TCR signal strength instructs αβ versus γδ lineage decision in immature T cells. Increased signal strength of γδTCR with respect to pre-TCR results in induction of the γδ differentiation program. Extracellular ATP evokes physiological responses through purinergic P2 receptors expressed in the plasma membrane of virtually all cell types. In peripheral T cells, ATP released upon TCR stimulation enhances MAPK activation through P2X receptors. We investigated whether extracellular ATP and P2X receptors signaling tuned TCR signaling at the αβ/γδ lineage bifurcation checkpoint. We show that P2X7 expression was selectively increased in immature γδ+CD25+ cells. These cells were much more competent to release ATP than pre–TCR-expressing cells following TCR stimulation and Ca2+ influx. Genetic ablation as well as pharmacological antagonism of P2X7 resulted in impaired ERK phosphorylation, reduction of early growth response (Egr) transcripts induction, and diversion of γδTCR-expressing thymocytes toward the αβ lineage fate. The impairment of the ERK-Egr-inhibitor of differentiation 3 (Id3) signaling pathway in γδ cells from p2rx7−/− mice resulted in increased representation of the Id3-independent NK1.1-expressing γδ T cell subset in the periphery. Our results indicate that ATP release and P2X7 signaling upon γδTCR expression in immature thymocytes constitutes an important costimulus in T cell lineage choice through the ERK-Egr-Id3 signaling pathway and contributes to shaping the peripheral γδ T cell compartment.
Productive rearrangement of gene segments at Tcrβ, Tcrγ, and Tcrδ loci is instrumental in T cell development in the thymus. The protein products of these stochastic recombination events, the pre-TCR (constituted by the rearranged TCRβ-chain in covalent association with the invariant pre–TCRα-chain) (1) and γδTCR, influence the thymocyte to commit to the αβ and γδ lineage, respectively (2). Whereas differentiation of γδ T cells is generally characterized by lack of expression of CD4 and CD8 coreceptor molecules (3, 4), development of αβ T cells is characterized by the transition of thymocytes through an ordered sequence of phenotypes defined by the expression of CD4 and CD8. Cells progress from the most immature CD4−8− double-negative (DN) stage to the mature either CD4+8− or CD4−8+ single positive (SP) stage through an intermediate CD4+8+ double-positive (DP) stage. The DN stage can be further subdivided according to the expression of CD25 and CD44 with maturation characterized phenotypically by the following sequence: CD44+25− (DN1) > CD44+25+ (DN2) > CD44−25+ (DN3) > CD44−25− (DN4) (5). In particular, the DN3 stage is a fundamental checkpoint at which pre-TCR and γδ TCR signals instruct commitment to αβ and γδ lineage, respectively (6). β-Selection is associated with higher CD5 and CD27 expression and increase in cell size, which define the DN3b stage (7, 8). Two elegant studies have shown that TCR signal strength rather than TCR signal quality determines lineage choice, a strong TCR signal resulting in γδ and a weak signal in αβ commitment. In fact, attenuation of γδTCR signal strength or expression level induced differentiation of γδTCR-expressing thymocytes along the αβ pathway to the DP stage (9, 10). Whether concomitant costimuli contribute to TCR signal strength in γδ lineage choice is at present unknown. In addition, differential signaling by the γδTCR can result in development of γδ NKT cells, which express NK.1.1 and, in contrast with conventional γδ cells, are relatively independent from the ERK-early growth response (Egr)-inhibitor of differentiation 3 (Id3) signaling axis.
We have previously shown that signal transduction by the TCR in peripheral CD4 T cells determines the increase in ATP synthesis and the release through calcium-sensitive pannexin-1 hemichannels, which in turn activates in an autocrine fashion purinergic P2X receptors (ATP-gated nonselective cationic channels) in the plasma membrane. This signaling loop acts as a costimulus for MAPK signaling and implements cell cycling as well as IL-2 secretion (11, 12). Pioneering experiments have shown that immature thymocytes are responsive to extracellular ATP through P2X receptors (13). Rearrangement competent but not incompetent DN3 thymocytes undergo spontaneous cytosolic Ca2+ oscillations, which might lead to ATP release. We tested whether ATP released as a result of pre-TCR and/or γδTCR expression affected thymocyte differentiation through P2X receptors signaling. We show that P2X7 receptor activation contributes to γδTCR signal strength and that genetic ablation as well as pharmacological antagonism of P2X7 diverts γδTCR-expressing cells toward the αβ fate. Moreover, impaired γδTCR signaling in p2rx7−/− mice resulted in increased representation of Id3-independent NK1.1+TCR Vδ6.3+ cells in the periphery.
