Gads-deficient thymocytes are blocked at the transitional single positive CD4⁺ stage

Summary

Positive selection of T cell precursors is the process by which a diverse T cell repertoire is established. Positive selection begins at the CD4⁺CD8⁺ double positive (DP) stage of development and involves at least two steps. First, DP thymocytes downregulate CD8 to become transitional single positive (TSP) CD4⁺ thymocytes. Then, cells are selected to become either mature SP CD4⁺ or mature SP CD8⁺ thymocytes. We sought to define the function of Gads during the two steps of positive selection by analyzing a Gads-deficient mouse line. In Gads^+/+ mice, most TSP CD4⁺ thymocytes are TCR^hiBcl-2^hiCD69⁺, suggesting that essential steps in positive selection occurred in the DP stage. Despite that Gads^−/− mice could readily generate TSP CD4⁺ thymocytes, many Gads^−/− TSP CD4⁺ cells were TCR^loBcl-2^loCD69⁻, suggesting that Gads^−/− cells proceeded to the TSP CD4⁺ stage prior to being positively selected. These data suggest that positive selection is not a prerequisite for the differentiation of DP thymocytes into TSP CD4⁺ thymocytes. We propose a model in which positive selection and differentiation into the TSP CD4⁺ stage are separable events and Gads is only required for positive selection.

Introduction

A diverse, yet restricted, peripheral T cell repertoire is critical to fighting infection while preventing autoimmune disease. The T cell repertoire is established when thymocytes pass from the CD4⁺CD8⁺ double positive (DP) stage of development to the single positive (SP) CD4⁺ or SP CD8⁺ stage. During the transition from DP thymocytes to SP thymocytes, thymocytes down-regulate CD8 and then become transitional SP (TSP) CD4⁺ thymocytes [1-3]. Some TSP CD4⁺ thymocytes maintain the SP CD4⁺ phenotype and become mature SP CD4⁺ thymocytes. Other TSP CD4⁺ thymocytes can re-express CD8 and become mature SP CD8⁺ thymocytes. The existence of an intermediate population between the DP and mature SP stages implies that the signaling pathways that regulate the differentiation of DP thymocytes into TSP CD4⁺ thymocytes may not be the same as those required for CD4/CD8 lineage commitment.

We investigated the signaling pathways that regulate the differentiation of DP thymocytes into mature SP thymocytes by focusing on Gads, an adaptor protein required for TCR-mediated calcium mobilization [4, 5]. TCR ligation results in the activation of Src-family and Syk/ZAP-70 protein tyrosine kinases. Activation of these kinases results in the phosphorylation of tyrosine residues of LAT. Gads binds phosphorylated LAT and recruits SLP-76 into the signaling complex [6-13]. The formation of the LAT-Gads-SLP-76 complex enables the activation of Phospholipase C-γ1, resulting in calcium mobilization.

Mice lacking Gads or over-expressing a dominant negative form of Gads have multiple, but partial defects in T cell development [4, 5, 14]. The first checkpoint during thymocyte development that is dependent on Gads correlates with TCRβ expression in CD4⁻CD8⁻ double negative (DN) thymocytes [4, 5, 14, 15]. The primary function of Gads at the DN stage of development is to promote survival of TCRβ⁺ DN thymocytes.

Gads is also required for the transition from DP thymocytes to mature SP thymocytes. The percentages of SP CD4⁺ to SP CD8⁺ thymocytes are markedly reduced in Gads^−/− mice, as compared to wild-type mice [4, 5, 14]. In addition, Gads^−/− thymocytes that expressed an MHC class I-restricted TCR could not generate as many SP CD8⁺ thymocytes than Gads-expressing cells, suggesting a defect in positive selection [4]. While these studies indicated that Gads is important for the generation of mature SP thymocytes, it remains unclear where Gads functions in the multi-step process of generating mature SP thymocytes from DP thymocytes. In this report, we extended previous studies by determining the precise nature of the block in T cell development seen in Gads^−/− mice.

