Abstract
To better apprehend γ/δ T cell biological functions in the periphery, it appears crucial to identify markers highlighting the existence of distinct phenotypic and functional γ/δ T cell subsets. Interestingly, the expression of CD44 and Ly-6C subdivides murine peripheral γ/δ T cells into several subsets, with Ly-6C− CD44hi γ/δ T cells corresponding to the IL-17–producing CD27− γ/δ T cell subset exhibiting innate-like features. By comparing the other subsets to naive and memory CD8+ α/β T cells, in this study, we show that Ly-6C− or + CD44lo and Ly-6C+CD44hi γ/δ T cells greatly resemble, and behave like, their CD8+ α/β T cell counterparts. First, like memory CD8+ α/β T cells, Ly-6C+CD44hi γ/δ T cells are sparse in the thymus but largely increased in proportion in tissues. Second, similarly to naive CD8 α/β T cells, CD44lo γ/δ T cells are poorly cycling in vivo in the steady state, and their proportion declines with age in secondary lymphoid organs. Third, CD44lo γ/δ T cells undergo spontaneous proliferation and convert to a memory-like Ly-6C+CD44hi phenotype in response to lymphopenia. Finally, CD44lo γ/δ T cells have an intrinsic high plasticity as, upon appropriate stimulation, they are capable of differentiating nonetheless into Th17-like and Th1-like cells but also into fully functional Foxp3+ induced regulatory T cell–like γ/δ T cells. Thus, peripheral CD27+ γ/δ T cells, commonly considered as a functionally related T cell compartment, actually share many common features with adaptive α/β T cells, as both lineages include naive-like and memory-like lymphocytes with distinct phenotypic, functional, and homeostatic characteristics.
This article is featured in In This Issue, p.1337
Introduction
The γ/δ T cells are a well-conserved T cell population throughout evolution and across species. Indeed, in every vertebrate in which T cell ontogeny has been examined, γ/δ T cells appear to be the first T cells to develop (1, 2). γ/δ T cells are unique and distinct from other lymphocyte subsets, such as NK cells, B cells, and α/β T cells, in that they combine adaptive features with rapid, innate-like responses that allow them to play an important role in all phases of an immune response.
Unlike α/β T cells, γ/δ T cells are produced in the thymus in sequential waves during defined periods of fetal, neonatal, or adult life (3). Most γ/δ T cells are not restricted to the recognition of peptides bound to conventional MHC molecules, which, once again, distinguishes them from the great majority of α/β T cells. Indeed, it has been shown that γ/δ T cells are reactive either to self-MHC molecules (4), to non–peptide-binding MHC class Ib molecules such as mouse H2-T10 and H2-T22 (5, 6), or to stress-induced proteins with or without conformational similarities with MHC molecules (7, 8). However, the paucity of the present knowledge on γ/δ TCR ligands confounds the full dissection of γ/δ T cell activation processes and thus compromises the in-depth characterization of the functions of these cells in the periphery. To overcome this, a strong effort has been made recently to identify markers highlighting the existence of distinct phenotypic and functional T cell subsets within the peripheral γ/δ T cell compartment (9, 10). In secondary lymphoid organs, this distinction appears predominantly based on γ/δ T cell cytokine production profiles and the expression of specific surface markers. Indeed, peripheral γ/δ T cells are currently subdivided into two subsets depending on their expression of CD27 and according to their ability to produce IL-17 or IFN-γ (11–14). In this way, robust IFN-γ production is associated with a CD27+ phenotype, whereas secretion of IL-17 is restricted to CD27− γ/δ T cells (11). This phenotypic and functional dichotomy is established during thymic development, as γ/δ T cells that do not express the costimulatory receptor CD27 are considered to emerge by default from thymic progenitor cells receiving only weak TCR signals during development, in contrast to the strong signals required for CD27+ γ/δ T cell generation (15, 16). However, in analogy to the α/β T cell compartment, the existence of naive or memory γ/δ T cells remains to be determined.
In this study, we show that murine peripheral γ/δ T cells can be subdivided into four subsets according to CD44 and Ly-6C expression. Ly-6C−CD44hi γ/δ T cells correspond to the CD27−IL-17+ γ/δ T cell subset that exhibit innate-like features (12). By comparing the other subsets to naive and memory CD8+ α/β T cells, we found that the CD27+ γ/δ T cell compartment include naive-like and memory-like lymphocytes that share many phenotypic, functional, and homeostatic characteristics with their adaptive α/β T cell counterparts.
Materials and Methods
Mice
C57BL/6 mice (CD45.2) were obtained from Charles River Laboratories. C57BL/6 CD45.1 mice, C57BL/6 CD3ε−/− mice (17), and C57BL/6 CD3ε−/−MHC class IIΔ/Δ mice (18) were maintained in our own animal facilities, under specific pathogen-free conditions. C57BL/6 Foxp3-GFP CD45.2 mice were initially obtained from Dr. Bernard Malissen (Centre d'Immunologie de Marseille-Luminy, France) (19). Donor and recipient mice were sex-matched. Six- to 8 wk-old mice were used for experiments unless otherwise indicated (Fig. 5). Animal housing, care, and research were carried out in accordance with the guidelines of the French Veterinary Department. All procedures performed were approved by the Paris-Descartes Ethical Committee for Animal Experimentation (decision CEEA34.BL.002.12).
Peripheral homeostasis of γ/δ T and TCRβ+ CD8+ T cell subsets. (A) Ki-67 expression by pLNs γ/δ T cell subsets is shown for a representative C57BL/6 mouse. (B) Proportions of Ki-67+ cells among γ/δ T cell and TCRβ+ CD8+ T cell subsets recovered from pLNs, mLNs, and spleen from C57/BL/6 mice. Results are shown as means ± SEM. (C) Proportions of γ/δ T cell and TCRβ+ CD8+ T cell subsets from LNs of 6–8-wk-, 8–10-mo-, or 18-mo-old C57BL/6 mice. Each symbol represents an individual mouse. *p < 0.05, **p < 0.01, ***p < 0.001.