Materials and Methods
C57BL/6 (wild-type [WT]) mice were obtained from Charles River Laboratories Germany and p2rx7−/− (14), and rag1−/− C57BL/6 mice were obtained from The Jackson Laboratory. Mice were used at 4 wk to isolate thymocytes and at 8 wk to analyze γδ cells in peripheral tissues. The animals were bred and treated in accordance with the Swiss Federal Veterinary Office guidelines and were kept in specific pathogen-free animal facility.
Flow cytometry and cell culture
Thymi were homogenized and washed in RPMI 1640 medium containing 10% (v/v) FBS. In some cases, thymocyte samples were depleted of CD8+ and CD4+ cells by treatment with anti-CD8 (31.M) and anti-CD4 (RL172) mAbs with low-tox-M Rabbit Complement (Cederlane Laboratories). TCRδ− DN3a thymocytes were sorted as CD44−CD25+CD27low and TCRδ− DN3b thymocytes as CD44−CD25+CD27high. Thymocytes subsets were sorted with a FACSAria (BD Biosciences).
Sorted γδ+CD25+cells were plated onto subconfluent OP9 BM stromal cells transduced with the Notch ligand Delta-like 1 (OP9-DL1) monolayers previously subjected to gamma irradiation (40 Gy) at 5 × 104 cells/well in a 96-well plate. All cocultures were performed in the presence of 1 ng/ml IL-7 and 5 ng/ml Flt3 ligand (PeproTech) in α-MEM containing 20% (v/v) FBS. When indicated, plates were coated with mAb specific for TCRδ (clone 3A10; final concentration, 10 μg/ml) before OP9 cells were plated. OP9-DL1 cells were maintained as described previously (15
For analysis of the extent of phosphorylation of ERK, sorted γδ+CD25+ cells (5 × 104 cells/well) were stimulated for 16 h with the indicated stimuli, permeabilized, and incubated with rabbit mAbs against phosphorylated ERK (Thr202/Tyr204) (D13.14.4E) (Cell Signaling Technology).
p2rx1 (Mm00435460_m1), p2rx2 (Mm01202369_g1), p2rx3 (Mm00523699_m1), p2rx4 (Mm00501787_m1), p2rx5 (Mm00473677_m1), p2rx7 (Mm00440582_m1), Egr1 (Mm00656724_m1), Egr2 (Mm00456650_m1), Egr3 (Mm00516979_m1), and Rn18s (EUK 18S rRNA [DQ] Mix). For each sample, mRNA abundance was normalized to that of 18S rRNA and is presented in arbitrary units.
Ca2+ imaging and measurement of ATP release
Sorted thymocytes were loaded for 30 min at room temperature with 5 μM Fura-2 pentacetoxy-methylester in α-MEM, washed in the same solution, and plated on a monolayer of OP9-DL1 or OP9-GFP cells previously seeded on poly-l-lysine–coated coverslips. Coverslips were then washed and transferred to the recording chamber of an inverted microscope (Axiovert 200; Zeiss) equipped with an Andor 885 JCS iXON classic camera. For Ca2+ measurements, Polychrome V (Till Photonics) was used as a light source. After excitation at 340 and 380 nm, the emitted light was acquired at 505 nm. The sampling rate was 1 Hz. Ca2+ concentrations are expressed as the 340/380 fluorescence ratio. The experiments were performed in a static modified Krebs–Ringer solution (155 mM NaCl, 4.5 mM KCl, 10 mM glucose, 5 mM Hepes [pH 7.4], 1 mM MgCl2, and 2 mM CaCl2) at 28–30°C. To inhibit TCR signaling in DN3 thymocytes, the src-like kinase inhibitor PP2 (VWR International) was used at 10 μM.
