Research ArticleImmunology

Regulation of Zap70 Expression During Thymocyte Development Enables Temporal Separation of CD4 and CD8 Repertoire Selection at Different Signaling Thresholds

See allHide authors and affiliations

Science Signaling  23 Mar 2010:
Vol. 3, Issue 114, pp. ra23
DOI: 10.1126/scisignal.2000702

Abstract

To investigate the temporal regulation of the commitment of immature thymocytes to either the CD4+ or the CD8+ lineage in the thymus, we developed a transgenic mouse that expressed a tetracycline-inducible gene encoding the tyrosine kinase ζ chain–associated protein kinase of 70 kD (Zap70), which restored development in Zap70−/− thymocytes arrested at the preselection, CD4+CD8+ double-positive (DP) stage. After induction of the expression of Zap70 and the production of Zap70 protein, CD4+ single-positive (SP) cells that expressed Zbtb7b (which encodes the CD4+ T cell–associated transcription factor ThPOK) became abundant within 30 hours, whereas CD8+ SP cells were not detectable until day 4. We found that mature CD4+ and CD8+ cells arose from phenotypically distinct subsets of DP thymocytes that developed with different kinetics and contrasting sensitivities to stimulation of the T cell antigen receptor (TCR). In wild-type mice, expression of endogenous Zap70 progressively increased during maturation of the DP subsets, and the abundance of Zap70 protein determined the sensitivity of the cells to stimulation of the TCR. This temporal gradient in the amount of Zap70 protein enabled the selection of CD4+ and CD8+ repertoires in separate temporal windows and at different TCR signaling thresholds, thereby facilitating discrimination of distinct positive selection signals in these lineages.

Introduction

Hematopoietic progenitor cells that enter the thymus can give rise to multiple lineages during their maturation. The presence of CD4 or CD8 receptors at the cell surface defines the distinct developmental stages of thymocyte maturation. CD4CD8 double-negative (DN) cells are the most immature and can be further divided, on the basis of the amounts of CD44 and CD25 found on the cell surface, into the DN1 through DN4 subsets. Examination of these different stages of maturation has identified branch points for different lineages. Natural killer (NK) cells that have the interleukin-7 receptor (IL-7R+) are generated within the DN population (1), most likely at the pre-DN3 stage, whereas T cells that have a rearranged γδ T cell antigen receptor (TCR) specifically arise at the DN3 stage (2). Thymocytes destined for the αβ T cell lineage rearrange their α and β TCR genes at the DN3 to DN4 stages and then increase the abundance of both CD4 and CD8 co-receptors. At this CD4+CD8+ double-positive (DP) stage, thymocytes undergo a process of selection in which those few cells that have TCRs with the appropriate low avidity for self-peptide presented by the major histocompatibility complex (MHC) (spMHC) are stimulated to differentiate into either CD4+ or CD8+ lineage T cells (3).

Various studies have identified a number of key signaling molecules and transcription factors that are involved in the genetic programs that promote development of the CD4+ and CD8+ T cell lineages. Runx3 is important for silencing the expression of the gene encoding CD4 (4, 5) during the development of CD8+ T cells, whereas the nuclear factors GATA-3 (6, 7), TOX (8), and ThPOK (also known as cKrox) (9, 10) are required for the generation of the CD4+ lineage. How these distinct genetic programs are triggered in DP precursors is less clear, and it has been suggested that the qualitative nature of the positive selection signal instructs the development of thymocytes. Experiments that tuned proximal TCR signaling in vivo by manipulating the activity of Src family kinases suggested that “strong” signals promote the development of CD4+ T cells (1113), whereas the development of CD8+ T cells requires “weaker” signals (11). Other studies also implicate the duration rather than the strength of TCR signaling in triggering cell-fate decisions, such that sustained signaling promotes generation of the CD4+ lineage (1416), whereas shorter (16) or intermittent signaling (14, 15) is required for the generation of CD8+ T cells. What remains unclear, however, is how DP precursors integrate these proximal TCR signals to trigger specific fate decisions or which downstream pathways are ultimately responsible for distinct developmental pathways.

The tyrosine kinase ζ chain–associated protein kinase of 70 kD (Zap70) is a direct downstream target of Src family kinases and is essential for TCR signaling. Because of redundancy in function with its related family member Syk throughout the DN stage of development, thymocytes from Zap70-deficient mice are blocked at the DP stage rather than at the earlier β-selection checkpoint, a common developmental block for mice deficient in many other TCR signaling molecules. This absolute requirement for Zap70 at the DP stage has been exploited in previous studies to examine the requirement for continued TCR signaling during selection by developmentally limiting expression of Zap70 to DP thymocytes with adenosine deaminase expression elements (15, 17); such expression was sufficient for the development of the CD8+, but not CD4+, lineage. The expression of endogenous Zap70 is under developmental regulation in the thymus, being switched on in DN thymocytes and reaching its maximum extent in SP thymocytes (18); however, whether this regulation plays any role in positive selection or in CD4+ and CD8+ lineage development is not known.

A key feature of hematopoietic development is the temporal separation of developmental cues that enable multiple lineages to be derived from single pluripotent progenitors. We therefore examined the timing of CD4-CD8 lineage commitment to determine whether the Zap70-dependent TCR signals required for the development of CD4+ and CD8+ cells were distinguished temporally as well as qualitatively and, if so, to examine how Zap70 was itself involved in the discrimination of lineage-specific signaling events. To do this, we developed an inducible Zap70 transgenic mouse strain in which positive selection of Zap70-deficient thymocytes, arrested at the DP stage of development, could be restored after induction of the expression of Zap70 to observe the kinetics and dynamics of CD4+ and CD8+ lineage development. With this system, we found that CD4-CD8 lineage commitment was indeed subject to temporal regulation. CD4+ lineage cells were generated rapidly after resumption of selection, whereas the appearance of the CD8+ lineage was delayed by 3 or more days. By measuring the abundance of both the TCR and CD5 on the cell surface, we identified DP thymocytes at temporally distinct stages of maturation that contained precursors of different lineages. Finally, we showed that regulation of the expression of Zap70 during the maturation of DP thymocytes underlies the temporal regulation of the different lineages to create distinct temporal windows in which each lineage is selected at different TCR signaling thresholds.

Results

Generation of inducible Zap70 transgenic mice

To investigate the temporal regulation of positive selection, we developed a system in which we could control the selection of a synchronized population of CD4+CD8+ DP thymocytes in vivo. Thymocytes of Zap70−/− mice are completely arrested at the DP stage because of a failure to transduce the TCR signals that are required for positive selection (19). Therefore, we used the tetracycline-inducible system (20) to develop a mouse strain in which we could control expression of Zap70 and, therefore, thymic development by administration of the inducing antibiotic doxycycline (dox) in vivo. We generated mice bearing inducible tetracycline transgenes of Zap70 and a tailless human CD2 reporter construct (TreZap70 mice), and inducible expression was targeted within the T cell lineage by breeding these mice with mice that constitutively expressed reverse tetracycline transactivator (rtTA) under the control of human CD2 expression elements (rtTAhuCD2) (12). Mice were additionally backcrossed onto an endogenous Zap70−/− background (19) (generating TreZap70 rtTAhuCD2 Zap70−/− mice, referred to as TetZap70 mice hereon), so that all of the expression of Zap70 in thymocytes came from the induction of TetZap70.