Results

To examine the function of Gads during the differentiation of DP thymocytes into mature SP thymocytes, we analyzed CD4 and CD8 expression on total wild-type and Gads^−/− thymocytes. As previously reported [4, 5, 14], a smaller percentage of thymocytes in Gads^−/− mice were SP CD4⁺ or SP CD8⁺ thymocytes than in wild-type mice (Fig. 1A). We investigated the site of the developmental block in Gads^−/− mice by analyzing CD24 and TCRβ expression on SP CD4⁺ and SP CD8⁺ thymocytes (Fig. 1B). The SP CD8⁺ thymocyte population can be divided into immature SP (ISP) CD8⁺ thymocytes, which are CD24^hiTCR^lo and represent the developmental stage between the DN and DP populations, and mature thymocytes, which are TCR^hi. There were no statistically significant differences in the percentage of SP CD8⁺ thymocytes that were ISP CD8⁺ or mature SP CD8⁺ thymocytes in wild-type and Gads^−/− mice.

Fig. 1. Gads is required for TCR up-regulation in DP and SP thymocytes.

Gads^+/+ and Gads^−/− thymocytes were surface labeled with anti-CD4, anti-CD8, anti-CD24, and anti-TCRβ and analyzed by flow cytometry. A) The percentages of total thymocytes that were DP, SP CD4⁺, and SP CD8⁺ are shown. B) Thymocytes were gated on SP CD4⁺ or SP CD8⁺ cells (as defined in (A) and analyzed for CD24 and TCRβ expression. For SP CD8⁺ thymocytes, the percentages of SP CD8⁺ thymocytes that are ISP (CD24^hiTCRβ^lo) and mature (CD24^loTCRβ^hi) are shown. For SP CD4⁺ thymocytes, the percentages of cells that are TCR^lo TSP CD4⁺, TCR^hi TSP CD4⁺, and CD24^loTCR^hi, are shown. C) The percentages of Gads^+/+ and Gads^−/− DP thymocytes that were TCRβ^hi are shown. D) The absolute numbers of Gads^+/+ and Gads^−/−thymocytes in each population were divided by the absolute number of ISP CD8⁺ thymocytes to calculate the relative number of cells. *p < 0.005, n= 6 wild-type, 8 Gads^−/−, Wilcoxon signed-rank test. Mat, mature.

In contrast to the SP CD8⁺ population, there were dramatic differences in the subpopulations of SP CD4⁺ thymocytes (Fig. 1B). Fewer mature CD24^lo SP CD4⁺ thymocytes were detected in Gads^−/− mice, as compared to wild-type mice (49% ± 3.9% in wild-type mice and 14% ± 4.4% in Gads^−/− mice, p < 0.0001, n > 6). Within the TSP CD4⁺ thymocyte population (CD24^hi SP CD4⁺), an average of 86% ± 4.4% of Gads^−/− cells expressed low levels of TCR on their surface, as compared to 4.3% ± 1.6% of wild-type TSP CD4⁺ cells (p < 0.0001). We next analyzed TCR expression on DP thymocytes and found that only 5.3% ± 1.2% of Gads^−/− DP thymocytes expressed high levels of surface TCR, as compared to 14% ± 2.9% of wild-type DP thymocytes (p = 0.0066) (Fig. 1C).

The reduced percentages of SP CD4⁺ and SP CD8⁺ thymocytes in Gads^−/− mice and the reduced surface TCR levels on Gads^−/− DP and SP CD4⁺ thymocytes were consistent with impaired positive selection in Gads^−/− mice. These data also suggested that DP thymocytes could proceed from the DP stage to the TSP CD4⁺ stage independently of Gads, but TCR up-regulation was Gads-dependent. To test this model, we sought to analyze the absolute numbers of thymocytes in each subpopulation of wild-type and Gads^−/− thymocytes. However, the defects in CD4⁻CD8⁻ double negative subsets seen in Gads^−/− mice [5, 15] prohibited us from directly comparing the absolute number of DP and SP populations in wild-type and Gads^−/− mice. To circumvent this problem, we normalized the number of DP and SP populations to the absolute number of ISP CD8⁺ thymocytes in each mouse line (Fig. 1D). The relative numbers of DP thymocytes were comparable in wild-type and Gads^−/− mice. The relative number of total TSP CD4⁺ thymocytes in Gads^−/− mice was half the number in wild-type mice (13 ± 8.0 in wild-type mice (n = 6) and 6.5 ± 1.9 in Gads^−/− mice (n = 8, p = 0.024, Wilcoxon signed-rank test)). Within the TSP CD4⁺ population, there were more TCR^lo cells and fewer TCR^hi cells in Gads^−/− mice than wild-type mice (Fig. 1D). These data suggested that the differentiation of ISP CD8⁺thymocytes into DP thymocytes was Gads-independent and Gads^−/− DP thymocytes could differentiate into TSP CD4⁺ thymocytes. However, Gads was required for TCR up-regulation in DP thymocytes.