ATP release was measured by means of a two-enzyme assay, as described previously (16). Briefly, cells were plated on poly-L-lysine–coated coverslips, which were fixed in the recording chamber of an inverted microscope. Cells were incubated in modified Krebs–Ringer solution supplemented with 8 U/ml each of hexokinase and glucose 6-phosphate dehydrogenase and 5 mM NADP. In the presence of ATP, these enzymes catalyze the formation of NADPH, a fluorescent molecule that was visualized using fluorescence microscopy with an excitation wavelength of 340 nm and an emission at 460 nm. The fluorescence intensity of regions on or outside the cells was measured using TILLvisION software, and ATP concentrations were estimated using ATP standard solutions of known concentrations.
For measurement of cellular ATP, cells were lysed in 1% Triton X-100 and frozen on dry ice until analyzed with the ATP determination kit (Molecular Probes).
Fetal thymus organ culture
The medium for fetal thymus organ culture (FTOC) was IMDM plus GlutaMAX medium with 20% (v/v) FBS, 50 μM 2-ME, 1 mM sodium pyruvate, penicillin, and streptomycin. WT or p2rx7−/− fetal thymic lobes at E14 were cultured for various times in FTOC medium on isopore membrane filters (Millipore) and then were analyzed by flow cytometry. Cultures were provided fresh medium every 2 d.
Statistical analysis was performed with the Student t test. Data are reported as means ± SEM or SD. The p values <0.05 were considered significant (*p < 0.05, **p < 0.01, ***p < 0.001).
TCR-dependent Ca2+ oscillations, ATP synthesis, and release in immature thymocytes
We labeled sorted DN3 thymocytes from WT and rag1−/− C57BL/6 mice with Fura-2 and analyzed them for cytosolic Ca2+ variations in live imaging. Because Notch signaling is required for survival of DN3 cells (17) and pre–TCR-driven transition to the DP stage (18, 19), we cocultured DN3 cells with OP9-DL1 (15). Thymocytes from WT mice displayed spontaneous Ca2+ oscillations. In contrast, recombinase deficiency resulted in complete absence of spontaneous calcium waves, indicating that expression of pre-TCR and/or γδTCR was required for cytosolic Ca2+ elevations (Fig. 1A). Calcium spikes detected in WT thymocytes were also observed by coculturing WT cells with nontransduced OP9 cells. No significant differences were observed both in percentage of responding cells and in number of Ca2+ elevations per cell between thymocytes cocultured with OP9-DL1 and DL1-null OP9 cells, thereby suggesting that calcium oscillations were not dependent on Notch activation (Supplemental Fig. 1A, 1B). Because γδTCR signaling in immature γδ thymocytes and progression along the γδ developmental pathway are independent from Notch signaling relative to the Notch dependent pre-TCR signaling, these results suggest that γδTCR induces cytosolic Ca2+ elevations in immature thymocytes. Accordingly, the percentage of cells displaying Ca2+ elevations was dramatically reduced by treatment with PP2, which inhibits src-like (e.g., lck/fyn) tyrosine kinase activity, suggesting that increases in cytosolic Ca2+ were dependent on most proximal signaling by γδTCR and possibly pre-TCR (Fig. 1B).
We have previously shown that cytosolic Ca2+ increases in T cells determines ATP release through Ca2+-sensitive pannexin-1 hemichannels (11). We compared ATP release in TCRδ−CD27+ DN3b cells (>97% TCRβ+) and CD25+TCRδ+ cells upon exposure to the ionophore ionomycin to provoke massive Ca2+-sensitive release of ATP. This treatment resulted in robust ATP release by γδ but not pre–TCR-expressing DN3b cells and CD27− DN3a cells (Fig. 2A), thereby suggesting that expression of the γδTCR induces Ca2+ dependent enhanced ATP release in the developing T cell. To see whether pre-TCR and γδTCR signaling differently affected ATP levels in immature DN thymocytes, we measured intracellular ATP increases upon stimulation with anti-CD3 mAb in CD25+TCRδ+, DN3a, and DN3b cells. Fig. 2B shows that γδTCR signaling in CD25+ cells induced significantly increased intracellular ATP levels with respect to pre-TCR. Moreover, ATP release could be readily measured upon stimulation with anti-CD3 mAb-coated beads in CD25+TCRδ+ (Fig. 2C) but not DN3a and DN3b thymocytes (Supplemental Fig. 1C) by a two-enzyme assay in live-cell imaging experiments (16). Altogether, these results indicate that γδTCR in immature thymocyte is a more potent inducer of ATP synthesis and release than the pre-TCR, and is responsible of enhanced purinergic signaling in the γδTCR expressing cells.
