cell division. The failure of cytokine-stimulated TEMRA cells to accumulate was associated with a high rate of cell death (Fig. 2B) and low Bcl-2 expression (Table 1), suggesting that the intrinsic high propensity to cell death compromises the accumulation of TEMRA cells in response to homeostatic cytokines, despite their high cytokine responsiveness.
Altogether, these results show that whereas antigen-dependent expansion, cell viability and Bcl-2 expression are progressively lost from Tnaive and TCM to TEM and TEMRA cells, IL-15R expression and cytokine responsiveness have a reciprocal pattern and are progressively acquired with differentiation.
Generation of TEMRA cells by cytokine-stimulated TCM cells
To assess the potential differentiation of Vγ9Vδ2 T lymphocytes, sorted Tnaive, TCM, TEM and TEMRA cells were labeled with CFSE and stimulated with either phosphoantigen or IL-15. On day 7, CD27 and CD45RA expression were analyzed on cells that had performed the same number of divisions. Typical results from one experiment out of five are shown in Fig. 4. In agreement with our previously published data 2, antigenic stimulation of Tnaive cells resulted in the generation of TCM cells, while antigenic stimulation of TCM and TEM cells resulted in a homogeneous population of the latter cells. Antigen-stimulated TEMRA cells maintained their phenotype, but their number consistently decreased over the 7-day culture period. In no case did antigen stimulation of Tnaive, TCM and TEM cells give rise to TEMRA cells.
In contrast, whereas Tnaive, TEM and TEMRA cells proliferating in response to IL-15 largely maintained their phenotype, TCM cells gave rise to TCM cells and to TEM and TEMRA cells, the latter showing high perforin and granulysin contents (data not shown). Thus, some cytokine-stimulated TCM cells maintained their phenotype, whereas others generated cells with a TEM or a TEMRA phenotype.
Altogether, these results suggest that cytokine-stimulated TCM cells, in the absence of antigen, can self-renew and generate different types of effector cells, including TEMRA cells.
Vγ9Vδ2 TEMRA cells generated by IL-15-stimulated TCM cells exert potent anti-tumor and antibacterial activity
Vγ9Vδ2 T cell lines and clones exert potent cytotoxic activities against tumors and Mycobacterium tuberculosis (Mtb)-infected macrophages and also kill intracellular bacteria 1, 19, 20. Additionally, when challenged, they produce large amounts of IFN-γ. We therefore compared IFN-γ production and cytotoxic activity of TEM and TEMRA cells generated by IL-15-stimulated TCM.
As shown in Fig. 5A, only Vγ9Vδ2 TEM but not Vγ9Vδ2 TCM and TEMRA cells were induced to produce IFN-γ following co-culture with Daudi Burkitt lymphoma cells or Mtb-infected human macrophages. However, Vγ9Vδ2 TEMRA cells exerted potent cytotoxicity against Daudi cells (Fig. 5B) and Mtb-infected macrophages (Fig. 5C) and were also able to consistently reduce the viability of intracellular Mtb (Fig. 5D). Vγ9Vδ2 TCM and TEM cells did not show either cytotoxic or microbicidal activities. Similar results were obtained using TEM and TEMRA cells generated from IL-15-stimulated Vγ9Vδ2 TCM cells from three different donors.
Together, these features clearly demonstrate that the pool of Vγ9Vδ2 TCM cells is able to generate two subsets of functionally active effector T cells with different phenotypic and functional properties.
Differential in vivo turnover of different subsets of Vγ9Vδ2 T cells
Memory T cells slowly turn over under steady state conditions in vivo21–23. Cells that pass the G1 cell cycle checkpoint are committed to replicate DNA and complete the cell cycle 22. The in vivo turnover of T cells can therefore be assessed by ex vivo bromodeoxyuridine (BrdU) incorporation 21. Freshly purified PBMC from five healthy donors were cultured in the presence of BrdU, and Vγ9Vδ2 T cells were purified and analyzed for BrdU incorporation in relation to CD27 and CD45RA expression. Typical results are shown in Fig. 6. Most cells that incorporated BrdU expressed the CD45RA– and CD27+ TCM and the CD45RA–CD27– TEM phenotype, while very few Tnaive (CD45RA+CD27+) and TEMRA (CD45RA+CD27–) cells incorporated BrdU (Fig. 6B). Consequently, TCM cells had the highest and TEMRA cells the lowest turnover among memory cells, whereas the turnover of Tnaive cells was below the detection limit. The mean percentages of BrdU+ cells, in each subset, of the five donors analyzed are shown in Fig. 6C.