Analysis of steady-state T cell development in the thymi of TetZap70 mice that were continuously fed dox throughout life confirmed that the expression of TreZap70 could rescue normal thymic development. Both huCD2 and Zap70 proteins were readily detectable in the thymi of TetZap70 mice, and the combined detection of surface huCD2 and intracellular Zap70 proteins revealed that huCD2 was a faithful reporter of the expression of Zap70 (Fig. 1A). After administration of dox, Zap70 was detectable in typically half of all thymocytes (Fig. 1B) as a result of transgene variegation (21). However, the abundance of Zap70 protein was consistent in thymocytes that successfully induced expression of the transgene (Fig. 1B). Analysis of thymocyte populations in TetZap70 mice revealed the presence of CD4+ SP cells, albeit at a lower frequency than in wild-type mice, which were not present in Zap70−/− mice (Fig. 1C). CD4+ SP and CD8+ SP cells were evident when huCD2+ cells were specifically analyzed (Fig. 1D), and they were mature TCRhi cells (Fig. 1F). In the absence of Zap70, the abundance of both TCR and CD5 on thymocytes is much reduced relative to that in wild-type mice (19, 22). In huCD2+ DP thymocytes from dox-fed TetZap70 mice, induction of the expression of Zap70 completely restored the normal abundance of TCR and CD5, whereas huCD2 DP thymocytes closely resembled Zap70−/− DP cells, confirming the absence of Zap70 in huCD2 cells (Fig. 1E). Absolute numbers of TCRhiCD4+ SP and TCRhiCD8+ SP cells in TetZap70 mice were substantially restored compared to those of Zap70−/− mice (Fig. 1G). The numbers of SP cells did not reach those of the wild-type animals, but this was largely a consequence of the failure to induce Zap70 in all of the TetZap70 thymocytes, because analysis of the frequencies of CD4+ and CD8+ SP thymocytes amongst huCD2+ cells revealed a near-normal representation of the subsets compared with those of wild-type mice (Fig. 1H).

Fig. 1

Reconstitution of thymic development in Zap70−/− mice by Tet-inducible expression of Zap70. TetZap70 mice were fed dox-containing food continuously and their thymi were characterized at 6 to 8 weeks of age (n = 13 mice). (A) The histogram on the left shows the presence of huCD2 in live thymocytes from WT mice (gray fill) and dox-fed TetZap70 mice (solid line), whereas the histogram on the right shows Zap70 in thymocytes from WT mice (broken line), Zap70−/− mice (gray fill), and TetZap70 mice (solid line). The contour plot shows the correlation between huCD2 and Zap70 in thymocytes from TetZap70 mice. (B) Box plots show the percentage of thymocytes from TetZap70 mice that express huCD2 and the mean fluorescence intensity (MFI) of Zap70 after TetZap70 mice were fed with dox-containing food for between 1 and 9 days. (C) Analysis of CD4 and CD8 on thymocytes from WT, Zap70−/−, and TetZap70 mice. The frequencies of the CD4+ SP, DP, and CD8+ SP populations, expressed as a percentage of the total cells analyzed, are shown for the indicated gates. (D) Histogram shows the gates used to identify huCD2+ and huCD2 TetZap70 thymocytes. The two-dimensional contour plot shows CD4 and CD8 on huCD2+ thymocytes from TetZap70 mice. (E) Histograms show TCR and CD5 on thymocytes from Zap70−/− mice (gray fill) compared to huCD2 DP thymocytes from TetZap70 mice (solid line), as well as TCR and CD5 on DP thymocytes from WT mice (gray fill) in comparison to huCD2+ DP thymocytes from TetZap70 mice (solid line). The huCD2+ and huCD2 populations were determined as shown in (D). (F) Histograms show TCR on CD4+ SP and CD8+ SP cells from WT (gray fill), Zap70−/− (broken line), and TetZap70 (solid line) mice. (G) Box plots show the total numbers of TCRhiCD4+ SP cells and TCRhiCD8+ SP cells in the thymi of WT (n = 8 mice), TetZap70 (n = 6 mice), and Zap70−/− mice (n = 11 mice). (H) Box plots are of cell frequencies, expressed as a percentage of the total number of cells, of TCRhiCD4+ SP and TCRhiCD8+ SP cells from WT (n = 8 mice), huCD2+ TetZap70 (n = 6 mice), and total Zap70−/− thymocytes (n = 11 mice). Data shown in (A) to (H) are representative of five or more experiments.

CD4+ SP and CD8+ SP cells develop with distinct kinetics in TetZap70 mice

To examine the kinetics of thymocyte development in vivo, we analyzed the thymi of dox-free TetZap70 mice at various time points after the dox-induced expression of Zap70. In the absence of dox, the thymocytes of TetZap70 mice were indistinguishable from those of Zap70−/− mice, lacking any mature SP populations. Administration of dox to the mice resulted in a rapid increase in the abundance of Zap70 protein, which was detectable at 8 hours and was at maximal amounts at 16 hours, by which time a normal phenotype had been restored among DP thymocytes (fig. S1). Remarkably, at day 1, after Zap70 protein had been produced for only a few hours, a small population of TCRhiCD5hiCD4+CD8lo thymocytes was already detectable (Fig. 2A), and by day 2, CD4+ SP cells were abundant (Fig. 2A). In contrast, CD8+ SP cells did not appear until day 4 or later (Fig. 2, A and D). The distinct kinetics of the generation of CD4+ and CD8+ SP cells was also reflected by the ratio of CD4+ SP cells to CD8+ SP cells, which remained high for the first 4 days after the induction of Zap70 (Fig. 2E).

Fig. 2

Mature CD4+ SP and CD8+ SP cells develop with distinct kinetics. Dox-free TetZap70 mice (n = 3 or more mice per day) were fed dox-containing food, and their thymi were characterized by flow cytometry on the indicated days. (A) Density plots of CD4 against CD8 are shown for total live huCD2+ (upper row) and CD5hiTCRhi gated populations (lower row) at the indicated days. (B) Density plots of CD5 against TCR are shown for DP cells gated from huCD2+ thymocytes from TetZap70 mice at the indicated days after dox feeding. The lower row of plots of CD5 against TCR show magnifications of the areas depicted in the boxes indicated on each of the dot plots above. (C) Density plots of HSA against TCR for CD4+ SP and CD8+ SP cells at the indicated days. (D) Graph shows the frequency of induced huCD2+ thymocytes from TetZap70 mice that were TCRhi CD4+ SP cells (filled circles) or TCRhi CD8+ SP cells (white diamonds) over time compared with the frequencies of TCRhiCD4+ or CD8+SP cells in thymi from WT mice (far right). (E) Graph of the ratio of TCRhiCD4+ SP cells to TCRhiCD8+ SP cells against time after the induction of Zap70 expression in TetZap70 thymi. (F) Density plot of TCR against CD5 on DP thymocytes from WT mice. Gates show the three subsets of DPs, namely, DP1 (TCRloCD5lo), DP2 (TCRintCD5hi), and DP3 (TCRhi CD5int). Data shown in (A) to (F) are representative of four or more independent experiments.

During selection, thymocytes modulate the abundance of CD4 and CD8 co-receptors, and cells of both lineages can arise through a CD4hiCD8lo subset of cells (23). To be sure that the CD4+ SP cells that we were analyzing were indeed cells that were committing to the CD4+ lineage, we examined thymocytes from MHC class II–deficient mice, which have immature CD4hiCD8lo cells but do not have committed CD4+ SP thymocytes. The CD4+ SP cells that we found in the dox-treated TetZap70 mice were indeed distinct from the CD4hiCD8lo cells in MHC class II–deficient mice (fig. S2). Furthermore, analysis of gene expression in CD4+ SP cells purified from dox-treated TetZap70 mice revealed greater amounts of Zbtb7b (ThPOK hereon) messenger RNA (mRNA) at day 2 and day 3 than were detected in CD4hiCD8lo cells from MHC class II–deficient mice (fig. S3), which confirmed their commitment to the CD4+ lineage. Finally, as thymocytes entered the SP population, they retained an immature heat-stable antigen (HSA)hi phenotype until ~4 days after the induction of Zap70 expression (Fig. 2C), after which they started to lose HSA. At about the same time, mature peripheral CD4+ and CD8+ T cells were first detected at days 4 and 5, respectively (fig. S4). The development of T regulatory cells (Tregs) was also restored in TetZap70 mice. Analysis of the abundance of the Treg-specific transcription factor Foxp3 revealed the appearance of substantial numbers of thymic Tregs by day 5 (fig. S5), whose timing was consistent with studies of neonatal mice (24). In contrast, the development of NK T cells was not restored in TetZap70 mice, and these cells were completely absent even from TetZap70 mice that were continuously treated with dox (fig. S6A).