Peripheral homeostasis of γ/δ T and TCRβ+ CD8+ T cell subsets. (A) Ki-67 expression by pLNs γ/δ T cell subsets is shown for a representative C57BL/6 mouse. (B) Proportions of Ki-67+ cells among γ/δ T cell and TCRβ+ CD8+ T cell subsets recovered from pLNs, mLNs, and spleen from C57/BL/6 mice. Results are shown as means ± SEM. (C) Proportions of γ/δ T cell and TCRβ+ CD8+ T cell subsets from LNs of 6–8-wk-, 8–10-mo-, or 18-mo-old C57BL/6 mice. Each symbol represents an individual mouse. *p < 0.05, **p < 0.01, ***p < 0.001.
Cell suspensions
Peripheral lymph nodes (pLNs), mesenteric LNs (mLNs), cervical LNs, Peyer’s patches (PP), spleen, and thymus were homogenized and passed through a nylon cell strainer (BD Falcon) in RPMI 1640 Glutamax (Life Technologies) supplemented with 10% FCS (Biochrom) for adoptive transfer and cell culture (LNs only) or in 5% FCS and 0.1% NaN3 (Sigma-Aldrich) in PBS for flow cytometry.
Tissue processing
Ketamine-xylazine–anesthetized mice were first perfused intracardiacally with cold PBS. Small intestines were opened longitudinally after removal of PP, washed in successive HBSS 5% FCS baths, and incubated in RPMI 1640 20% FCS, with collagenase D (Roche) 1 mg/ml and DNase I (Roche) 0.1 mg/ml for 30 min at 37°C. Cells were then recovered after Percoll gradient separation (40–70%). Livers were dissected and mashed in HBSS 2% FCS and 10 mmol HEPES. Cells were then recovered after Percoll gradient separation (35%). Lungs were cut in small parts and digested with collagenase D (Roche) 1 mg/ml and DNase I (Roche) 0.1 mg/ml for 30 min at 37°C.
Fluorescence staining and flow cytometry
Cell suspensions were collected and dispensed into 96-well round-bottom microtiter plates (Greiner Bioscience; 6 × 106 cells/well). Surface staining was performed as previously described (20, 21). Briefly, cells were incubated on ice, for 15 min per step, with Abs in 5% FCS (Biochrom) and 0.1% NaN3 (Sigma-Aldrich) PBS. Each cell-staining reaction was preceded by a 15-min incubation with a purified anti-mouse CD16/32 Ab (FcγRII/III block; 2.4G2). For determination of intracellular cytokine production, cells were stimulated with 0.5 μg/ml PMA (Sigma-Aldrich), 0.5 μg/ml ionomycin (Sigma-Aldrich), and 10 μg/ml brefeldin A (Sigma-Aldrich) for 2 h at 37°C. Cells were then stained for surface markers, fixed in 2% paraformaldehyde in PBS, and permeabilized with 0.5% saponin, followed by labeling with specific cytokine Abs. For Foxp3 and Ki-67 intracellular staining, the Foxp3 Staining Buffer Set (eBioscience) was used. Multicolor immunofluorescence was analyzed using BD LSR2 and BD Fortessa cytometers (BD Biosciences). List-mode data files were analyzed using Diva software (BD Biosciences). Data acquisition and cell sorting were performed at the Cochin Immunobiology facility.
Adoptive transfer of γ/δ T cells
γ/δ T cells were purified from LNs (pooled superficial cervical, axillary, brachial, inguinal, and mLNs) of C57BL/6 CD45.1 mice by incubating cell suspensions on ice for 20 min with a mixture of anti-CD8α (53-6.7), anti-CD4 (GK1.5), anti-CD11b (Mac-1), and anti-CD19 (1D3) Abs obtained from hybridoma supernatants and then with magnetic beads coupled to anti-rat Igs (Dynal Biotech). Purified cells were then labeled with biotinylated anti–Ly-6C (AL21), PE-conjugated anti-NK1.1 (PK136), anti-TCRβ (H57-597), anti-B220 (RA3-6B2), anti-CD11b (M-1/70), anti-CD11c (HL3), anti-CD8β (53-7.8), and allophycocyanin-conjugated anti-CD44 (IM7), all from BD Biosciences, and Pacific Blue–conjugated streptavidin (Invitrogen). CD44hiLy6C−, CD44hiLy6C+, CD44loLy6C−, and CD44loLy6C+ γ/δ T cells were flow cytometry sorted as NK1.1−TCRβ−CD8β−B220−CD11b−CD11c− cells using an FACSAria 3 flow cytometer (BD Biosciences) and injected i.v. into sex-matched recipient mice (5 × 104 cells/recipient mice).
In vitro polarization assays
After magnetic enrichment as described above, purified cells from LNs of C57BL/6 Foxp3-GFP mice were then labeled with biotinylated anti–Ly-6C (AL21), PE-conjugated anti-TCRγ/δ (GL3), and allophycocyanin-conjugated anti-CD44 (IM7), all from BD Biosciences, and Pacific Blue-conjugated streptavidin (Invitrogen). CD44hiLy6C−, CD44hiLy6C+, CD44loLy6C−, and CD44loLy6C+ γ/δ T cells were flow cytometry sorted as GFP−TCRγ/δ+ cells and then stimulated for 4 d with immobilized anti-CD3 (clone 145.2C11; 4 μg/ml; obtained from hybridoma supernatants) and anti-CD28 (clone 37.51; eBioscience; 4 μg/ml) Abs, in the presence of LEAF-purified anti–IFN-γ neutralizing Abs (R4-6A2; BioLegend) and graded concentrations of exogenous recombinant human TGF-β1 (Invitrogen) in the presence of 20 ng/ml recombinant mouse IL-6 (Invitrogen) and 10 ng/ml recombinant mouse IL-23 (R&D Systems) (γ/δ 17+) or in the presence of 13 ng/ml recombinant human IL-2 (R&D Systems) (γ/δ regulatory T cell [Treg]). To study the ability of γ/δ T cells to differentiate into γ/δIFN+ effector cells, stimulation was performed in the presence of graded concentrations of recombinant mouse IL-12 (R&D Systems).