The selective expansion, in adult peripheral blood, of γδ T cells bearing the Vγ9Vδ2 rearrangement reflects their reactivity towards common environmental antigens; indeed, most Vγ9Vδ2 T cells share identical antigen specificity. Thus, contrary to αβ T cells where the frequency of cells specific for any given antigen is very low, Vγ9Vδ2 T cells represent a unique system to follow T cell differentiation and acquisition of memory following encounter with antigen. A clear understanding of the mechanisms that underlie differentiation and maturation of memory T cell subsets is a necessary prerequisite to effectively select and drive the more "suitable" subset for therapeutic purposes.
In this paper, we show that different subsets of human Vγ9Vδ2 T cells have different capacities to proliferate and differentiate in response to antigen stimulation or homeostatic cytokines. In particular, we show that Vγ9Vδ2 TCM cells can be expanded with IL-15, in the absence of antigen stimulation, and that they give rise to TEMRA cells with a characteristic CD27–CD45RA+ phenotype, expressing perforin and granulysin (data not shown) and killing tumor and mycobacteria-infected targets.
Both constitutive and lymphopenia-induced proliferation of mouse T cells depend on the homeostatic cytokines IL-7 and IL-15 15, 16, 24–27. Although the relative contributions of different cytokines to homeostatic proliferation in humans are difficult to address directly, recent reports are consistent with the view that human and murine T cells have similar requirements 13, 14, 21, 28–32. Thus, antigen-independent naive T cell proliferation in the mouse requires TCR tickling, IL-7, and lymphopenia, whereas CD8 memory T cells turn over constitutively, express high levels of the IL-2/15Rβ chain and require IL-15 or IL-7.
Consistent with the above quoted data, we found that human Vγ9Vδ2 TCM but not Tnaive cells incorporate BrdU ex vivo, proliferate in response to IL-15 and express the IL-15R.
Moreover, we show here that expression of the IL-15R and cytokine responsiveness are associated with differentiation and increase from Tnaive to TCM and TEM Vγ9Vδ2 T cells. Thus, cytokine responses of human Vγ9Vδ2 T cells recapitulate many aspects of homeostatic T cell proliferation in vivo and provide a valuable tool to extend our knowledge of T cell maintenance in the human system.
Memory T cells have an enhanced susceptibility to cell death 33, 34 and die upon bystander activation 35, although surviving cells may proliferate and repopulate the memory pool. We show that the resistance of human Vγ9Vδ2 T cell subsets to cell death is associated with the expression of the anti-apoptotic Bcl-2 protein, which protects T cells from multiple forms of cell death 36. Both antigen and cytokines up-regulate Bcl-2 in all subsets, but the differences in Bcl-2 expression between the subsets are maintained, supporting the notion that high susceptibility to cell death is an intrinsic feature of TEMRA cells. Consequently, Vγ9Vδ2 TEMRA cells have a low expansion potential in response to both antigen and cytokines in vitro and a low turnover in vivo.
Naive T cells expanding in lymphopenic mice differentiate, acquiring some characteristics of effector/memory cells 37–39: similarly, and in agreement with data reported for CD4 13 and CD8 40 T cells, cytokine-stimulated human Vγ9Vδ2 Tnaive cells maintain a naive phenotype because they retain CD45RA and CD27 expression. In contrast, cytokine-stimulated TCM cells differentiate upon cytokine stimulation and generate various types of effector cells, which express CD27 and CD45RA in various combinations.
CD45RA+ effector cells (TEMRA) are a dynamic and enigmatic population of the Vγ9Vδ2 memory pool, with potent cytotoxic activities against tumors and intracellular bacteria 1, 19, 20. Several reports have described the appearance of αβ CD8 TEMRA lymphocytes following viral infections 41, 42. TEMRA cells were shown to represent the main Vγ9Vδ2 T cell subset present in ascites and cerebrospinal fluids of tuberculosis patients 2, confirming the migratory capability of these end-stage effectors.
Various lines of evidence, such as the loss of CD27 and CCR7 2, the low proliferative capacity, the high susceptibility to apoptosis, the presence of high levels of perforin and, indirectly, the shorter telomeres of TEM Vγ9Vδ2 T cells 2, indicate that they represent the most differentiated type of Vγ9Vδ2 memory T cells.