Temporally and developmentally distinct populations of DP thymocytes are defined by TCR and CD5

We next wished to investigate whether the markedly different temporal kinetics of CD4+ and CD8+ lineage development were regulated or influenced by Zap70 signaling. To do this, we wanted to analyze the abundance and signaling properties of Zap70 in thymocytes at different times after the initiation of the positive selection of CD4+ and CD8+ SP lineages. Subtle changes in the abundance of co-receptors are associated with different stages of selection (23). Both lineages initially become CD4loCD8lo and then CD4+CD8lo. CD4+ lineage cells continue to down-regulate CD8 to become CD4+ SP cells, whereas CD8+ lineage cells that are CD4+CD8lo return to the CD4loCD8lo state before becoming CD8+ SP cells. However, the CD4loCD8lo and CD4hiCD8lo populations of cells are heterogeneous in terms of both lineage precursors and the temporal stage of development, for example, early- and late-stage CD4loCD8lo CD8+ lineage cells. Furthermore, even though CD4+ and CD8+ lineages both involve a CD4+CD8lo intermediate, it is not known whether this is an obligate step or whether both lineages pass through this stage with the same timing.

We therefore took advantage of the TetZap70 model to analyze the time-dependent maturation of DP thymocytes to determine a means of identifying “signaled” thymocytes, which have recognized selecting spMHC ligands and have initiated selection, at temporally distinct stages of selection by examining their phenotype at different times after restoring the expression of Zap70. We examined the surface densities of TCR and CD5 on DP thymocytes, because both molecules are regulated during thymic development (2529). The TCR is low in abundance on DP thymocytes and increases in abundance during and after the selection process to generate SP cells (19). CD5 is an inhibitor of TCR signaling (30), and its abundance is controlled by TCR signaling both in the thymus (31) and in mature, peripheral T cells (32, 33). The optimal abundance of both receptors depends on the expression of Zap70 (19, 22) (Fig. 1E). In wild-type mice, we noted three distinct subsets of DP cells: TCRloCD5lo (DP1 hereon), TCRintCD5hi (DP2 hereon), and TCRhiCD5int (DP3 hereon) (Fig. 2D). Induction of the expression of Zap70 in TetZap70 mice rapidly restored DP1 and DP2 populations (fig. S1 and Fig. 2B), whereas the DP3 subset did not develop until 3 days after the induction of Zap70 expression (Fig. 2B). The kinetics of the development of DP2 and DP3 cells correlated with the generation of mature SP cells. The DP2 subset appeared shortly before the generation of CD4+ SP cells, whereas the development of the DP3 population on day 3 preceded the first appearance of CD8+ SP cells from day 4 onward (Fig. 2A), which was suggestive of possible precursor-product relationships.

To directly test the lineage relationships between the DP and SP subsets and to provide independent confirmation of the kinetics of the development of CD4+ and CD8+ SP cells observed in dox-treated TetZap70 mice, DP1, DP2, and DP3 subsets were isolated by sorting from CD45.2 wild-type mice and were directly injected into the thymi of CD45.1 wild-type mice. DP1 and DP2 thymocytes contained preselection and selecting precursors, respectively, of both CD4+ and CD8+ lineages because they gave rise to both CD4+ SP cells and CD8+ SP cells (Fig. 3, A and B). In contrast, transferred DP3 thymocytes gave rise to only CD8+ SP cells (Fig. 3C). The kinetics of development of CD4+ and CD8+ SP cells after intrathymic injection of DP1 cells were virtually identical to those observed in dox-treated TetZap70 mice, as was development of the DP2 and DP3 subsets (Fig. 3A). These kinetics were also confirmed by following the development of DP2 thymocytes; the development of the CD4+ and CD8+ lineages were advanced by about a day relative to that of DP1 cells (Fig. 3B). That developing CD8+ lineage cells underwent serial transition between DP1, DP2, and DP3 subsets was further confirmed by following the development in vivo of a cohort of MHC class I–restricted F5 DP1 thymocytes after intrathymic injection into normal CD45.1 mice. F5 Rag1−/− β2-microglobulin–deficient (b2m−/−) mice were used as donors of the DP1 cells because development was arrested at the DP1 stage in the absence of selecting ligand (Fig. 3D). By day 1 after intrathymic injection into wild-type mice, virtually all of the F5 thymocytes had already entered the DP2 stage and they gradually entered the DP3 stage by day 3, but did not down-regulate CD4 to become CD8+ SP cells until day 4 (Fig. 3D). Finally, analysis of TCR transgenic strains revealed that DP3 populations were present only in thymocytes that expressed MHC class I–restricted receptors (fig. S7A), whereas analysis of gene expression by wild-type DP subsets provided further evidence of the lineage relations between the DP and SP subsets (fig. S7B).

Fig. 3

Lineage development from different subpopulations of DP thymocytes. Sorted DP1, DP2, and DP3 thymocytes from WT mice were injected (5 × 105 to 10 × 105 cells per mouse) directly into the thymi of CD45.1 WT mice (n = 3 mice per day). At different days after transfer of the cells, host thymi were recovered and analyzed by flow cytometry to characterize the CD45.2+ donor and CD45.1+ host thymocytes. (A) The upper row shows density plots of CD4 against CD8, whereas the lower row shows plots of CD5 against TCR. DP1 thymocytes from donor mice before (Pre-Sort) and after (Post-Sort) purification by sorting are shown on the left, and donor thymocytes recovered from recipient mice on the indicated days are shown on the right. Gates show the DP, CD4+ SP, and CD8+ SP subsets as defined by CD45.1 host mice. (B and C) Density plots show CD4 against CD8 and CD5 against TCR of donor DP2 (B) and DP3 (C) thymocytes after sorting and of donor thymocytes recovered from recipient mice on the indicated days after transfer. Pre- and post-sort acquisitions were performed with BD Aria or BD Canto flow cytometers, whereas transferred thymocytes were analyzed with CyAn flow cytometers. Data are representative of three or more independent experiments. (D) Thymocytes from F5 Rag1−/−b2m−/− mice were injected (1 × 106 cells per mouse) directly into the thymi of CD45.1 WT mice (n = 3 mice per day). At different days after transfer, host thymi were recovered and characterized by flow cytometry for the presence of CD45.2+ F5 donor and CD45.1+ host thymocytes. The upper row shows density plots of CD5 against TCR, whereas the lower row shows plots of CD4 against CD8 at the indicated days compared to donor thymocytes before injection (day 0, d0) and to F5 Rag1−/− control thymi. Gates show the DP1 through DP3 subsets as defined by analysis of CD45.1 host (d1 to d6) or WT control mice on the same day. Data are representative of three independent experiments.

Transgenic Zap70 signals influence developmental progression through the DP subsets

Having established that we could analyze DP thymocytes at distinct temporal stages of development through the use of TCR and CD5 as markers, we used TetZap70 mice to examine the influence of the signal strength of Zap70 on the selection of thymocytes through different DP stages and to different lineages. Measurement of calcium ion (Ca2+) flux and the phosphorylation of extracellular signal–regulated kinase (ERK) after the ligation of CD3 on TetZap70 DP thymocytes showed that the abundance of transgenic Zap70 determined the efficiency of TCR signaling (Fig. 4A). Therefore, given the extensive evidence that suggests that “stronger” or sustained selection signals favor the development of the CD4+ lineage and that “weaker” or intermittent signals promote the development of CD8+ cells (11, 12, 15, 16, 34, 35), we reasoned that the development of TetZap70 DP1 thymocytes to the DP2 and DP3 stages would be influenced by their amount of Zap70. Analysis of the abundance of Zap70 in TetZap70 DP thymocytes revealed that this was indeed the case. Whereas DP1 thymocytes exhibited a broad range of Zap70 abundance, the amount of Zap70 in DP2 cells was at the higher end of that observed in the DP1 population. In contrast, the amount of Zap70 in DP3 cells was within the lower range observed in both DP2 and DP1 cells (Fig. 4B), consistent with the view that strong Zap70 signals tended to favor the development of CD4+ cells from the DP2 subset, whereas weak Zap70 signals appeared to favor the continued development of the CD8+ lineage development into the DP3 stage.