Expression of purinergic P2X receptors during T cell development
Analysis of P2X transcripts by real-time PCR at different stages of T cell development in the thymus revealed the higher expression of p2rx4 and p2rx7 genes at the DN and CD4 SP stages (Fig. 3A). We did not test p2rx6 because its transcript was not detectable in any thymocyte subset (data not shown). To get insight into transcriptional regulation of p2rx4 and p2rx7 genes in the DN compartment during αβ and γδ lineage commitment, we sorted DN3a and DN3b cells as well as immature CD25+ γδTCR cells and analyzed transcription of p2rx genes by real-time PCR. This experiment revealed the selective expression of p2rx7 in γδ+CD25+ cells (Fig. 3B).
P2X7-mediated regulation of γδ T cell lineage commitment
To see whether P2X7 contributed to lineage choice of γδTCR-expressing immature thymocytes, we performed FTOC with E14 thymi from p2rx7−/− embryos. Analysis of thymocyte subsets for CD4 and CD8 expression revealed no differences in the distribution of thymocytes as DN, DP, and SP cells (Fig. 4A) as well as DN1 to four subsets defined by CD44 and CD25 expression (Supplemental Fig. 1D). However, p2rx7−/− FTOC showed the significant increase of γδTCR-expressing DP cells both as percentage and absolute number (Fig. 4B), suggesting that lack of P2X7 signaling caused the aberrant transition of γδTCR-expressing cells to an αβ-lineage like DP stage. Then, we performed FTOC with WT E14 thymi in the presence of oATP, a pharmacological P2X antagonist (20). Also, in the presence of oATP DP cells expressing TCRδ chains were significantly increased (Fig. 4C), suggesting aberrant αβ commitment of γδTCR-expressing cells by pharmacological P2X antagonism. To more specifically address the contribution of P2X7 in γδ commitment, we performed FTOC from both C57BL/6 and BALB/c embryos in the presence of the selective P2X7 antagonist A438079. These experiments revealed the significant decrease of γδ DN cells (p < 0.001) as well as the significant increase of γδ DP cells (p < 0.001) in both strains (Supplemental Fig. 2). In addition, we addressed the effect of oATP on γδ cell differentiation by coculturing γδ+CD25+ cells with OP9-DL1 cells in the presence of γδTCR mAb, which implements the differentiation of CD25+ γδTCR-bearing cells to the DN mature stage (6). These experiments confirmed the aberrant transition of γδ+ cells to the DP stage in the presence of oATP in a dose-dependent fashion (Fig. 4C). Finally, to show that P2X7 signaling contributes to γδ commitment, we performed FTOC in the presence of the selective P2X7 agonist benzoyl-ATP (BzATP) (21). Whereas BzATP did not influence the development of TCRβ-expressing cells (Supplemental Fig. 1E), it determined the significant increase of γδ cells in the DN compartment (Fig. 4D), suggesting that P2X7 stimulation promoted the acquisition of the mature DN phenotype by γδTCR-bearing cells. A polymorphism in the cytoplasmic domain of P2X7, which expresses either a proline (found in BALB/c mice) or leucine (present in C57BL/6 mice) at position 451 confers high and low sensitivity to stimulation, respectively (22). Consistent with increased P2X7 activity in BALB/c mice, 6-d FTOC from E14 embryos revealed the significant increase in γδ cells development in BALB/c versus C57BL/6 cultures (Fig. 5A).