In vitro suppression assay
FACS-sorted Foxp3−CD44loTCRγ/δ+ T cells or Foxp3−CD25−CD44lo naive CD4 T cells from LNs of C57BL/6 CD45.1 Foxp3-GFP mice were stimulated for 3 d with coated anti-CD3 and anti-CD28 Abs in the presence of 1 ng/ml TGF-β1 and then allowed to rest for 3 more d in the presence of recombinant human IL-2 (13 ng/ml; R&D Systems). GFP-expressing cells were then flow cytometry sorted, and the suppressive capacities of these highly purified Foxp3-expressing cells were then assessed as previously described (18). Briefly, conventional CD4 T cells (GFP−CD4+ T cells) were purified from LNs of C57BL/6 CD45.2 Foxp3-GFP mice, labeled with Cell Trace Violet (Invitrogen), and stimulated for 64 h, alone or together with Tregs at various Treg/conventional T cell ratios.
TLR stimulation in vitro assay
Flow cytometry–sorted TCRγ/δ+ T cells from LNs of CD45.1 C57BL/6 were stimulated for 16 h in the presence of LPS (100 ng/ml), Pam3CSK4 (1 μg/ml), polyinosinic-polycytidylic acid (Poly I:C; 5 μg/ml), or CpG (1 μg/ml), all from Invivogen.
Intracellular Ca2+ mobilization assay
Cells were loaded for 30 min at 37°C with the membrane-permeable fluorescent Ca2+ indicator dye Indo-1 AM (Invitrogen) at a concentration of 1 μmol in RPMI 1640 medium with no FCS. Thereafter, cells were stained for surface markers and kept on ice. Before stimulation, cell aliquots were allowed to equilibrate to 37°C for 5 min and then were analyzed by flow cytometry. After acquisition of background intracellular Ca2+ concentrations for 1 min, cells were stimulated with 3 μg/ml hamster anti-CD3ε (145-2C11; obtained from hybridoma supernatants) and then were crosslinked by the addition of 20 μg/ml goat anti-hamster Abs (Thermo Fisher).
Statistics
Data are expressed as mean ± SEM, and the significance of differences between two series of results was assessed using the Student unpaired or paired t test. The p values < 0.05 were considered significant (*p < 0.05, **p < 0.01, and ***p < 0.001).
Results
Heterogeneity of Ly-6C and CD44 expression in peripheral γ/δ T cells
γ/δ T cells as well as CD8+ α/β T cells from (LNs of C57BL/6 mice can be subdivided into several subsets according to Ly-6C and CD44 expression (Fig. 1A). More precisely, whereas Ly-6C and CD44 expression split the CD8+ T cell compartment into three subsets (naive Ly-6C− or Ly-6C+ CD44lo CD8+ T cells and effector/memory Ly-6C+CD44hi CD8+ T cells), these markers allow to subdivide peripheral γ/δ T cells into four different subsets. Indeed, in addition to Ly-6C− or Ly-6C+CD44lo cells and Ly-6C+CD44hi cells, a subset of Ly-6C−CD44hi cells can be defined. In the literature, peripheral γ/δ T cells are currently dichotomized into two subsets according to CD27 and CD44 expression (CD27−CD44hi and CD27+CD44int/lo γ/δ T cell subsets; Fig. 1B). We observed that Ly-6C−CD44hi γ/δ T cells correspond to CD27−CD44hi γ/δ T cells, whereas the CD27+CD44int/lo γ/δ T cell compartment, commonly considered as functionally related, can be subdivided into three different subsets on the basis of CD44 and Ly-6C expression (Fig. 1B). Although the proportion of these newly defined subsets varied strongly according to the analyzed organ, this applied also to γ/δ T cells and CD8+ T cells from the spleen (Fig. 1C), peritoneal cavity, liver, and lung (Fig. 1D). By contrast, in the thymus, nearly all CD8+ α/β T cells were Ly-6C−CD44lo, and only two γ/δ T cell subsets (Ly-6C−CD44lo and Ly-6C−CD44hi) were detected (Fig. 1E).
γ/δ T cells can be subdivided into four subsets on the basis of Ly-6C and CD44 molecule expression. (A) CD44/Ly-6C dot plots for gated TCRγ/δ+ or TCRβ+ CD8+ T cells recovered from LNs of C57BL/6 mice. (B) CD44/CD27 dot plots for gated TCRγ/δ+ recovered from LNs of C57BL/6 mice. (C–E) Pie charts illustrating the proportions of T cell subsets among TCRγ/δ+ and TCRβ+ CD8+ T cells recovered from LNs and spleen, peritoneal cavity (PC), liver, and lungs, and thymus of C57BL/6 mice. (F) Proportions of Vγ1.1-, Vγ2-, Vγ3-, and Vγ5-expressing cells among TCRγ/δ+ T cell subsets recovered from LNs and spleen of C57BL/6 mice are shown as means ± SEM. ***p < 0.001.
γ/δ T cells can be subdivided into four subsets on the basis of Ly-6C and CD44 molecule expression. (A) CD44/Ly-6C dot plots for gated TCRγ/δ+ or TCRβ+ CD8+ T cells recovered from LNs of C57BL/6 mice. (B) CD44/CD27 dot plots for gated TCRγ/δ+ recovered from LNs of C57BL/6 mice. (C–E) Pie charts illustrating the proportions of T cell subsets among TCRγ/δ+ and TCRβ+ CD8+ T cells recovered from LNs and spleen, peritoneal cavity (PC), liver, and lungs, and thymus of C57BL/6 mice. (F) Proportions of Vγ1.1-, Vγ2-, Vγ3-, and Vγ5-expressing cells among TCRγ/δ+ T cell subsets recovered from LNs and spleen of C57BL/6 mice are shown as means ± SEM. ***p < 0.001.