Fig. 4

Low expression of transgenic Zap70 favors the development of DP3 thymocytes. (A) TetZap70 mice were fed dox-containing food for 2 to 4 days. The histogram on the left shows the expression of huCD2 by DPs and the gates (1 to 5) used for signaling analysis. The line graph on the right shows the percentage of thymocytes (from each of the five gates) over the threshold that fluxed Ca2+ as a function of the time after cross-linking of CD3 with antibody. Histograms show pERK in TetZap70 thymocytes after cross-linking with antibody against CD3 stimulation for 180 s. Solid lines represent pERK in the thymocytes from the indicated huCD2 gate. The gray fill shows pERK in huCD2 thymocytes (gate 5). (B) The abundance of Zap70 protein was examined by flow cytometry in thymocytes from TetZap70 mice that had been fed dox-containing food for 7 days. The density plot of CD5 against TCR is of huCD2+ gated DP thymocytes from TetZap70 mice and shows gates for the DP1 to DP3 subsets used for subsequent analysis. Histograms show the abundance of Zap70 protein in huCD2+ DP1 (black line), DP2 (red line), and DP3 (blue line) thymocytes from TetZap70 mice and in DP thymocytes from Zap70−/− mice (gray fill) as controls. The histogram on the right shows Zap70 in the same DP1, DP2, and DP3 subpopulations unscaled, unnormalized, and with a biexponential y-axis scale. Numbers indicate the MFI of Zap70 in DP1 (black), DP2 (red), and DP3 cells (blue). Data are representative of three or more independent experiments.

Endogenous regulation of Zap70 abundance during DP thymocyte selection creates a gradient of TCR signaling sensitivity

We next examined the expression patterns of endogenous Zap70 to determine whether lineage specification in the wild-type thymus was controlled in a manner similar to that in TetZap70 mice. As observed in TetZap70 thymocytes, Zap70 was more abundant in wild-type DP2 cells than in DP1 cells (Fig. 5A). Surprisingly, however, in contrast to the amount of Zap70 in TetZap70 mice, endogenous Zap70 was more abundant in wild-type DP3 cells than in either DP2 or DP1 cells (Fig. 5A). The abundance of Zap70 was greatest in the SP thymocytes (Fig. 5A and fig. S8A), as previously reported (18). The striking difference in the abundance of Zap70 in DP3 cells between TetZap70 mice and wild-type mice likely reflects fundamental differences in the regulation of transgenic and endogenous Zap70 in these strains, respectively. In TetZap70 mice, the expression of Zap70 is under the control of a transgenic promotor, and differences in the abundance of Zap70 protein among the different DP subsets likely reflects the preferential selection of DP thymocytes on the basis of the extent of expression of the Zap70 transgene attained in the precursor DP pool. In contrast, the expression of Zap70 in wild-type mice is under developmental control, because it is switched on in DN1 thymocytes and is maximally expressed by the SP stages (18). Analysis of DP thymocytes as they gradually increased the amount of TCR and modulated the abundance of CD5 through the DP1 to DP3 subsets revealed a progressive gradient in the abundance of Zap70 protein (Fig. 5B). Analysis of the expression of Zap70 in DP subsets confirmed that changes in the abundance of Zap70 protein were also associated with increased amounts of mRNA in the DP2 and DP3 subsets (Fig. 5C).

Fig. 5

The abundance of Zap70 in developing WT DP thymocytes determines their sensitivity to stimulation of the TCR. (A) The histogram on the left shows Zap70 in DP1 (black), DP2 (red), and DP3 (blue) subsets of WT thymocytes and in thymocytes from Zap70−/− mice (gray fill). Numbers indicate the MFIs for DP1 (black), DP2 (red), and DP3 cells (blue). The histogram on the right shows Zap70 in DP1 (black), CD4+ SP (red), and CD8+ SP (blue) thymocytes from WT mice and in thymocytes from Zap70−/− (gray fill) mice. Numbers indicate the MFIs for DP1 (black), CD4+ SP (red), and CD8+ SP cells (blue). (B) Density plot shows TCR against CD5 on DP thymocytes from WT mice and the electronic gates (1 to 10) used to classify the changing abundance of Zap70 protein in developing DP thymocytes. The bar chart shows the abundance of Zap70 expressed as MFI ± the coefficient of variation (CV) for the indicated gates. (C) DP1, DP2, DP3, CD4+ SP, and CD8+ SP subsets from WT mice were purified by sorting and the relative abundance of Zap70 mRNA was determined in these samples and in total Rag1−/− thymus by real-time quantitative PCR. (D) Histograms show CD69 and Zap70 in DP thymocytes from WT mice stimulated with PMA (red line), ionomycin (blue line), or both (black line) for 15 hours followed by 7 hours of rest, compared to unstimulated thymocytes (gray fill). (E) The histogram shows Zap70 in DP thymocytes from WT mice stimulated with PMA and ionomycin as in (D) and additionally treated with inhibitors of transcription (actinomycinD, red line) or translation (cycloheximide, blue line) compared to untreated controls (gray fill). (F) Thymocytes from WT or TetZap70 mice that were fed dox-containing food for 7 days were analyzed for the presence of CD4, CD8, TCR, CD5, and huCD2 and labeled with antibody against CD3. Cells required for the analysis of Ca2+ flux were additionally loaded with Indo dye (see fig. S9A for loading controls). Thymocytes were then stimulated by cross-linking antibodies against CD3. Line graphs (left column) show the percentage of thymocytes that contained pERK at 20-s intervals after cross-linking of CD3. Lines indicate the responses of DP1 (broken black line), DP2 (red line), and DP3 (blue line) thymocytes from the indicated mice. Line graphs (right column) show the percentage of thymocytes from WT mice or TetZap70 mice fluxing Ca2+ over a threshold, set from unstimulated thymocytes, as a function of time. Traces shown are for DP1 (broken black line), DP2 (red line), and DP3 (blue line) thymocytes. Data are representative of two (C and E) or more (A, B, D, and F) independent experiments.

The progressive increase in the abundance of Zap70 that we observed in DP cells may have represented continued up-regulation of gene expression initiated in the DN population that continued autonomously through to the SP stages. However, because the substantial increase in the quantity of mRNA was limited to the DP2 and DP3 subsets, it was also possible that the induction of gene expression was directly modulated by TCR signaling during positive selection. We therefore examined whether expression of Zap70 in DP thymocytes could be induced by activating pathways downstream of the TCR. Stimulation of purified DP thymocytes with low concentrations of phorbol ester and Ca2+ ionophore induces the development of CD4+ SP cells (36). The same stimulation of purified DP1 thymocytes resulted in a substantial increase in the expression of Zap70 (Fig. 5D) that was completely blocked by inhibitors of transcription and translation (Fig. 5E). Therefore, our data suggest that the developmental gradient in Zap70 abundance may by regulated by TCR signals during positive selection as part of a potential feed-forward regulatory circuit.

The abundance of Zap70 determines thymocyte signaling sensitivity during development

During the development of wild-type CD8+ cells, it appears that instead of reducing the abundance of Zap70 and thus reducing their sensitivity to TCR stimulation (changes that the “signal strength” models of lineage commitment would have predicted to favor the development of the CD8+ lineage), wild-type thymocytes appear to do exactly the opposite, progressively increasing the abundance of Zap70 and therefore increasing sensitivity to TCR signaling. In TetZap70 mice, the abundance of transgenic Zap70 was a crucial determinant of signaling competency in DP1 thymocytes (Fig. 4A).