The relative number of mature SP CD4⁺ thymocytes in Gads^−/− mice was approximately 10% the number in wild-type mice while the relative numbers of mature CD8⁺ thymocytes were comparable in wild-type and Gads^−/− mice (Fig. 1D). These data suggested that Gads is more required for positive selection of CD4⁺ T cells than CD8⁺ T cells.

TCR^lo TSP CD4⁺ thymocytes have not undergone positive selection

In wild-type mice, most TSP CD4⁺ thymocytes had already up-regulated their surface TCR levels while the cells were in the DP stage of development (Fig. 1). However, many Gads^−/− TSP CD4⁺ thymocytes expressed low levels of TCR on their surface. This indicated that either Gads^−/− thymocytes entered the TSP CD4⁺ stage without undergoing positive selection or Gads^−/− thymocytes had dysregulated TCR expression. To distinguish between these possibilities, we examined other markers of positive selection and maturation, CD69 and CD62L. On average, 3.9% ± 0.71% of Gads^−/− DP thymocytes expressed CD69, as compared to 9.0% ± 0.41% of wild-type DP cells (Fig. 2A). Among wild-type TSP CD4⁺ thymocytes, nearly all of the cells that were TCR^hi also expressed CD69 and nearly half the TCR^lo TSP CD4⁺ T cells expressed CD69 (Fig. 2B). These data indicate that the vast majority of wild-type TSP CD4⁺ thymocytes had already undergone positive selection, but a small number of cells can enter the TSP CD4⁺ population without undergoing positive selection. In Gads^−/− mice, only a few TCR^lo TSP CD4⁺ thymocytes had undergone positive selection. These data suggest that the DP to TSP CD4⁺ transition is distinct from positive selection.

Fig. 2. Gads^−/− thymocytes proceed to the TSP CD4⁺ stage without undergoing positive selection.

Gads^+/+ and Gads^−/− thymocytes were surface labeled with anti-CD4, anti-CD8, anti-CD24, anti-TCRβ, anti-CD69, and anti-CD62L and analyzed by flow cytometry. A) The percentages of Gads^+/+ and Gads^−/− DP thymocytes that were CD69⁺ are shown. B) CD69 and CD62L expression was analyzed in each SP CD4⁺ subpopulation defined in Fig. 1. The percentages of cells in each quadrant are shown. C) CD69 and CD62L expression was analyzed on CD24^loTCRβ^hi SP CD8⁺ thymocytes. The percentages of cells in each gate are shown. D) The percentages of CD24^loTCRβ⁺ SP CD8⁺ thymocytes that expressed CD122 are shown.

As SP thymocytes mature, they down-regulate CD69 and up-regulate CD62L. Among wild-type CD24^lo SP CD4⁺ thymocytes, 42% ± 7.5% were of the most mature phenotype (CD69-CD62L^hi), as compared to only 25% ± 2.4% of Gads^−/− CD24^lo SP CD4⁺ thymocytes that were CD69⁻CD62L^hi), p = 0.0017. These data indicate that positive selection and maturation of SP CD4⁺ thymocytes were markedly impaired in Gads^−/− mice.

Gads^−/− mature SP CD8⁺ thymocytes are of a nonconventional phenotype

We next examined the phenotype of the mature SP CD8⁺ thymocyte subset in wild-type and Gads^−/− mice. TCR^hi SP CD8⁺ thymocytes appeared more mature, as defined by CD69⁻CD62L^hi, in Gads^−/− mice than in wild-type mice (Fig. 2C). We also tested whether the mature SP CD8⁺ thymocytes produced in Gads^−/− mice were of a conventional, CD122⁻ phenotype, or of an innate or memory-like, CD122⁺ phenotype [16]. In wild-type mice, 0.98% ± 0.39% of SP CD8⁺ CD24^lo thymocytes expressed CD122 (Fig. 2D). However, 44% ± 5.9% of Gads^−/− mature SP CD8⁺ thymocytes expressed CD122 (p < 0.0001, n = 7). These data suggest that many CD8⁺ thymocytes generated in Gads^−/− mice are of an innate or memory-like phenotype, further emphasizing the importance of Gads in the development of mature T cells.