Narayan et al. (22) previously showed that distinct cell types in the γ/δ T cell lineage could be identified on the basis of their TCR repertoire usage. We then analyzed the Vγ-chain repertoire of the four γ/δ T cell subsets recovered from LNs and spleen (Fig. 1F). Interestingly, a lower Vγ1.1- and a higher Vγ2-expressing cell proportion could distinguish Ly-6C−CD44hi γ/δ T cells from the other three subsets.
Consistent with their CD27− phenotype, Ly-6C−CD44hi γ/δ T cells were the only cells able to produce IL-17 in response to stimulation (Fig. 2A, 2B). By contrast, the other γ/δ T cell subsets produced IFN-γ rather than IL-17, with a pattern similar to that observed in their CD8+ α/β T cell counterparts.
Cytokine production patterns in TCRγ/δ+ and TCRβ+ CD8+ T cell subsets. (A) IL-17/IFN-γ dot plots for gated TCRγ/δ+ or TCRβ+ CD8+ T cells recovered from LNs of C57BL/6 mice. (B) Proportions of IFN-γ– and IL-17–producing cells among TCRγ/δ+ and TCRβ+ CD8+ T cell subsets recovered from LNs of C57BL/6 mice are shown as means ± SEM. ***p < 0.001.
Cytokine production patterns in TCRγ/δ+ and TCRβ+ CD8+ T cell subsets. (A) IL-17/IFN-γ dot plots for gated TCRγ/δ+ or TCRβ+ CD8+ T cells recovered from LNs of C57BL/6 mice. (B) Proportions of IFN-γ– and IL-17–producing cells among TCRγ/δ+ and TCRβ+ CD8+ T cell subsets recovered from LNs of C57BL/6 mice are shown as means ± SEM. ***p < 0.001.
To go further, we then examined the expression of several surface markers by γ/δ T and CD8+ α/β T cells. Once again, Ly-6C−CD44hi γ/δ T cells seemed to stand out from all other γ/δ T cell subsets with high CD25 and low CD5 and CD27 expression levels (Fig. 3A). Similarly to effector/memory Ly-6C+CD44hi CD8+ T cells, we found that Ly-6C+CD44hi γ/δ T cells expressed high surface amounts of CD122, IL-15Rα, Fas ligand, CD137, and OX40 (Fig. 3B). By contrast, all of these molecules were poorly expressed by either Ly-6C− or Ly-6C+CD44lo γ/δ T cells, which exhibit a phenotype comparable to that of naive Ly-6C−/+CD44loCD8+ T cells.
Activation marker expression patterns in TCRγ/δ+ and TCRβ+ CD8+ T cell subsets. (A) CD5, CD25, and CD27 fluorescence histograms of TCRγ/δ+ and TCRβ+ CD8+ T cell subsets recovered from LNs of C57BL/6 mice. (B) CD122, IL-15Rα, Fas ligand (Fas-L), CD137, and OX40 fluorescence histograms of TCRγ/δ+ and TCRβ+ CD8+ T cell subsets recovered from LNs of C57BL/6 mice.
Activation marker expression patterns in TCRγ/δ+ and TCRβ+ CD8+ T cell subsets. (A) CD5, CD25, and CD27 fluorescence histograms of TCRγ/δ+ and TCRβ+ CD8+ T cell subsets recovered from LNs of C57BL/6 mice. (B) CD122, IL-15Rα, Fas ligand (Fas-L), CD137, and OX40 fluorescence histograms of TCRγ/δ+ and TCRβ+ CD8+ T cell subsets recovered from LNs of C57BL/6 mice.
Therefore, phenotypic, Vγ-chain repertoire, cytokine production, and thymic cell distribution analysis suggest that Ly-6C−CD44hi γ/δ T cells could result from a T cell lineage distinct from the three other γ/δ T cell subsets.
Hyporesponsive TCR signaling in CD44hiLy-6C− γ/δ T cells
We then examined the capacity of the aforementioned γ/δ T cell subsets to respond to TCR stimulation by measuring the intracellular calcium increase and compared them to their CD8+ α/β T cell counterparts (Fig. 4A). We observed strong and similar calcium fluxes in both naive Ly-6C− or Ly-6C+CD44loCD8+ α/β T cells and Ly-6C− or Ly-6C+CD44lo γ/δ T cells and a weaker signal in Ly-6C+CD44hi CD8+ α/β and γ/δ T cell subsets. By contrast, no intracellular calcium mobilization was observed in Ly-6C−CD44hi γ/δ T cells (Fig. 4A).
Hyporesponsive TCR signaling in Ly-6C−CD44hi γ/δ T cells. (A) Intracellular Ca2+ mobilization in γ/δ T and TCRβ+ CD8+ T cell subsets from LNs of C57BL/6 mice. (B) Flow cytometry–sorted TCRγ/δ+ T cells from LNs of C57BL/6 mice were stimulated for 16 h in the absence or presence of LPS (100 ng/ml), Pam3CSK4 (1 μg/ml), Poly I:C (5 μg/ml), or CpG (1 μg/ml). Proportions of IL-17–producing Ly-6C−CD44hi γ/δ T cells and IFN-γ–producing Ly-6C+ CD44hi, Ly-6C−CD44lo, and Ly-6C+ CD44lo γ/δ T cells are shown as means ± SEM. *p < 0.05, **p < 0.01.
Hyporesponsive TCR signaling in Ly-6C−CD44hi γ/δ T cells. (A) Intracellular Ca2+ mobilization in γ/δ T and TCRβ+ CD8+ T cell subsets from LNs of C57BL/6 mice. (B) Flow cytometry–sorted TCRγ/δ+ T cells from LNs of C57BL/6 mice were stimulated for 16 h in the absence or presence of LPS (100 ng/ml), Pam3CSK4 (1 μg/ml), Poly I:C (5 μg/ml), or CpG (1 μg/ml). Proportions of IL-17–producing Ly-6C−CD44hi γ/δ T cells and IFN-γ–producing Ly-6C+ CD44hi, Ly-6C−CD44lo, and Ly-6C+ CD44lo γ/δ T cells are shown as means ± SEM. *p < 0.05, **p < 0.01.