To ask whether the increased abundance of Zap70 in wild-type DP thymocytes could determine sensitivity to TCR signaling, we measured two independent TCR induced responses in the different DP subsets: phosphorylation of ERK and Ca2+ flux. In wild-type mice, stimulation of CD3 in DP1 thymocytes gave rise to modest Ca2+ and pERK responses, which were marginally increased in DP2 cells (Fig. 5F). In contrast, wild-type DP3 cells were markedly more sensitive to TCR stimuli than were either DP2 or DP1 thymocytes (Fig. 5F). Because the abundance of the TCR and CD5 in DP3 cells are also modulated in ways that would be expected to enhance sensitivity to TCR stimulation in this subset, we also examined Ca2+ and ERK activation responses in TetZap70 mice. Transgenic DP thymocytes cannot increase the abundance of Zap70 in DP3 cells but do undergo identical changes in the abundance of the TCR and CD5 as do wild-type cells and they have other proteins involved in proximal TCR signaling at a similar abundance to those in wild-type cells (fig. S8B). Similar to wild-type cells, TetZap70 DP1 and DP2 thymocytes made modest responses (Fig. 5F); however, in contrast to cells from wild-type mice, DP3 cells from TetZap70 mice, in which Zap70 is less abundant than it is in wild-type cells, did not exhibit the same substantial increase in TCR sensitivity that was observed in wild-type DP3 cells. ERK and Ca2+ signaling pathways were intact in TetZap70 thymocytes because the positive controls of phorbol ester and ionophore had identical effects on ERK activation and Ca2+ flux, respectively, in DP3 cells from TetZap70 and wild-type mice (fig. S9). Thus, wild-type DP thymocytes appear to progressively increase their sensitivity to TCR stimulation during selection by a mechanism that is critically dependent on the increased abundance of Zap70.

The increased abundance of Zap70 in DP3 thymocytes is required for efficient generation of CD8+ SP cells

Because DP3 cells were the most sensitive DP subset to TCR stimulation, we next asked whether this increase in sensitivity was required to sense the weaker stimuli that drive the positive selection of CD8+ lineage cells. First, we examined development of CD8+ cells in MHC class I–restricted F5 TCR transgenic mice expressing the TetZap70 transgene. DP thymocytes from Zap70−/− F5 mice exhibited a similar loss of surface TCR and CD5 as was seen in Zap70−/− mice that expressed polyclonal TCRs (Fig. 6A). Induction of the expression of transgenic Zap70 restored the normal abundance of TCR and CD5 (Fig. 6A). An examination of the DP subsets revealed a normal DP2-to-DP3 transition in F5 TetZap70 mice compared to that in control F5 mice (Fig. 6, A and B). In contrast, we observed a substantial reduction in the number of mature CD8+ SP cells in F5 TetZap70 mice (Fig. 6, A and B), which suggested a specific block in maturation between the DP3 and CD8+ SP stages.

Fig. 6

Expression of TetZap70 impairs the development of F5 TCR transgenic T cells. F5 Rag1−/− TetZap70 mice were fed dox-containing food throughout life and huCD2+ thymocytes were analyzed by flow cytometry at 6 to 8 weeks of age. (A) Density plots show CD5 against TCR on cells in the DP thymocyte gate and CD4 against CD8 on total live thymocytes from the indicated mouse strains. Cells from F5 Rag1−/− TetZap70 mice were additionally gated on huCD2. (B) Bar chart shows the ratio of the numbers of DP3 to DP2 cells and of CD8+ SP to DP2 cells for F5 control (F5 Con) and F5 Rag1−/− TetZap70 mice. *P < 0.005 (n = 8 mice of each strain). Data are representative of four independent experiments.

Second, we looked for evidence of defective CD8+ cell development in dox-treated polyclonal TetZap70 mice. The ratio of CD4+ SP cells to CD8+ SP cells in the thymi of TetZap70 mice continuously treated with dox was moderately, but significantly, increased relative to that in the thymi of wild-type mice [CD4 SP–CD8 SP ratios of 4.5 ± 1.0 for wild-type mice (n = 15) and 6.0 ± 1.4 for TetZap70 mice (n = 11 mice), P < 0.025] and also remained increased for the duration of experiments in which TetZap70 mice were fed dox-containing food as adults (Fig. 2E). To investigate this potential defect in the generation of CD8+ SP cells more closely, we examined the development of TetZap70 thymocytes in competitive mixed irradiation chimeric mice in which differences in selection efficiency at different developmental stages would be more sensitively revealed by direct comparison with partnered thymocytes from wild-type mice. Host mice reconstituted with mixtures of CD45.1 wild-type and CD45.2 TetZap70 bone marrow cells were fed dox-containing food for 10 to 16 weeks and the chimeric thymi were characterized by flow cytometry. The generation of CD8+ SP cells among the cells of CD45.2 TetZap70 origin was greatly reduced when compared with that of control, CD45.1 wild-type thymocytes (Fig. 7A), and there was a substantial increase in the ratio of TetZap70 CD4+ SP cells to CD8+ SP cells [CD4 SP–CD8 SP ratio of 4.1 ± 1.1 for wild-type mice (n = 7) and 9.2 ± 2.0 for TetZap70 mice (n = 7 mice), P < 0.0005].

Fig. 7

The increased abundance of Zap70 in DP3 subsets is essential for the efficient development of CD8+ lineage cells. (A to C) Irradiated Rag1−/− recipient mice were reconstituted with a 1:1 mixture of CD45.1 WT and CD45.2 TetZap70 bone marrow cells and were fed dox-containing food. Between 10 to 16 weeks after reconstitution, thymi (from n = 8 mice) were analyzed by flow cytometry. (A) Density plots show CD4 against CD8 for CD5hiTCRhi gated thymocytes of WT (CD45.1) or TetZap70 (CD45.2) origin, as indicated. The difference in the percentage of recovered CD8+ SP cells from WT or TetZap70 donors was significantly different (P < 0.0006, n = 8). (B) Density plots show CD5 against TCR for DP cells in the thymus and the gates used to define the DP1 through DP3 subsets, as well as CD4 against CD8 on total thymocytes from chimeras and the gates used to define DP, CD4+ SP, and CD8+ SP populations. huCD2-CD45.2+ thymocytes were excluded from analysis by negative gating. Histograms show CD45.1 on cells from the indicated thymic subset. Numbers indicate the percentage of CD45.2+ TetZap70 cells in each subset. (C) The graph shows the ratio of the number of TetZap70 cells to WT cells, normalized to the starting ratio in DP1, during the developmental progression of CD4 (DP1, DP2, and CD4+ SP, filled circles) and CD8+ (DP1 to DP3, CD8+ SP, open circles) lineages. Statistics for the ratio of TetZap70 cells to WT cells: DP2 compared to DP3, P < 0.001; DP2 compared to CD8+ SP and DP3 compared to CD8+ SP, P <0.0001. Data are representative of three independent experiments. (D to F) In similar experiments, mixed bone marrow chimeras were generated by reconstituting lethally irradiated Thy1.1 BALB/c mice with a 1:1 mixture of Thy1.1 BALB/c and Thy1.2 SKG bone marrow cells. Between 10 to 16 weeks after reconstitution, thymi (from n = 9 mice) were analyzed by flow cytometry. (D) Density plots are of CD4 against CD8 for CD5hiTCRhi gated thymocytes of WT (Thy1.1) or SKG (Thy1.2) origin, as indicated. (E) Density plots show CD5 against TCR on DP cells in the thymus and the gates used to define the DP1 to DP3 subsets, as well as CD4 against CD8 on total thymocytes from chimeras and the gates used to define the DP, CD4+ SP, and CD8+ SP populations. Histograms show Thy1.1 on the indicated thymic subsets. Numbers indicate the percentage of Thy1.2 TetZap70 cells in each subset. (F) The graph shows the ratio of SKG cells to WT cells normalized to the starting ratio in DP1, during the developmental progression of CD4+ (DP1, DP2, and CD4+ SP, filled circles) and CD8 (DP1 to DP3, CD8+ SP, open circles) lineages. Data are pooled from three independent experiments.