T cell development in Gads^−/− mice is blocked at the TSP CD4⁺ stage

To further characterize the function of Gads in positive selection, we crossed our Gads^−/− mice with mice expressing one of three different transgenic TCRs. The TCRs expressed were either the MHC class II-restricted TCR OT-II or one of two MHC class I-restricted TCRs, P14 and OTI. Gads^−/− OT-II mice had a dramatic deficiency in the percentage of thymocytes that were SP CD4⁺ (Fig. 3A). Only 1.0% ± 0.15% of total thymocytes from Gads^−/− OT-II mice were SP CD4⁺ cells, as compared to 15% ± 5.7% of Gads^+/- OT-II mice (p < 0.0001, n = 7). Furthermore, greater than 90% of SP CD4⁺ from Gads^−/− OT-II mice were CD24^hiTCR^lo (Fig. 3B), as compared to fewer than 10% of Gads-expressing SP CD4⁺ cells that were CD24^hiTCR^lo. Of the mature SP CD4⁺ thymocytes and CD4⁺ splenocytes produced in Gads^−/− mice, only a small percentage expressed TCR Vβ5 and TCR Vα2 (Figs. 3C and 3D), the two chains that comprise the OT-II TCR [17]. This low frequency of OT-II-expressing mature SP CD4⁺ thymocytes and CD4⁺ splenocytes is in stark contrast to the high percentage of OT-II-expressing cells from Gads-expressing OT-II mice. These data indicate that Gads was required for positive selection of OTII-expressing cells.

Fig. 3. T cell development in Gads^−/− OT-II mice is blocked at the TSP CD4⁺ stage.

A-C) Total thymocytes from Gads^+/− OT-II and Gads^−/− OT-II mice were labeled with anti-CD4, anti-CD8, anti-CD24, anti-TCRβ, anti-TCRβ5, and anti-TCRVα2 and analyzed by flow cytometry. A) Cells were analyzed for CD4 and CD8 expression and the percentages of thymocytes that were SP CD4⁺ and SP CD8⁺ are shown. B) DP, SP CD4⁺, and SP CD8⁺ cells from (A) and analyzed for CD24 and TCRVβ5 expression. The percentages of each population in the indicated gates are shown. C) DN, DP, SP CD4⁺, and SP CD8⁺ thymocytes were analyzed for TCRVβ5 and TCRVα2 expression. The percentages of each population that expressed TCRVβ5 and TCRVα2 are shown. D) Splenocytes from Gads^+/− OT-II and Gads^−/− OT-II mice were labeled with anti-CD4, anti-CD8, anti-TCRVβ5, and anti-TCRVα2 and analyzed by flow cytometry. Left panels, the percentages of total splenocytes that were CD4⁺ or CD8⁺ are shown. Middle and right panels, CD4⁺ and CD8⁺ splenocytes were analyzed for TCRVβ5 and TCRVβ2 expression and the percentage of cells that expressed both proteins are shown.

To determine if Gads^−/− OT-II mice expressed the transgenic TCR, we examined TCR Vβ5 and TCR Vα2 expression throughout T cell development (Fig. 3C). A comparable percentage of DN thymocytes from Gads^+/− OT-II and Gads^−/− OT-II expressed the OT-II TCR. The low percentages of DN and DP thymocytes that expressed the OT-II TCR were similar to that reported by others [17]. These data indicate that Gads^−/− OT-II mice expressed the OT-II TCR, but that Gads^−/− OT-II-expressing thymocytes did not survive T cell development.