Extracellular ATP and P2X7 signaling contribute to TCR signal strength in γδ lineage commitment
P2X7 expression in cells otherwise devoid of P2X7 increases intracellular ATP content and supports cell growth through P2X7 stimulation by secreted ATP (23). Analysis of intracellular ATP levels following CD3 stimulation of γδ thymocytes showed lack of ATP increase in p2rx7−/− with respect to WT cells (Fig. 6A). We have previously shown that reduced ATP levels in p2rx7−/− regulatory T cells correlated with reduced ERK phosphorylation compared with the WT counterpart (24). Attenuation of ERK phosphorylation was shown to divert γδ thymocytes to the αβ lineage (9, 10). Then, we addressed whether P2X7 expression affected ERK phosphorylation in γδ+CD25+ thymocytes cocultured with OP9-DL1 cells. P2rx7−/− γδ cells displayed impaired phospho-ERK and cell size increases after 16-h stimulation with TCRδ mAb. These effects were also observed in WT γδ thymocytes upon addition of oATP (Fig. 6B). Altogether, our results point to a role of P2X7 in tuning TCR signal strength in γδ lineage commitment. Accordingly, egr1, egr2, and egr3 transcripts, which are controlled by γδTCR signal strength (9), were diminished in ex vivo-purified p2rx7−/−γδ+CD25+ thymocytes with respect to WT cells (Fig. 6C).
Impact of P2X7 activity on peripheral γδ T cells
We investigated whether p2rx7 deletion affected the γδ cells peripheral pool. We did not detect significant differences in total γδ cells in peripheral lymph nodes, uterus, and lungs from 8-wk-old p2rx7−/− and WT mice. However, γδ cells were significantly decreased in mesenteric lymph nodes from p2rx7−/− mice (Supplemental Fig. 3). Peripheral γδ cells are heterogeneous and can be distinguished by expression of TCRs using distinct TCRγ and TCRδ variable (V) regions. For example, expression of TCRVδ4 characterizes principally cells in the intestinal epithelium, spleen, thymus, and lactating mammary glands. This cell subset was not differently represented in the thymus, spleen, and mesenteric lymph nodes from p2rx7−/− and WT mice (data not shown). The defect of ERK signaling we observed in developing p2rx7−/− γδ cells should affect the ERK-Egr-Id3 signaling axis, which is required for the development of cells expressing TCR Vγ4 and Vγ5 (according to Heilig and Tonegawa nomenclature; Refs. 25 and 26). Cells expressing TCR Vγ4 were not significantly different in the spleen and lungs; however, they were significantly reduced in mesenteric lymph nodes of 8-wk-old p2rx7−/− mice, and cells expressing TCR Vγ5 were dramatically reduced in the uterus of p2rx7−/− females (Supplemental Fig. 4). Consistent with increased P2X7 activity in BALB/c mice, we detected a significant increase of γδ cells expressing TCR Vγ4 in BALB/c lymph nodes with respect to C57BL/6 mice (Fig. 5B). These results suggest that signaling by P2X7 substantially contributes to the development of the γδ T cells peripheral pool that depends on the ERK-Egr-Id3 axis.
Lack of Id3 provokes an outgrowth of cells expressing TCR Vγ1 and Vδ6.3 together with the NK cells marker NK1.1, referred to as γδ NKT cells (27). Consistent with an impairment of the ERK-Egr-Id3 pathway in developing γδ cells from p2rx7−/− mice, we detected a significant increase in NK1.1+ cells expressing Vδ6.3/2 in the uterus (Fig. 7) and Peyer’s patches albeit nonsignificant (data not shown) of 8-wk-old p2rx7−/− mice. Altogether, these observations point to a role of P2X7 activity in contributing to lineage choice in γδ T cell development and shaping of the peripheral pool of γδTCR-expressing cells.
In 1997, Ross et al. (13) have shown that immature thymocytes are responsive to P2X7 agonists, and they hypothesized that extracellular ATP could act as a signal for growth and differentiation of DN thymocytes. DN3 thymocytes, which include γδ and αβ committed cells, showed Notch-independent but lck/fyn-dependent (e.g., TCR dependent) spontaneous Ca2+ oscillations. Cytosolic Ca2+ elevations in peripheral T cells determine ATP release through pannexin-1 hemichannels (11). Notably, we could measure ATP release in CD25+ γδ but not pre–TCR-expressing DN3b cells upon exposure to the ionophore ionomycin. In addition, significant increases in intracellular ATP were observed in γδ versus pre–TCR-expressing cells upon CD3 stimulation, suggesting that γδTCR signaling determines enhanced ATP synthesis and release in the developing T cell.