Analysis of chimerism at different stages of thymic development pinpointed a specific block in CD8+ cell development. Relative to earlier DP1 and DP2 stages, the representation of CD45.2+ TetZap70 cells was significantly increased at the DP3 stage (DP3 compared to DP2, P < 0.001), but was subsequently reduced in the mature CD8+ SP population (Fig. 7, B and C). In contrast, chimerism at the DP1, DP2, and CD4+ SP stages revealed the relatively efficient development of CD4+ lineage TetZap70 cells. Thus, although the low abundance of transgenic Zap70 favored the development of DP3 cells, the block in the subsequent development of CD8+ SP cells suggests that the increased abundance of Zap70 in DP3 thymocytes is ultimately essential for successful selection of CD8+ SP cells.

Finally, to determine whether reduced Zap70 function alone was sufficient to replicate the specific block in the development of CD8+ lineage cells that we observed in TetZap70 mice, or whether this block was specifically associated with a failure to increase the abundance of Zap70 in DP3 cells, we examined lineage development in the SKG strain of mice, which have a hypomorphic Zap70 allele. A Trp163→Cys (W163C) mutation in the C-terminal Src homology 2 (SH2) domain of Zap70 in SKG mice causes a defect in TCR signaling in thymocytes that impairs positive selection, which results in few SP cells and reduced numbers of peripheral T cells (37). Thymocytes of these mice are hyporesponsive to TCR stimulation (37) and have a reduced abundance of TCR and CD5, but exhibit a similar developmental increase in the expression of Zap70 to that of wild-type BALB/c thymocytes (fig. S10). Competitive mixed bone marrow chimeras were generated by reconstituting lethally irradiated Thy1.1 BALB/c mice with a mixture of Thy1.2+ SKG and Thy1.1+ BALB/c control bone marrow that had been depleted of T cells. Ten weeks later, the cell populations in the thymi were determined. Analysis of mature thymocytes that originated from Thy1.1+ wild-type mice or Thy1.2+ SKG mice revealed similar frequencies of CD4+ and CD8+ SPs (Fig. 7D). Analysis of chimerism revealed an initial drop in the representation of Thy1.2+ SKG thymocytes between the DP1 and DP2 stages, consistent with the expected block in the initiation of positive selection in these mice (Fig. 7, E and F). However, representation of SKG thymocytes in the CD4+ SP and CD8+ SP populations was similar, indicating that the hypomorphic Zap70 allele supported positive selection of CD4+ and CD8+ lineages equally (Fig. 7, E and F). Hypomorphic Zap70 is most likely compensated for by selecting a higher-avidity repertoire in these mice; hence, successful positive selection signals are generated by using different TCRs. However, because Zap70 was still increased in abundance in SKG mice (fig. S10), DP3 cells could still deliver stronger TCR signals than could DP1 cells, even if the Zap70 mutant was hypomorphic. Thus, reduced Zap70 function alone was insufficient to replicate the specific block in CD8+ lineage development observed in TetZap70 mice, which was associated rather with a failure to increase the expression of Zap70 from the point at which positive selection was initiated.

Discussion

During thymocyte development, DP precursors commit to either a CD4+ or a CD8+ lineage in a complex process that requires the decoding of TCR-dependent selection signals on the basis of parameters such as strength and duration (1113, 15, 16). In this study, we found that the CD4-CD8 lineage fate decision was also subject to temporal regulation such that CD4+ SP and CD8+ SP cells were generated from distinct subsets of DP precursors at different times. Regulation of the abundance of Zap70, the TCR, and CD5 in DP thymocytes during selection defined phenotypically distinct DP subsets with different signaling sensitivities. This developmental gradation of TCR sensitivity permitted selection of the CD4 and CD8 repertoires in separate temporal windows and at distinct TCR signaling thresholds, thereby facilitating the discrimination of positive selection signals associated with each lineage.

Previous studies that examined fetal thymic organ cultures (38), development of BrdU (5-bromo-2′-deoxyuridine)–labeled thymocytes in vivo (39), and reconstitution of irradiated hosts (40) have all suggested distinct kinetics for the development of CD4+ and CD8+ cells. Through two independent systems, either by restoring the expression of Zap70 in Zap70−/− thymocytes or by the intrathymic injection of preselection DP cells, we directly observed the kinetics of thymic development and found that CD4+ thymocytes underwent rapid selection, whereas the selection of CD8+ cells was delayed for several days. Our time-course analysis also revealed that it was possible to robustly identify thymocytes at different temporal stages of selection on the basis of the coexpression of the TCR and CD5. We identified three distinct populations, DP1 through DP3, that represented preselection (DP1), early selection, up to ~48 hours (DP2), and late selection thymocytes (DP3), beyond 48 hours. Consistent with the timing of their first appearance, we found that CD4+ cells arose directly from the DP2 population, whereas CD8+ lineage cells transited through both the DP2 and the DP3 stages. However, the full significance of this temporal component to thymic development has not hitherto been explored, and our data suggest that the timing of signaling is as important a component of positive selection as is either the strength or the frequency of signaling.

Analysis of the expression of Zap70 in TetZap70 transgenic and wild-type DP thymocytes provided insight into the nature of the TCR signals that control early development of the CD4+ and CD8+ lineages during the different DP stages. There is evidence supporting both quantitative (1113) and kinetic (1416, 40) models of lineage commitment. Consistent with the quantitative models, we demonstrated that the Zap70 transgene was highly expressed in DP2 cells and that a low abundance of Zap70 predisposed thymocytes to progress onto the DP3 stage, because the broad range of expression of transgenic Zap70 provided an additional selection parameter during development in TetZap70 mice. The abundance of Zap70 affects the signaling efficiency of thymocytes, and, given the dynamic changes in the sensitivity of thymocytes to TCR signaling as a result of changes in the abundance of CD5, the TCR, and Zap70, it is clear that the strength of TCR signaling evoked by spMHC undergoes considerable tuning during positive selection. However, the relatively low abundance of Zap70 in DP3 cells is also consistent with purely kinetic selection models, because it could be viewed as exacerbating the “break” in positive selection signaling that is proposed to take place as a consequence of co-receptor reversal (14). However, endogenous control of the expression of Zap70 in wild-type mice was different from that in TetZap70 mice. We found that the abundance of Zap70 progressively increased during selection such that the amount of Zap70 protein was greatest in DP3 cells, intermediate in DP2 cells, and least in DP1 cells, in contrast to the pattern observed in TetZap70 mice. Although the low expression of the Zap70 transgene facilitated continued development to the DP3 stage in TetZap70 mice and was successful in distinguishing between the development of CD4+ and CD8+ cells at early stages of selection, this appeared to be a developmental “dead end.”

A key consequence of the inducible transgenic expression system was that it did not enable the same developmental regulation of the expression of Zap70 as was observed in wild-type DP3 cells. The importance of this regulation was evident from the substantial and specific block in the development from the DP3 stage to the CD8+ SP stage in F5 Rag1−/− TetZap70 mice and competitive chimeras, which suggested that the increased expression of Zap70 at the DP3 stage was necessary for the efficient development of the CD8+ lineage and that the generation of CD8+ SPs from DP3 cells was not a fait accompli. Also, the block in the development of CD8+ cells in TetZap70 mice could not be attributed to inefficient TCR signaling by transgenic Zap70, as might be predicted by purely quantitative models of development, because SKG mice with a hypomorphic Zap70 allele generated CD4+ and CD8+ lineages with equal efficiency even though positive selection was greatly reduced overall. These mice exhibited the normal increase in the abundance of Zap70 protein during development.