Analysis of Gads^−/− P14 mice and Gads^−/− OT-I mice also revealed a block in development at the TSP CD4⁺ stage (Figs. 4 and 5). Only 2.5% ± 1.9% of thymocytes from Gads^−/− P14 mice were SP CD8⁺, as compared to 12% ± 2.0% of thymocytes from Gads^+/+ P14 mice (p < 0.0001, n = 12) (Figs. 4A). Further, only 20% ± 13% of the SP CD8⁺ thymocytes in Gads^−/− P14 were of the mature phenotype, as compared to 50% ± 8.9% of Gads^+/+ SP CD8⁺ cells (Fig. 4B). Similarly, 1.6% ± 0.59% of Gads^−/− OT-I thymocytes were SP CD8⁺ thymocytes, as compared to 13% ± 3.3% of Gads^+/− OT-I thymocytes (p = 0.0003, n = 6) (Fig. 5A). The SP CD4⁺ thymocytes in Gads-expressing and Gads-deficient P14 and OT-I mice were nearly exclusively TSP CD4⁺ thymocytes (Figs. 4B and 5B). However, whereas Gads-expressing SP CD4⁺ thymocytes from P14 and OT-I mice had high levels of surface TCR expression, nearly half of the Gads^−/− TSP CD4⁺ thymocytes had low levels of TCR expression (Figs. 4B and 5B). Further, TCR^lo TSP CD4⁺ thymocytes from Gads^−/− OT-I mice were CD69-negative (data not shown), indicating that the cells reached the TSP CD4⁺ stage without undergoing positive selection.

Fig. 4. T cell development in Gads^−/− P14 mice is blocked at the TSP CD4⁺ stage.

A-C) Total thymocytes from Gads^+/− P14 and Gads^−/− P14 mice were labeled with anti-CD4, anti-CD8, anti-CD24, anti-TCRβ, anti-TCRVβ8, and anti-TCRVα2 and analyzed as described in Fig. 3A-C.

Fig. 5. T cell development in Gads^−/− OT-I mice is blocked at the TSP CD4⁺ stage.

A-C) Total thymocytes from Gads^+/− OT-I and Gads^−/− OT-I mice were labeled with anti-CD4, anti-CD8, anti-CD24, anti-TCRβ, anti-TCRVβ5, and anti-TCRVα2 and analyzed as described in Fig. 3AC. D) Splenocytes from Gads^+/− OT-I and Gads^−/− OT-I mice were labeled and analyzed as described in Fig. 3D.

In summary, the data from the three TCR transgenic mouse lines support a model in which Gads is required for TCR up-regulation, but is not required for progression into the TSP CD4⁺ stage of development. Further, these data demonstrate that the defect in positive selection was much more severe when Gads^−/− mice were crossed to a mouse line expressing an MHC class II-restricted TCR than when crossed to a mouse line expressing an MHC class I-restricted TCR.

Gads^−/− thymocytes have dysregulated CD28 and CD127 expression

To investigate potential mechanisms by which Gads regulates the differentiation of DP thymocytes, we examined the expression of CD28, CD127 (IL-7Rα), and CD132 (γc) on each population of wild-type and Gads^−/− thymocytes. We previously demonstrated that CD28 expression was very high on DN thymocytes [15]. As cells progressed through the ISP CD8⁺, DP and SP stages of development, CD28 expression decreased (Fig. 6A). However, the few TCR^lo TSP CD4⁺ thymocytes that could be found in wild-type mice had much higher CD28 expression than DP cells or TCR^hi TSP CD4⁺ thymocytes. In Gads^−/− mice, CD28 expression across the DP and SP populations was fairly constant, except for the slightly lower expression on the most mature thymocytes (Fig. 6A). These data indicate that Gads regulates the changes in CD28 expression seen in wild-type mice.

Fig. 6. Gads^−/− thymocytes have dysregulated CD28 and CD127 expression.

Total thymocytes from Gads^+/+ and Gads^−/− mice were labeled with anti-CD4, anti-CD8, anti-CD24, anti-TCRβ, and either (A) anti-CD28 (B) anti-CD127, or (C) anti-CD132. The MFIs of CD28, CD127, and CD132 expression for each population of DP and SP thymocytes are shown.