We then examined the sensitivity of γ/δ T cell subsets to respond to TCR-independent signals such as TLR-mediated stimulation (Fig. 4B). To do so, FACS-sorted γ/δ T cells were cultured for 16 h in the presence of different TLR ligands (LPS, Pam3CSK4, Poly I:C, or CpG). Cytokine production analysis revealed that the presence of TLR ligands in the culture medium potentiated the IL-17 production capacity of Ly-6C−CD44hi γ/δ T cells. By contrast, IFN-γ production by the other three γ/δ T cell subsets did not appeared to be modulated by the presence of these ligands. Thus, these results suggest a heterogeneity of the peripheral γ/δ T cell compartment comprising: 1) TCR-hyporesponsive innate-like γ/δ T cells (Ly-6C−CD44hi γ/δ T cells); and 2) TCR-responsive γ/δ T cell subsets (Ly-6C+CD44hi, Ly-6C−CD44lo, and Ly-6C+CD44lo γ/δ T cells), sharing phenotypic and functional features with their adaptive effector/memory CD44hi or naive CD44loCD8+ α/β T cell counterparts.
Peripheral homeostasis of γ/δ T cell subsets according to Ly-6C and CD44 expression
We then examined whether we could distinguish various differentiation stages (naive-like or effector/memory-like) within the peripheral γ/δ T cell pool. Effector/memory α/β T cells have a higher proliferation rate than naive T cells (23). As assessed by KI-67 staining (Fig. 5A, 5B), similar to their effector/memory CD8+ α/β T cell counterparts, a significantly greater proportion of Ly-6C+CD44hi γ/δ T cells were cycling compared with Ly-6C− or + CD44lo γ/δ T cells. Of note, although hyporesponsive to in vitro TCR stimulation, Ly-6C−CD44hi γ/δ T cells contained an important proportion of cycling cells, which is in agreement with the work of Haas et al. (24), suggesting that IL-17–producing γ/δ T cells persist in adult mice as self-renewing, long-lived cells.
We then decided to study the impact of ageing in the homeostasis of peripheral γ/δ T and conventional CD8+ α/β T cell subsets. Indeed, during ageing, an enrichment in the proportion of effector/memory T cells to the detriment of naive T lymphocytes can be observed in the periphery as a consequence of thymic involution (25, 26). As depicted in Fig. 5C, the proportion of Ly-6C− or + CD44lo cells within both γ/δ T and CD8+ α/β T cells was significantly decreasing with age. In sharp contrast, the percentage of Ly-6C+CD44hi T lymphocytes among both T cell types was greatly increased in 18-mo-old C57BL/6 mice. Of note, we observed that the proportion of Ly-6C−CD44hi γ/δ T cells recovered from LNs was strongly increasing with age.
CD44lo γ/δ T cells convert to Ly-6C+CD44hi γ/δ T cells in response to lymphopenia
When transferred into lymphopenic environment, naive α/β T cells proliferate strongly and acquire a memory-like phenotype (27, 28). To study the fate of γ/δ T lymphocytes in response to lymphopenia, FACS-sorted γ/δ T cell subsets were transferred separately into C57BL/6 CD3ε−/− mice (Fig. 6A). Absolute numbers of recovered γ/δ T cells were then calculated 2 and 3 mo after transfer (Fig. 6B, 6C). We first observed that, after Ly-6C−CD44hi γ/δ T cell transfer, the absolute numbers of γ/δ T cells recovered from secondary lymphoid organs were extremely low, suggesting survival and/or migration defects of this γ/δ T cell subset in this setting. By contrast, after transfer of the three other γ/δ T cell subsets, the absolute numbers of γ/δ T cells recovered from the spleen and LNs were higher than the number of cells initially injected, suggesting a proliferation and subsequent expansion of these T cell subsets in response to lymphopenia (Fig. 6B). Of note, the absolute numbers of γ/δ T cells recovered from PP and small intestine were relatively low, suggesting that there was no preferential migration of these T cell subsets toward these organs in this setting (Fig. 6C).
CD44lo γ/δ T cells convert to Ly-6C+ CD44hi γ/δ T cells in response to lymphopenia. Total of 5 × 104 flow cytometry–sorted Ly-6C−CD44hi, Ly-6C+ D44hi, Ly-6C−CD44lo, or Ly-6C+ D44lo γ/δ T cells from C57BL/6 CD45.1 mice were injected i.v. into C57BL/6 CD45.2 CD3ε−/− mice. (A) Diagram illustrating the experimental model. (B) Absolute numbers of γ/δ T cells recovered 2 and 3 mo after transfer from pLNs, mLNs, or spleen. Absolute numbers of total γ/δ T cells recovered from pooled LNs (pLNs and mLNs) and spleen are also represented. Horizontal dotted line corresponds to the initial absolute number (5 × 104) of transferred cells. Each symbol represents an individual mouse. (C) Absolute numbers of γ/δ T cells recovered 3 mo after transfer from PP and intestine of recipient mice. Each symbol represents an individual mouse. (D) CD44/Ly-6C dot plots for gated TCRγ/δ+ recovered 2 mo after transfer from pLNs, mLNs, and spleen of representative recipient mice. (E) IL-17/IFN-γ dot plots for gated TCRγ/δ+ recovered 2 mo after transfer from pLNs, mLNs, and spleen of representative recipient mice.