Strongly suggestive of a progressive selection for the fittest effectors 43, differentiation into Vγ9Vδ2 TEMRA cells produces the most highly active anti-tumoral effectors among Vγ9Vδ2 T cells. Given the negative influence of phosphoantigens on the in vitro induction of Vγ9Vδ2 TEMRA cells from TCM cells or even whole PBMC 44, their presence in vivo most probably reflects a prolonged absence of stimulation by microbial phosphoantigens, as proposed for αβ CD8 TEMRA cells 40. Indeed, the pool of proliferating non-effector Vγ9Vδ2 cells must generate either TCR-dependent helpers (TEM) or cytolytic effectors (TEMRA) by some as yet incompletely defined terminal maturation switch. While phosphoantigen clearly drives the proliferation of precursor pools and the maturation of Vγ9Vδ2 TEM cells 2 (and this manuscript), we found that cells with a phenotype corresponding to TEMRA, i.e. CD45RA+ CD27– perforin+, are generated rather exclusively by a Vγ9Vδ2 TCM subset upon IL-15 stimulation, suggesting that Vγ9Vδ2 TEMRA cell generation requires homeostatic proliferation in the absence of antigen. Ex vivo BrdU incorporation results are consistent with the view that TEMRA cells are continuously replenished from proliferating precursors of the TCM pool.
Altogether, our results indicate that human Vγ9Vδ2 memory T cell subsets have different capacities to proliferate, differentiate and resist to cell death in response to antigen and homeostatic cytokines. Moreover, we suggest that Vγ9Vδ2 TCM cells, due to their high expansion and differentiation potential and in vivo turnover, play an important role in maintaining the heterogeneous memory pool.
Future studies will now aim at elucidating the conditions that favor the selective generation of human Vγ9Vδ2 TEMRA lymphocytes, due to the need of generating cytotoxic effectors for anticancer immunotherapy purposes.
Materials and methods
FACS staining and sorting
PBMC were isolated from heparinized blood of healthy donors by Ficoll-Hypaque (Pharmacia Biotech, Uppsala, Sweden). PBMC were incubated with the following antibodies in different combinations: anti-Vδ2-FITC (Coulter, Miami, FL), anti-CD27-PE (BD PharMingen, San Diego, CA), anti-CD45RA-PE-Cy5 (Coulter), anti-CD45RO-PE-Cy5 (Coulter), anti-CD3-PE (Sigma, St. Louis, MO), and unconjugated anti-Vγ9 (BD PharMingen), anti-CD69 (BD PharMingen), anti-CD25 (BD PharMingen), anti-CD122 (IL-2/IL-15Rβ; BD PharMingen), anti-CD132 (γc; BD PharMingen), anti-IL-7Rα (R&D, Minneapolis, MN), anti-IL-15Rα (R&D). Bcl-2 expression was assessed by anti-Bcl-2 mAb (BD PharMingen) after fixation with paraformaldehyde and permeabilization with saponin. Data were acquired on a FACSCalibur or a FACSCanto instrument (BD Biosciences) and analyzed using CellQuest software (BD Immunocytometry Systems, San Jose, CA) or FlowJo (Treestar, OR). Labeling of Vδ2 T cells with CFSE (Molecular Probes, Eugene, OR) was performed as described 45.
Vγ9Vδ2 T cells were isolated by positive selection with magnetic beads (MACS; Miltenyi, Bergisch Gladbach, Germany). The cells obtained were more than 98% Vγ9Vδ2+ CD3+. Tnaive, TCM, TEM and TEMRA Vγ9Vδ2 T lymphocytes were then purified to more than 99% by cell sorting, using anti-CD45RA and anti-CD27 antibodies. Cell sorting was performed on a FACSVantage (BD Biosciences). The distribution between the four Vγ9Vδ2 T lymphocyte subsets of ten donors was the following: Tnaive, 15±4%; TCM, 46±10%; TEM, 32±8%; and TEMRA, 7±2%.