In contrast to CD28 expression, CD127 surface levels increased upon maturation of wild-type thymocytes (Fig. 6B). Wild-type DP thymocytes expressed little or no CD127 protein and TCR^lo and TCR^hi TSP CD4⁺ thymocytes had higher CD127 expression than DP cells. In Gads^−/− mice, CD127 was only expressed on those SP CD4⁺ thymocytes that were TCR^hi. In addition, Gads^−/− CD24^lo SP thymocytes had higher levels of CD127 than the corresponding wild-type populations.

CD132 surface levels were highest on ISP CD8⁺ and mature SP CD8⁺ thymocytes, but comparable across the DP and SP CD4⁺ populations (Fig. 6C). There were no statistically significant differences in the mean fluorescence intensities (MFIs) for CD132 expression between wild-type and Gads^−/− cells in any population, supporting our conclusion that expression of CD28 and CD127 expression was dysregulated in Gads^−/− thymocytes. The MFI of the negative control stain for each thymic subpopulation was less than 100 (data not shown).

Gads^−/− TCR^lo TSP CD4⁺ thymocytes fail to up-regulate Bcl-2

Next, we analyzed Bcl-2 expression in each DP and SP thymocyte population in wild-type and Gads^−/− mice because increased Bcl-2 expression correlates with cells having undergone positive selection [18, 19]. Bcl-2 was not expressed by DP thymocytes (Fig. 7). Bcl-2 expression was higher in wild-type TCR^lo and TCR^hi TSP CD4⁺ thymocytes than in the corresponding Gads^−/− populations. The MFI of Bcl-2 expression for Gads^−/− TCR^lo TSP CD4⁺ thymocytes was 220 ± 40, as compared to 560 ± 51 for wild-type TCR^lo TSP CD4⁺ cells (p < 0.0001, n = 6). Only a subpopulation of Gads^−/− CD24^hiTCRβ^hi SP CD4⁺ thymocytes expressed Bcl-2 at levels comparable to wild-type cells. However, wild-type and Gads^−/− mature CD4⁺ T cells expressed Bcl-2 at comparable levels. These data indicate that Gads^−/− TSP CD4⁺ thymocytes that had reduced TCR, CD28, and CD127 expression also had reduced Bcl-2 expression. Interestingly, the expression pattern of Bcl-2 in Gads^−/− mature CD8⁺ thymocytes suggested that there were two populations of mature SP CD8⁺ thymocytes in Gads^−/− mice, just as two populations were evident when we analyzed CD122 expression (Fig. 2D).

Fig. 7. Gads^−/− TSP CD4⁺ thymocytes haved reduced Bcl-2 expression.

Total thymocytes from Gads^+/+ and Gads^−/− mice were labeled with anti-CD4, anti-CD8, anti-CD24, anti-TCRβ, and anti-Bcl-2. In each histogram, the light lines represent the isotype control staining, the shaded histograms represent Bcl-2 staining in wild-type cells, and the dark lines represent Bcl-2 staining in Gads^−/− cells.

Discussion

The data presented here are consistent with a model of T cell development in which there are at least two distinct steps in the differentiation of DP thymocytes into mature SP thymocytes. Between the DP and mature SP developmental stages is the TSP CD4⁺ stage of development [1-3]. The block in development seen in Gads^−/− mice is at the TSP CD4⁺ stage of development, a unique site for a block in T cell development. More commonly, blocks in positive selection are seen at the DP stage of development. These data suggest that the signaling pathways regulating the differentiation of DP thymocytes are different than those that regulate the differentiation of TSP CD4⁺ thymocytes.

As seen in Fig. 1D, the ratios of DP thymocytes to ISP CD8⁺ thymocytes were identical in wild-type and Gads^−/− mice and the ratio of TSP CD4⁺ cells to ISP CD8⁺ cells was only slightly less in Gads^−/− mice than wild-type mice. However, despite the ability of Gads^−/− thymocytes to proceed to the TSP CD4⁺ step, many Gads^−/− TSP CD4⁺ thymocytes had not up-regulated TCR and CD69 expression (Figs. 1 and 2), which are hallmarks of positive selection [18-22]. Because nearly all wild-type TSP CD4⁺ thymocytes had high levels of TCR and CD69, it is likely that these aspects of positive selection normally occur prior to the TSP CD4⁺ stage and most likely in the DP stage. We conclude that Gads is required for positive selection, but positive selection is not a prerequisite for advancement to the TSP CD4⁺ stage. However, only those cells that were positively selected matured into CD24^loCD62L^hi SP thymocytes.