CD44lo γ/δ T cells convert to Ly-6C+ CD44hi γ/δ T cells in response to lymphopenia. Total of 5 × 104 flow cytometry–sorted Ly-6C−CD44hi, Ly-6C+ D44hi, Ly-6C−CD44lo, or Ly-6C+ D44lo γ/δ T cells from C57BL/6 CD45.1 mice were injected i.v. into C57BL/6 CD45.2 CD3ε−/− mice. (A) Diagram illustrating the experimental model. (B) Absolute numbers of γ/δ T cells recovered 2 and 3 mo after transfer from pLNs, mLNs, or spleen. Absolute numbers of total γ/δ T cells recovered from pooled LNs (pLNs and mLNs) and spleen are also represented. Horizontal dotted line corresponds to the initial absolute number (5 × 104) of transferred cells. Each symbol represents an individual mouse. (C) Absolute numbers of γ/δ T cells recovered 3 mo after transfer from PP and intestine of recipient mice. Each symbol represents an individual mouse. (D) CD44/Ly-6C dot plots for gated TCRγ/δ+ recovered 2 mo after transfer from pLNs, mLNs, and spleen of representative recipient mice. (E) IL-17/IFN-γ dot plots for gated TCRγ/δ+ recovered 2 mo after transfer from pLNs, mLNs, and spleen of representative recipient mice.
Ly-6C and CD44 expression on recovered γ/δ T cells was then analyzed (Fig. 6D). First, we noticed that the phenotype of Ly-6C− or Ly-6C+CD44hi cells was stable over time. In addition, 2 mo after transfer, initially injected Ly-6C−CD44hi γ/δ T cells were still producing IL-17, whereas Ly-6C+CD44hi γ/δ T cells produced IFN-γ (Fig. 6E). By contrast, such a phenotypic stability was not observed for Ly-6C− or + CD44lo γ/δ T cells. Indeed, all γ/δ T cells recovered 2 mo after the initial injection of Ly-6C− or Ly-6C+CD44lo cells were expressing high surface amounts of both Ly-6C and CD44 (Fig. 6D) and gave rise to a proportion of IFN-γ–producing cells, comparable to what can be observed after Ly-6C+CD44hi γ/δ T cell transfer (Fig. 6E).
Thus, these results indicate that, as observed for naive α/β T cells, CD44lo γ/δ T cells convert to memory-like CD44hi cells in response to lymphopenia, suggesting that these three γ/δ T cell subsets (Ly-6C− or + CD44lo and Ly-6C+CD44hi) would actually represent various differentiation stages of a single T cell lineage.
CD44lo γ/δ T cells are highly plastic in vitro
Following activation by APCs in the periphery, naive α/β T cells can differentiate into a variety of well-documented Th cell subsets, such as Th1, Th2, Th17, or induced Tregs (iTreg), characterized by their cytokine production profiles and specific effector functions (29). Thus, we assessed whether this characteristic feature could also be observed for γ/δ T cells and first compared the capacity of the different γ/δ T cell subsets to polarize into Th17-like (IL-17–producing) and Th1-like (IFN-γ–producing) cells in vitro (Fig. 7). FACS-sorted γ/δ T cell subsets were stimulated with anti-CD3– and anti-CD28–coated Abs in the presence of either IL-6, IL-23, and graded doses of TGF-β (Th17-like conditions, Fig. 7A) or IL-2 and graded doses of IL-12 (Th1-like conditions, Fig. 7B). In these polarization conditions, Ly-6C−CD44hi and Ly-6C+CD44hi γ/δ T cell subsets mainly retained their natural effector cell profile, by producing IL-17 or IFN-γ respectively, suggesting that both CD44hi γ/δ T cell subsets would be at an end stage of differentiation. By contrast, Ly-6C− and Ly-6C+CD44lo γ/δ T cells were able to differentiate into both IL-17– and IFN-γ–producing cells according to polarization assays. Of note, under Th17-like polarization conditions, part of Ly-6C− and Ly-6C+CD44lo γ/δ T cells acquired the expression of the iTreg lineage-defining transcription factor Foxp3.
CD44lo γ/δ T cells are highly plastic in vitro. Flow cytometry–sorted Ly6C−CD44hi, Ly6C+CD44hi, Ly6C−CD44lo, and Ly6C+ D44lo γ/δ T cells from LNs of C57BL/6 Foxp3-GFP mice were stimulated for 4 d with coated anti-CD3 and anti-CD28 in the presence of IL-6 (20 ng/ml), IL-23 (10 ng/ml), and graded concentrations of TGF-β1 (A) or in the presence of graded concentrations of IL-12 (B) or TGF-β1 (C). Proportions of IL-17–, IFN-γ–producing γ/δ T cells and the proportion of Foxp3+ cells among γ/δ T cells are shown as a function of TGF-β1 concentration (A and C) or as a function of IL-17 concentration (B). (D) FACS-sorted Foxp3−CD44lo TCRγ/δ+ T cells or Foxp3− D25−CD44lo naive CD4 T cells from LNs of C57BL/6 CD45.1 Foxp3-GFP mice were stimulated for 3 d with coated anti-CD3 and anti-CD28 Abs in the presence of 1 ng/ml of TGF-β1 and then let to rest for 3 more d in the presence of IL-2 (13 ng/ml). GFP-expressing cells were then flow cytometry sorted and their suppressive abilities to inhibit the proliferation of Cell Trace Violet (CTV)–labeled conventional CD4 T cells (LN GFP− CD4 T cells from C57BL/6 CD45.2 Foxp3-GFP mice) in response to anti-CD3 Abs in the presence of APCs were tested at various Treg/conventional T cell (Tconv) ratios.
CD44lo γ/δ T cells are highly plastic in vitro. Flow cytometry–sorted Ly6C−CD44hi, Ly6C+CD44hi, Ly6C−CD44lo, and Ly6C+ D44lo γ/δ T cells from LNs of C57BL/6 Foxp3-GFP mice were stimulated for 4 d with coated anti-CD3 and anti-CD28 in the presence of IL-6 (20 ng/ml), IL-23 (10 ng/ml), and graded concentrations of TGF-β1 (A) or in the presence of graded concentrations of IL-12 (B) or TGF-β1 (C). Proportions of IL-17–, IFN-γ–producing γ/δ T cells and the proportion of Foxp3+ cells among γ/δ T cells are shown as a function of TGF-β1 concentration (A and C) or as a function of IL-17 concentration (B). (D) FACS-sorted Foxp3−CD44lo TCRγ/δ+ T cells or Foxp3− D25−CD44lo naive CD4 T cells from LNs of C57BL/6 CD45.1 Foxp3-GFP mice were stimulated for 3 d with coated anti-CD3 and anti-CD28 Abs in the presence of 1 ng/ml of TGF-β1 and then let to rest for 3 more d in the presence of IL-2 (13 ng/ml). GFP-expressing cells were then flow cytometry sorted and their suppressive abilities to inhibit the proliferation of Cell Trace Violet (CTV)–labeled conventional CD4 T cells (LN GFP− CD4 T cells from C57BL/6 CD45.2 Foxp3-GFP mice) in response to anti-CD3 Abs in the presence of APCs were tested at various Treg/conventional T cell (Tconv) ratios.