Analysis of P2X transcripts by real-time PCR revealed higher expression of p2rx7 in γδ+CD25+ cells. Then, we analyzed T cell development in FTOC with E14 thymi from p2rx7−/− embryos. Despite normal β selection, these experiments showed a significant increase of γδTCR-expressing DP cells, indicating that lack of P2X7 signaling favored the aberrant transition of γδTCR-expressing cells to the DP stage. This cell subset was observed in pTα-deficient mice in which thymus cellularity is dramatically reduced and the absence of the pre-TCR complex severely impairs the development of αβ T cells. Conversely, γδ lineage differentiation is not impaired, and γδTCR expression together with Notch could support development of DP αβ-lineage like cells (28). The αβ-lineage–like development observed in γδTCR-bearing p2rx7−/− thymocytes was observed also in WT cells by pharmacological P2X antagonism with oATP or selective inhibition of P2X7 by A438079 both in C57BL/6 and BALB/c FTOC. These data indicate that ATP/P2X signaling contributes to the acquisition of the physiological CD4−8− γδ cells phenotype. Accordingly, FTOC performed in the presence of the prototypic P2X7 agonist, Bz-ATP, showed a significant increase of CD4−8− γδ cells.
Attenuation of ERK phosphorylation induced by γδTCR signaling diverted γδ thymocytes to the αβ lineage differentiation pathway indicating that TCR signal strength was crucial in the acquisition of the CD4−8− phenotype (9, 10). Because ATP/P2X signaling promotes ERK activation (24), we compared ERK phosphorylation in WT and p2rx7−/−CD25+ γδ cells upon TCR stimulation. P2rx7−/− γδ cells displayed impaired phospho-ERK and cell size increases after 16-h stimulation with TCRδ mAb, an indication of the contribution of P2X7 to γδTCR signaling in CD25+ γδ cells. In addition, P2X7 expression determined an increase in ATP synthesis upon TCR stimulation likely through an autocrine feed-forward loop, as already observed in regulatory T cells (24). These results point to a role of P2X7 in tuning TCR signal strength in T cell development at the αβ/γδ lineage bifurcation checkpoint. Thus, we propose a model whereby ATP released by γδTCR signaling in immature thymocytes activates P2X receptors in an autocrine fashion and contributes to ERK phosphorylation as well as to implement the transcriptional program required for γδ T cell lineage commitment. In contrast with murine γδ cells, which differentiate independently from Notch, human γδ thymocytes depend on increased levels of Notch signaling than αβTCR-expressing cells (29). Human Notch activation might be interpreted as a stronger signal, thus conforming human T cell development to the TCR signal strength concept in T cell lineage commitment. In this respect, it would be interesting to see whether purinergic signaling contributes to TCR signal strength not only in murine but also in human T cell development.
Although the ERK-Egr-Id3 signaling axis plays a role in γδ lineage development, mice lacking Id3 display elevated numbers of γδ T cells (26, 30, 31). In fact, Id3-null mice have increased generation of “innate-like” γδ cells expressing TCR Vγ1.1 Vδ6.3 with limited junctional diversity. This γδ T cell subset shares phenotypic and functional characteristics with αβ NKT cells (27) and is repressed by Id3 (30, 31). In line with a role of P2X7 activity in the ERK-Egr-Id3 signaling pathway, p2rx7−/− mice displayed increased peripheral representation of this cell subset, suggesting a role of extracellular ATP in modulating γδ NKT cell development.
The authors have no financial conflicts of interest.
We thank Juan Carlos Zúñiga-Pflücker for providing OP9-GFP and OP9-DL1 cells, David Jarrossay for cell sorting, and Enrica Mira Cató for technical assistance with mice.
This work was supported by Grant 310030-124745 from the Swiss National Science Foundation, Grant KFS 02445-08-2009 from the Swiss Cancer League, the Fondazione Ticinese per la Ricerca sul Cancro, the Fondazione Leonardo, the AGD Italia (Coordinamento Associazioni Italiane Giovani con Diabete), Leo Club Italia, and the Sixth Research Framework Program of the European Union, Project MUGEN (MUGEN LSHG-CT-2005-005203) (to F.G.).
The online version of this article contains supplemental material.
Abbreviations used in this article:
- early growth response
- extracellular signal regulated kinase
- fetal thymic organ culture
- inhibitor of differentiation 3
- periodate-oxidized ATP
- OP9 BM stromal cell transduced with the Notch ligand Delta-like 1
- Received June 2, 2011.
- Accepted April 28, 2012.
- Copyright © 2012 by The American Association of Immunologists, Inc.