Like Gads^−/− mice, transgenic mice expressing ZAP-70 only in DP thymocytes could readily generate TSP CD4⁺ thymocytes [23, 24]. Expressing ZAP-70 in DP thymocytes was necessary and sufficient to trigger TCR up-regulation. By contrast, ZAP-70^−/− DP thymocytes could not up-regulate their TCR levels, nor could they differentiate into TSP CD4⁺ thymocytes. These data indicate that ZAP-70-dependent, Gads-dependent signaling pathways are required for TCR up-regulation. However, a ZAP-70-dependent, Gads-independent pathway permits the differentiation of DP thymocytes into the TSP CD4⁺ stage.

Gads^−/− thymocytes had dysregulated CD28 expression (Fig. 6A). The precise role of CD28 in positive and negative selection of thymocytes is controversial. While some data suggest that CD28 has no role in positive selection [29, 30], other evidence suggest that CD28-mediated signaling inhibits positive selection [31]. The confusion over the function of CD28 in positive and negative selection could be explained by data showing that the intensity of CD28 signaling must be tightly controlled in order to optimize the balance between differentiation and apoptosis of DP thymocytes [32]. Only a moderate level of TCR and CD28 signaling in DP thymocytes might induce Bcl-2 expression; if the level of TCR and CD28 signaling were too high or too low, Bcl-2 expression might not be induced. These observations suggest that the defect in CD28 expression seen in Gads^−/− thymocytes could contribute to the lack of adequate Bcl-2 expression in TSP CD4⁺ thymocytes (Fig. 7).

Another contributing factor to the reduced level of Bcl-2 expression seen in Gads^−/− TSP CD4⁺ thymocytes is the reduced level of CD127 expression, particularly among TCR^lo TSP CD4⁺ cells (Fig. 6B). IL-7 supports Bcl-2 expression during co-receptor reversal, the transition from TSP CD4⁺ thymocytes to SP CD8⁺ thymocytes [33, 34].

The block in positive selection seen in Gads^−/− mice was more severe when an MHC class II-restricted TCR was expressed than when an MHC class I-restricted TCR was expressed (Figs. 3-5). These data support the model in which a stronger signal is required for the development of CD4⁺ thymocytes than CD8⁺ thymocytes [25, 26]. TCR ligation of comparable affinity in Gads-expressing and Gads-deficient thymocytes is predicted to yield a much less intense signal in Gads^−/− cells, where calcium mobilization is defective [4, 5]. Thus, in Gads^−/− mice, it is likely that only thymocytes expressing a TCR with high affinity for self-antigens can up-regulate their TCR levels and proceed through development. In addition, these high affinity TCR-antigen interactions were more able to produce mature SP CD8⁺ cells than mature SP CD4⁺ cells (Fig. 1).

The nature of the block in development appeared to be slightly different in Gads^−/− OT-I and Gads^−/− P14 mice. Gads^+/+ P14 mice had a much smaller percentage of SP CD4⁺ thymocytes than Gads^+/+ OT-I mice. In Gads^−/− P14 mice, the percentage of SP CD4⁺ thymocytes was similar to Gads^+/+ P14 mice, but in Gads^−/− OT-I mice, the SP CD4⁺ thymocyte population was smaller than in Gads^+/+ OT-I mice. This difference is likely explained by the differences in the affinities of the transgenic TCRs for their MHC-peptide ligands; the affinity of the P14 TCR for its cognate antigen is approximately twice that of the OT-I TCR for its ligand [27, 28]. When TSP CD4⁺thymocytes express a low affinity TCR, the cells may not receive an adequate signal to continue maturation. The survival of these TSP CD4⁺ cells may be Gads-dependent. Because many TSP CD4⁺ thymocytes had not undergone positive selection in Gads^−/− TCR transgenic mice, few cells proceeded to the mature SP CD8⁺ stage.