Therefore, we then compared the capacity of γ/δ T cell subsets to differentiate into iTreg-like cells in vitro (Fig. 7C). In contrast to CD44hi γ/δ T cell subsets, low doses of TGF-β were sufficient to induce the expression of Foxp3 by the vast majority of Ly-6C− or Ly-6C+CD44lo γ/δ T cells (Fig. 7C). Furthermore, we observed that Foxp3 expression in these cells was correlated with suppressive capacities, as induced Foxp3+ γ/δ T cells displayed similar suppressive capacities as CD4+ iTregs (Fig. 7D). Taken together, our data strongly suggest that, as described for naive α/β T cells, CD44lo γ/δ T cells have an intrinsic higher plasticity, illustrated by their capacity to differentiate into Th17-like, Th1-like, or Foxp3+ iTreg-like γ/δ T cells, than their CD44hi γ/δ T cell counterparts.
Discussion
In the last decade, several teams have used the expression level of the CD27 coreceptor to subdivide peripheral γ/δ T cells (11–14). In the present paper, we found that peripheral γ/δ T cells can be subdivided into four subsets, corresponding to two distinct functional γ/δ T cell lineages (i.e., Ly-6C−CD44hi versus Ly-6C− or + CD44lo and Ly-6C+CD44hi γ/δ T cells) according to CD44 and Ly-6C expression. We first noticed that Ly-6C−CD44hi γ/δ T cells were corresponding to the CD27− γ/δ T cell subset previously described by others (11, 12). Consistent with their CD27− phenotype, CD44hiLy-6C− γ/δ T cells seemed to stand out from all other γ/δ T cell subsets through their high constitutive expression of CD25 and their capacity to produce IL-17 (9) but also by their ability to proliferate strongly in secondary lymphoid organs in the steady state, a property that is in agreement with previous results suggesting that IL-17–producing γ/δ T cells persist in adult mice as self-renewing, long-lived cells (24).
Interestingly, Cai et al. (30) reported that dermal γ/δ T cells, which are capable of producing IL-17, were proliferating strongly few days after birth but exhibited a far lower proliferation rate (3%) in adult mice compared with the important proportion of cycling CD44hiLy-6C− γ/δ T cells that we observed in the secondary lymphoid organs of young adult mice. This apparent discrepancy could reflect an impact of γ/δ T cell localization (lymphoid organs versus tissues) on their functional and homeostatic properties. In addition, Shibata et al. (9) previously examined age-related changes in the frequency of IL-17–producing γ/δ T cells recovered from the peritoneal cavity. They observed that the proportion of this γ/δ T cell subset was decreasing from 1 wk after birth to adult age (17-wk-old mice). In the present paper, we have studied age-related changes in the frequency of Ly-6C−CD44hi γ/δ T cells recovered from the LNs of young (6–8 wk old), adult (8–10 mo old), or elderly mice (18 mo old) and observed that the proportion of this γ/δ T cell subset strongly increased in elderly mice when compared with 6/8-wk- or 8/10-mo-old mice. Thus, altogether these data suggest that Ly-6C−CD44hi γ/δ T cell homeostasis could vary according to cell localization and ageing.
Our study emphasized subdivision in the γ/δ T cell lineage as analysis of Vγ-chain repertoire, cytokine production, and cell distribution at the thymic level suggest that Ly-6C−CD44hi γ/δ T cells would result from a distinctive T cell lineage when compared with the other γ/δ T cells. Functionally, as previously shown by Wencker et al. (12) for CD27− γ/δ T cells, Ly-6C−CD44hi γ/δ T cells appear as hyporesponsive to TCR stimulation. In addition, they express low levels of CD5, which is in line with the fact that they will perceive weak TCR signals in the steady state (16, 31).
The activation of adaptive lymphocytes depends on Ag recognition, engagement of costimulatory ligands, and the presence of cytokines. By contrast, innate lymphocytes are able to respond to innate signals alone. For instance, it has been shown that some γ/δ T cells can produce IL-17 in a TCR-independent manner in response to IL-1β plus IL-23 (32) and TLR2 or Dectin-1 ligands (33). Moreover, it has also been observed that innate TLR/MyD88-dependent signals selectively expand IL-17–producing CD27- γ/δ T cells in vivo (13). In this study, we observed that the IL-17 production capacity of Ly-6C−CD44hi γ/δ T cells was potentiated in the presence of various TLR ligands, strongly supporting the innate-like hallmark of this γ/δ T cell subset. This result seems contradictory to the work from Ribot et al. (13) as, in their study, the authors did not observe a cis activation of CD27− γ/δ T cells by TLR agonists in vitro. This discrepancy could reside in the difference of culture duration between their study and ours (2 d versus 16 h). Indeed, the increased production of IL-17 by Ly-6C−CD44hi γ/δ T cells in response to TLR agonists could only be transitory and not detectable after an extended culture time.