The increased abundance of Zap70 during the maturation of DP cells and the enhanced sensitivity to TCR signaling that accompanies it therefore appears to provide a key mechanism to aid thymocytes distinguish the qualitatively distinct proximal TCR signals induced by MHC class I– and class II–restricted positive selection ligands. Whereas low amounts of Zap70 were sufficient to transduce signals from MHC class II–restricted selection ligands at the DP2 stage, it appeared that CD8+ lineage cells waited until the amount of Zap70 protein in the cells was sufficiently high to transduce the weaker stimuli associated with MHC class I–restricted ligands. That activated thymocytes require further signaling to trigger completion of the selection process is consistent with studies that described a proofreading mechanism that controls completion of positive selection (17). Our data suggest that the initiation of selection from DP1 thymocytes and the completion of selection by CD4+ and CD8+ lineage cells all occur at different signaling thresholds, because CD8+ lineage cells enter the DP2 stage as rapidly as do CD4+ lineage cells but fail to complete selection for several days until they enter the DP3 stage. The changes in the abundance of the TCR and CD5 that we observed as cells transited from the DP2 to DP3 stages may also contribute to tuning sensitivity to stimulation of the TCR during selection, because an increase in the amount of TCR and a loss of CD5, an inhibitor of TCR signaling, would also be expected to increase the sensitivity of the cells to stimulation of the TCR. Consistent with this, although DP3 cells from TetZap70 mice exhibited poor Ca2+ flux in response to TCR stimulation when compared to wild-type DP3 cells, the responses were still slightly stronger than were those from DP2 cells from the same mice, despite their lower abundance of Zap70. Whereas the amount of CD5 at the cell surface is tuned during thymic development, it is notable that down-regulation of CD5 by DP3 cells was also observed in monoclonal F5 and OTI thymocytes, which confirmed that modulation of CD5 is a specific developmental adaptation rather than a sensory one (33).

The regulation of the abundance of Zap70 appears to represent a specific point of control of TCR sensitivity because other TCR-proximal kinases, such as p56lck, are not so regulated during development (41). The abundance of Zap70 is under some degree of developmental control because the amount of protein is initially low in preselection DP thymocytes but is high in mature SPs (18). However, we found evidence that the expression of Zap70 in DP thymocytes was itself regulated by signaling pathways activated by the TCR, suggesting that positive selection signals may themselves induce the expression of Zap70 as part of a dynamic positive feedback loop. The avidity of spMHC-TCR interactions that induce positive selection in thymocytes are thought to be weak, and it is remarkable, given their low abundance of the TCR and Zap70 compared with those of mature T cells, that these interactions in thymocytes can give rise to signals that induce a differentiation response and that can be distinguished from background signaling from nonselecting ligands. However, a feed-forward regulatory circuit involving the induction of Zap70 expression could conceivably play a role in enabling thymocytes to effectively discriminate between background stimulation induced by nonproductive spMHC-TCR interactions and the weak interactions of the TCR with selecting ligands. Background “spikes” in signaling from nonselecting ligands may not reach the threshold required to achieve sustained activation of such a circuit, whereas selecting ligands might be distinguished by their ability to stimulate more sustained, low-level interactions that could trigger such a circuit, resulting in the induction of Zap70 expression that would feedback to reinforce selection signaling. However, linking the expression of Zap70 with ongoing signaling would have the consequence that expression would equally be lost if signaling was not maintained, as might be the case with nonselecting ligands. Furthermore, whereas we showed that the signaling capacity of thymocytes was directly dependent on their amount of Zap70 protein, another study highlighted the importance of key scaffolding residues within the interdomain B region of Zap70 for function during development (42). Therefore, it is possible that the limiting requirement for Zap70 for optimal signaling during selection could reflect a specific requirement for a sufficient abundance of Zap70 protein for optimal scaffold function to occur in addition to its simple kinase activity. Finally, the absence of the NK T lineage in the thymi of TetZap70 mice suggests that a Zap70-dependent feed-forward circuit may also be important in the development of these cells. Consistent with this, NK T cells in wild-type mice contained particularly high amounts of Zap70 protein (fig. S6B).

Whereas we and others (17) find evidence of an essential role for TCR signaling for the completion of CD8+ lineage development, other studies argue that it is IL-7R, and not TCR, signaling that is in fact required (43, 44). We have reported that CD8+ F5 T cells undergo positive selection entirely normally in the complete absence of IL-7R signaling, because F5 T cells do not reexpress IL-7Rα until they reach the peripheral lymphoid compartments after selection (45). Consistent with this, although thymopoiesis is reduced in Il7r−/− mice, there is no specific defect in CD8+ lineage development (46). However, in principle, it is possible that there is redundancy between IL-7 and other γc cytokines that could support CD8+ lineage development in the absence of IL-7Rα. In contrast, in the present study, we found that a decrease in the abundance of Zap70 in DP3 thymocytes was sufficient to cause a profound block late in CD8+ lineage development. Furthermore, because the DP2-DP3 is normal in F5 TetZap70 mice, reduced amounts of Zap70 in F5 DP3 thymocytes would not be anticipated to affect the final stages of CD8+ lineage commitment were it to occur independently of TCR signaling, by whatever mechanism. We found that this is not the case, however, because F5 TetZap70 thymocytes were specifically blocked at the DP3-to-CD8+ SP transition.

Finally, our data also suggest that the DP stage at which completion of selection takes place appears to be involved in determining lineage fate. For the CD4+ lineage, the production of GATA3 (6) in DP2 cells likely plays a key instructive role (47). For the CD8+ lineage, premature termination of selection signaling by co-receptor reversal has been suggested as a mechanism for preventing misdirected lineage development (14). We found that such modulation of co-receptor modulation did not occur until 48 to 72 hours after the initiation of selection of transferred F5 thymocytes, and coincided with the DP2-to-DP3 transition. The increase in the abundance of factors important for CD8 commitment, such as Runx3 (4, 5) and other CD8+ lineage–specific factors, did not occur until the DP3 stage. Consistent with other studies (17, 48), we found that cells in this stage still required selection signals to complete their development. However, it remains to be determined whether instigation of CD8+ lineage specification and the expression of associated genes such as Runx family members represent a default path of differentiation that occurs in the absence of mutually exclusive CD4 commitment (47, 49) or whether it is dependent on continued and successful selection signaling at the DP3 stage.

Materials and Methods

Mice

Inducible Zap70 transgenic mice were generated with the tetracycline regulatory system. Mouse Zap70 complementary DNA (cDNA) was subcloned into a Tet reporter construct featuring an upstream tetracycline response element (TRE), a minimal cytomegalovirus (CMV) promoter, an N-terminal tag, and a downstream IRES (internal ribosomal entry site)–tailless human CD2 reporter, simian virus40 splice donor and acceptor sites, and polyadenylation (polyA) signals. Fragments were prepared and injected into the pronuclei of fertilized oocytes from (CBA × C57Bl/10) F1 mice. Inducible Zap70 transgenic mice (TreZap70) were intercrossed with Zap70−/− mice (provided by A. Weiss, University of California, San Francisco) bearing the reverse tetracycline transactivator domain (rtTA) transgenes under the control of the human CD2 (huCD2) promoter (rtTAhuCD2) to generate TreZap70 rtTAhuCD2Zap70−/− mice (abbreviated to TetZap70 throughout). Breeders and weaned pups were fed dox in food (3 mg per gram of body weight) to induce continuous expression of Zap70. These strains and Rag1−/− mice, F5 Rag1−/−, F5 Rag1−/− TreZap70 rtTAhuCD2Zap70−/− (F5 Rag1−/− TetZap70), MHC class II–deficient mice (B6.129-H2<dlAb1-Ea>), C57Bl6/J, C57Bl6/J CD45.1, SKG, and BALB/c strains of mice were bred in a conventional colony free of pathogens at the National Institute for Medical Research (NIMR), London, UK. All of the TetZap70 lines used were of the H-2b haplotype. Animal experiments were performed according to institutional guidelines and Home Office regulations.