In the absence of a transgenic TCR, nearly all peripheral CD8⁺ T cells in Gads^−/− mice had a memory-like phenotype [5]. Consistent with this observation, nearly half of the mature SP CD8⁺thymocytes (CD24^loTCRβ^hi) in Gads^−/− mice expressed CD122 (Fig. 1D). In wild-type mice, CD122⁺ SP CD8⁺ thymocytes represent a rare population. Elevated CD122 expression in CD8⁺T cells is a characteristic shared between activated CD8⁺ T cells and innate-like CD8⁺ T cells [16]. Thus, our CD122⁺ mature CD8⁺ thymocytes could reflect a circulating subset of memory CD8⁺ T cells or a distinct population of innate-like CD8⁺ T cells that develop within the thymus. The phenotype of Gads^−/− CD24^lo SP CD8⁺ thymocytes was similar to that seen in mice lacking Itk, Rlk, or both kinases [35-37]. Studies using Itk^−/− and Itk^−/−Rlk^−/− mice demonstrated that CD122⁺ CD8⁺ T cells were present early in the life of the mouse and developed in the thymus, rather than being activated mature T cells. These data support the biochemical data showing that Gads, Itk, and Rlk share a common signaling pathway [38]. Strikingly, crossing Gads^−/− mice to mice expressing an MHC class I-restricted TCR rescued T cell development such that splenic CD8⁺ T cells from Gads^−/− OT-I and Gads^−/− P14 mice had a naïve phenotype. (Figs. 4 and 5 and data not shown). This recovery of naïve CD8⁺ T cell production was also reported in Itk^−/− OT-I mice [35]. These data indicate that expression of an MHC class I-restricted TCR can overcome the need for the Gads-Itk signaling pathway in the development of conventional SP CD8⁺thymocytes.

In summary, our analysis of Gads^−/− mice demonstrate that positive selection of DP thymocytes into mature SP thymocytes is a multi-step process and each step is regulated by different signaling pathways. Gads-dependent signals are required for positive selection, as determined by the lack of TCR, CD69, and Bcl-2 up-regulation. However, Gads is not for the differentiation of DP thymocytes into TSP CD4⁺ thymocytes. Without Gads-dependent TCR, CD69, and Bcl-2 up-regulation, TSP CD4⁺ thymocytes fail to progress to the mature thymocyte stage of development.

Materials and Methods

Mice

The generation of C57BL/6 Gads^−/− mice was previously described [5]. C57BL/6 mice expressing the OT-I, OT-II, and P14 transgenic TCRs have been described [17, 39, 40]. All mice were housed under specific pathogen-free conditions, and all experiments were performed in compliance with the University of Kansas Medical Center Institutional Care and Use Committee. Mice were used between the ages of three and five weeks.

Antibodies

Anti-CD8-Alexa 647, anti-CD8-FITC anti-CD24-PE-Cy5, anti-TCRβ-PE, anti-TCRβ-APC, anti-TCR Vβ5-FITC, anti-TCR Vβ8-FITC, anti-TCR Vα2-PE, anti-CD69-FITC, anti-CD62LPE-Cy7, anti-CD28-PE, anti-CD127-PE, anti-CD132-PE, and anti-Bcl-2-PE were purchased from BD Biosciences (San Diego, CA) or eBioscience (San Diego, CA). Anti-CD4-Pacific Blue was purchased from Invitrogen (Carlsbad, CA).

Surface and intracellular staining and flow cytometry

Single cell suspensions of thymocytes were isolated by mincing the tissue on a wire screen and filtering through 100 micron nylon mesh. Surface and intracellular staining was performed as previously described [15]. Briefly, cells were labeled in phosphate-buffered saline (PBS) containing 2% Fetal Clone I (HyClone Laboratories, Inc., Logan, UT) and 3 mM sodium azide. After washing, cells were resuspended in 1% paraformaldehyde in PBS. For intracellular staining, cells were permeabilized in PBS containing 10% Fetal Clone I and 0.1% saponin for 15 minutes. Cells were stained and washed in the saponin-containing buffer. Flow cytometry studies were performed using a BD LSR II (BD Immunocytometry Systems, San Jose, CA). Data were analyzed using BD FACSDiva software (BD Biosciences, San Jose, CA).

Statistics

Statistics were performed using Student's T test, except where indicated.

Abbreviations

ISP

immature single positive

DN

double negative

DP

double positive

MFI

mean fluorescence intensity

SP

single positive

TSP

transitional single positive