Unlike CD27− γ/δ T cells, the CD27+ γ/δ T cell compartment can be subdivided into three different subsets on the basis of Ly-6C and CD44 expression. Consistent with their CD27+ phenotype, these γ/δ T cell subsets appeared skewed toward IFN-γ production ex vivo. More interestingly, by comparing these subsets to naive and memory CD8+ α/β T cells, in this study, we reveal important similarities between naive Ly-6C− or Ly-6C+CD44lo CD8+ α/β T cells and Ly-6C− and Ly-6C+CD44lo γ/δ T cells as well as between memory Ly-6C+CD44hi CD8+ α/β T cells and Ly-6C+CD44hi γ/δ T cells. Indeed, with respect to the expression of several phenotypic markers, Ly-6C+CD44hi γ/δ T cells look like memory Ly-6C+CD44hi CD8+ α/β T cells. Like memory CD8+ α/β T cells, they are sparse in the thymus but largely increased in proportion in tissues. Furthermore, similar to CD8+ α/β T cells, we found that within the γ/δ T cell compartment, Ly-6C+CD44hi cells mobilized calcium less efficiently than Ly-6C− or + CD44lo cells upon in vitro stimulation but divided more in vivo in the steady state. In both the α/β and γ/δ T cell compartments, the proportion of CD44lo cells (naive/naive-like) decreased with age in secondary lymphoid organs, a process that may derive in both cases from thymic involution (25, 26). Finally, as observed for naive α/β T cells (34), Ly-6C− or + CD44lo γ/δ T cells exhibit a high intrinsic plasticity in vitro, illustrated by their capacity to differentiate efficiently into Th17-like, Th1-like or, more strikingly, into functional Foxp3+ iTreg-like γ/δ T cells. Altogether, these data strongly suggest that the peripheral CD27+ γ/δ T cell compartment comprises both naive-like and memory-like adaptive cells. Although Ag-specific memory γ/δ T cell responses following immunization or infection in mice and humans still need to be better characterized, our results appear consistent with previous studies describing γ/δ T cell memory-type responses in mice after Staphylococcus aureus infection (35), in macaques following a secondary challenge with bacillus Calmette-Guerin (36), or in humans in a CMV infection context (37).
When transferred into a lymphopenic environment, naive α/β T cells proliferate strongly and acquire a memory-like phenotype (20, 27, 38). In our study, we observed that naive-like Ly-6C− and Ly-6C+CD44lo γ/δ T cell subsets were able to undergo lymphopenia-induced spontaneous proliferation and to convert to a memory-like phenotype (Ly-6C+CD44hi) while increasing their capacity to produce IFN-γ, suggesting that these three γ/δ T cell subsets would actually correspond to various differentiation stages of a unique T cell lineage. It is well established that interactions with self-peptides or commensal bacterium-derived peptides presented by MHC molecules are required for the lymphopenia-induced spontaneous proliferation of α/β T cells (28, 39). By contrast, the homeostatic resources such as the cytokines [IL-7 and IL-15 (40)] or the TCR ligands driving this process for γ/δ T cells still need to be addressed.
We recently showed that peripheral Tregs and naive CD4+ α/β T cells can be subdivided into two subsets according to Ly-6C expression and that these newly defined subsets were functionally not equal (34, 41). In this study, we noticed that, in both the α/β CD8+ T cell and γ/δ T cell compartments, CD44loLy-6C+ cells exhibit a greater ability to produce IFN-γ than their Ly-6C− cell counterparts ex vivo. One can thus wonder whether, as observed for naive and regulatory CD4+ α/β T cells (34, 41), Ly-6C expression or nonexpression within the naive α/β CD8+ T cell and the naive-like γ/δ T cell compartments could reflect distinct functional features.
γ/δ T cells are crucially involved in host immune defense against infections (42) but are also known to have a strong clinical association with various autoimmune diseases such as inflammatory bowel disease (43, 44), rheumatoid arthritis (or collagen-induced arthritis, the murine model of rheumatoid arthritis) (45, 46), and multiple sclerosis (or experimental autoimmune encephalomyelitis, the murine model of multiple sclerosis) (32, 47). In addition, there is compelling evidence indicating that γ/δ T cells play an important role in immunity to cancer by sensing and reacting to cellular stress. This has been clearly demonstrated in murine models of spontaneous (48), chemically induced (49), transgenic (50), and transplantable tumors (51, 52). However, it seems that the activity of γ/δ T cells in response to tumors can differ radically according to tumor types or tumor environments (53, 54). This may reflect the high diversity of the γ/δ T cell compartment highlighted in this study.
γ/δ T cells hold promise for adoptive immunotherapy because of their reactivity to bacteria, viruses, and tumors. Although these cells represent a small fraction (1–5%) of the peripheral T cell pool, various methodologies (such as aminobiphosphonates and synthetic phosphoantigens to expand human Vγ9Vδ2 T cells or immobilized Ag, Abs, or artificial APCs to grow Vδ1 or other non-Vγ9Vδ2 T cells) have been developed to expand these cells ex vivo to achieve clinical benefit (55). These techniques have been transitioned to the clinic for investigational treatments of cancer (56, 57). It has been observed that adoptive transfer and in vivo expansions of different γ/δ T cell subsets (Vγ9Vδ2, Vδ1, or other non-Vγ9Vδ2 T cells) are safe therapeutic modalities and can result in objective clinical responses in the treatment of cancer such as renal cell carcinoma (58), colorectal cancer, and melanoma (59) or leukemia (60).
Future studies aimed at deciphering the diversity of the γ/δ T cell compartment and the molecular mechanisms that control its functional plasticity will be of major interest and would pave the way for the development of novel therapeutics required for γ/δ T cell–based immunotherapy.
Acknowledgements
We thank Dr. Pablo Pereira for providing anti-Vγ5 Ab. We also thank the Cochin Cytometry and Immunobiology and Animal Core facilities and, in particular, C. Lapert and O. Le Gall for animal care.
Footnotes
This work was supported by grants from the Association pour la Recherche contre le Cancer and the Initiative for Excellence “Université Sorbonne Paris Cité.” A.L. was supported by a master fellowship from “Ecole de l’Inserm Liliane Bettencourt.” M.R. was supported by a fellowship from Fondation pour la Recherche Médical.
References
Disclosures
The authors have no financial conflicts of interest.