Ex vivo BrdU labeling
This assay was carried out by using the procedure reported by Lempicki et al. 21. Vγ9Vδ2 T cells were isolated by positive selection as described above and incubated with 10 µM BrdU (Sigma) for 4 h at 37°C, 5% CO2. Cell surface staining was then performed by using anti-CD27 and anti-CD45RA antibodies. The stained cells were treated with lysing solution (BD) for 10 min at room temperature, followed by incubation with 1% paraformaldehyde and 1% Tween-20 in PBS for 15 min at 37°C to fix and permeabilize the cells. Cellular DNA in the permeabilized cells was partially digested with DNase I (Boehringer Mannheim) in DNase buffer (PBS with 4.2 mM MgCl2, pH 5) for 30 min at 37°C, then stained for 30 min with anti-BrdU-FITC (BD) antibody in PBS containing 1% BSA and 0.5% Tween-20. Cells were washed twice before flow cytometric analysis. A total of 50,000–100,000 events were acquired, resulting in a sensitivity of 0.01% BrdU+ events. Samples were analyzed in parallel with unlabeled cells (without BrdU) from the same individual, and this value was subtracted from the value obtained for BrdU-labeled cells. Data are presented as the percentage of cells in the specific lymphocyte pool that are BrdU+.
The medium used was complete RPMI 1640 medium (Gibco, Grand Island, NY) supplemented with 10% heat-inactivated FCS (Gibco), 2 mM L-glutamine, 20 nM Hepes and 100 U/ml penicillin/streptomycin. Tnaive, TCM, TEM and TEMRA Vγ9Vδ2 T lymphocytes, sorted as described above, were labeled with CFSE and cultured for 7 days at 37°C, in the presence of 5% CO2, at 106 cells/ml in 96-well flat-bottom plates (0.2 ml/well), with different concentrations of IPP (Sigma) and 20 U/ml human recombinant IL-2 added at the 3rd day of culture 2. Alternatively, sorted Vγ9Vδ2 subsets were cultured with recombinant IL-2 (20 U/ml, final concentration), IL-7 or IL-15 (both used at 25 ng/ml final concentration). These concentrations were determined in preliminary dose-response experiments. After the culture period, cells were recovered, and proliferation and surface marker expression were assessed by FACS. The Vγ9Vδ2 expansion factor (EF) was calculated by dividing the absolute number of Vγ9Vδ2 cells in antigen- or cytokine-stimulated cultures by the number of Vγ9Vδ2 cells put in at the beginning of the culture.
Infection of macrophages with Mtb
The myelomonocytic THP-1 target cells (4×103 cells/well) were incubated with phorbol 12-myristate 13-acetate (PMA; Sigma) at a final concentration of 10 ng/ml in 96-well round-bottom plates for 24 h at 37°C in the presence of 5% CO2. Nonadherent cells were removed and the macrophages were infected with Mtb strain H37Ra overnight at a multiplicity of infection of 10:1. After extensive washing, macrophages were detached and the efficiency of infection was determined by staining a sample portion with auramine-rhodamine acid-fast staining. Approximately 85% of the cells were infected with Mtb, with an average of three bacteria per cell.
Cytotoxicity assay and assessment of viability of Mtb bacilli in human macrophages
Mtb-infected macrophages (104) or Daudi lymphoma cells (104) were incubated for 5 h at 37°C in 96-well round-bottom plates with Vγ9Vδ2 T cells at E:T ratios of 30:1, 10:1 and 1:1. Assays were performed in triplicate for each E:T ratio. Cytotoxicity was analyzed using a nonradioactive colorimetric cytotoxicity assay (CytoTox 96; Promega).
IFN-γ levels in the 24-h culture supernatants were assessed by two-mAb sandwich ELISA assay following the manufacturer's recommendations (R&D Systems).
Mtb-infected macrophages and Vγ9Vδ2 T cells were incubated for 20 h at 37°C, washed three times to eliminate bacteria that were not cell associated, lysed with 0.1% saponin and sonicated for 20 s. Serial tenfold dilutions were made in 7H9 broth and plated on 7H10 agar plates. Plates were sealed in plastic, kept at 37°C, and the number of colonies was counted after 14–21 days.
Student's t-test was used to compare the significance of differences between cytokine receptor expression in different subsets of Vγ9Vδ2 T cells (see Table 1).
We thank Dr. Federica Sallusto for critically reviewing the manuscript. This work has been supported by grants from the Commission of the European Union, 6FP, contract LSHP-CT-2003–503367, "An integrated project for design and testing of vaccine candidates against tuberculosis: identification, development and clinical studies" (TB-VAC). F.D. is supported by a grant from the University of Palermo.