Flow cytometry

Fluorescein isothiocyanate (FITC)–conjugated monoclonal antibody (mAb) against HSA, phycoerythrin (PE)-conjugated mAb against NK1.1, PE-conjugated mAb against Zap70, PE- and PE-Cy5–conjugated mAbs against huCD2, Pacific Blue–conjugated mAb against CD4, allophycocyanin (APC)- and FITC-conjugated mAbs against CD5, PE-Cy5– and APC-conjugated mAbs against TCR, Alexa 780–conjugated mAb against CD44, PE Texas Red–conjugated mAb against CD8α, PE-Cy5–conjugated mAb against CD8β, PE Texas Red–conjugated mAb against Ly5.1, and FITC-conjugated mAb against SLP-76 were obtained from eBioscience. Rabbit antibodies against human LAT, human Zap70 (99F2), human glyceraldehyde 3-phosphate dehydrogenase (GAPDH, 14C10), and pERK1/2 were purchased from Cell Signaling; mouse antibody against human Lck was obtained from Pharmingen; and FITC-conjugated antibody against rabbit immunoglobulin G (IgG) was obtained from the Jackson Laboratories. Detection of CD8+ cells was always performed with an antibody against CD8α, unless otherwise stated in the figure legend. Flow cytometry was performed with 2 × 106 to 5 × 106 thymocytes or with the whole thymus in the case of recipients of intrathymic cellular injections. Cell concentrations were determined with a Scharf Instruments Casy Counter. Cells were incubated with saturating concentrations of antibodies in 100 μl of phosphate-buffered saline (PBS), 0.1% bovine serum albumin (BSA), 1 mM sodium azide (PBS-BSA-azide) for 1 hour at 4°C followed by three washes in PBS-BSA-azide. For intracellular staining, thymocytes were fixed in 4% paraformaldehyde (PFA) at room temperature for 30 min. To detect Zap70, cells were permeabilized with 0.1% NP-40 for 3 min and incubated overnight with PE-conjugated antibody against Zap70. To detect SLP-76, LAT, and Lck, cells were permeabilized with 90% methanol (to detect SLP-76 or LAT) or with 0.1% NP-40 (to detect Lck). To detect pERK, thymocytes were permeabilized with 90% methanol for 30 min on ice. We used FITC-conjugated antibody against SLP-76, whereas LAT, Lck, and pERK were detected with appropriate primary antibodies (listed above) followed by incubation with FITC-conjugated antibody against rabbit IgG. Detection of intracellular Foxp3 was performed with the Foxp3-APC kit (eBioscience) according to the manufacturer’s instructions. Measurement of TCR-stimulated Ca2+ flux in thymocytes was performed essentially as described previously (50). Six-color cytometric staining was analyzed on Canto II and LSR II Instruments (Becton Dickinson), whereas seven-color cytometric staining was analyzed on CyAn instruments (Dako), and data analysis and color compensation were performed with FlowJo V8 software (TreeStar). The data are presented on log and biexponential displays.

Cell culture

DP thymocytes for cell culture were purified by depletion of CD3hi and CD69hi cells with biotinylated mAb and dynal beads (Dynal). Thymocytes were cultured in complete Dulbecco’s modified Eagle’s medium (DMEM) supplemented with l-glutamine, β-mercaptoethanol, and antibiotics, with different combinations of ionomycin (200 ng/ml), phorbol 12-myristate 13-acetate (PMA, 0.2 ng/ml) for 15 hours followed by 7 hours of rest in complete medium. Cultures were also supplemented with actinomycin D (5 μg/ml), cycloheximide (1 μg/ml), and the MEK inhibitor UO126 (10 μM), where appropriate.

Bone marrow chimeras

Bone marrow was extracted from mice and depleted for T cells with Thy1.2bio (eBioscience) and streptavidin A–dynal beads (Dynal). Rag1−/− mice were sublethally irradiated with 500 rad cesium source and injected with 5 × 106 total bone marrow mixtures from wild-type (WT) and TetZap70 donors. Hosts were fed 3 mg of dox-containing food per gram body weight before analysis.

Real-time quantitative polymerase chain reaction assays

Thymocytes from WT mice were incubated with the appropriate antibodies against CD4, CD8, TCR, and CD5, and the following populations were purified to >90% purity by high-speed sorting on a Moflow cytometer (Dako Cytomation) or an Aria machine (Becton Dickinson): DP1 cells (CD4+CD8+TCRloCD5lo), DP2 cells (CD4+CD8+TCRintCD5hi), DP3 cells (CD4+CD8+TCRhiCD5int), CD4+ SP cells (CD4+CD8TCRhi), and CD8+ SP cells (CD4+CD8TCRhi). RNA was isolated by extraction with Trizol (Invitrogen) according to the manufacturer’s instructions and was treated with DNase (Ambion). cDNA was produced with reverse transcriptase (Invitrogen) with oligo(dT) 14-18 (GE Healthcare) and standard protocols. Expression of Zap70, Gata3, CtsW, Zbtb7b, and Runx3 was determined by real-time quantitative reverse transcription polymerase chain reaction (RT-PCR) on an Applied Biosystems ABI Prism 7900 Sequence Detection System with commercial FAM-labeled probes (Applied Biosystems). The abundances of the various mRNAs were normalized to that of Hprt mRNA.

Western blotting analysis

DP1, DP2, DP3, CD4+ SP, and CD8+ SP thymocytes were sorted from WT thymus as described earlier and resuspended in ice-cold lysis buffer [50 mM tris-HCl (pH 8.0), 150 mM NaCl, 1% Triton X-100, chymostatin, leupeptin, pepstatin A, and antipain (5 μg/ml each; Sigma), iodoacetamide (1 mg/ml; Sigma), Pefabloc (0.2 mg/ml; Boehringer Mannheim)] for 30 min, and postnuclear supernatants were obtained by centrifugation. SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and Western blotting were performed according to standard laboratory techniques. Total cell lysates were loaded at 2 × 105 cell equivalents per lane and resolved by 12.5% SDS-PAGE under reducing conditions, blotted onto Immobulin-p membranes, and incubated with rabbit mAb against human Zap70 (Cell Signaling) and rabbit antibody against GAPDH (Cell Signaling). Horseradish peroxidase (HRP)–conjugated antibody against rabbit IgG (Amersham) was used to detect the primary antibodies. Membranes were washed with PBS and 0.5% Tween between sequential incubations with antibody.

Statistical analysis

Statistically significant differences between groups of data were assessed with the Student’s t test assuming unpaired data with unequal variance, with Kaleidagraph V4.1 software.

Acknowledgments

Acknowledgments: We thank R. Murphy, A. Mathiot, P. Ganchevska, S. Tung, and the Biological Services staff for assistance with mouse breeding and typing. We also thank G. Kassiotis, J. Langhorne, A. O’Garra, B. Stockinger, and A. Weiss for providing mice; A. Potocnik and A. Saveliev for technical assistance and discussion; and S. Ley and V. Tybulewicz for critical reading of the manuscript, together with other members of the NIMR for their interest and discussions. Funding: This work was supported by the Medical Research Council UK under programme code U117573801. Author contributions: M.S., C.S., D.M., and M.T. performed experiments; M.S., C.S., and B.S. analyzed data; M.S. and C.S. generated figures; S.S. made available the SKG mouse strain; and B.S. wrote the paper. Competing interests: The authors declare no competing financial interests. Use of the TetZap70 mouse strains requires the signing of a material transfer agreement (MTA).

Supplementary Materials

www.sciencesignaling.org/cgi/content/full/3/114/ra23/DC1

Fig. S1. Rapid induction of Zap70 expression after the administration of dox to TetZap70 mice.

Fig. S2. Abundance of co-receptors on DP thymocytes after induced production of Zap70 protein.

Fig. S3. Rapid induction of ThPOK production in CD4+ SP cells from dox-fed TetZap70 mice.

Fig. S4. Cell populations of peripheral lymph nodes in dox-fed TetZap70 mice.

Fig. S5. Intrathymic development of Foxp3+ Tregs from TetZap70 thymocytes.

Fig. S6. NK1.1+ NK T cells fail to develop in TetZap70 mice fed a dox-containing diet.

Fig. S7. TCR and CD5 define developmentally distinct populations of DP thymocytes.

Fig. S8. The abundance of Zap70, Lck, SLP-76, LAT, and ERK in various thymocyte subsets from wild-type and TetZap70 mice.

Fig. S9. INDO dye loading, Ca2+ flux responses to ionomycin, and generation of pERK in response to phorbol ester by wild-type and TetZap70 thymocytes.

Fig. S10. The increased abundance of Zap70 protein in thymocytes from SKG mice.

References

References and Notes

Stay Connected to Science Signaling

Navigate This Article