Roles of the Src Tyrosine Kinases Lck and Fyn in Regulating γδTCR Signal Strength

Lck and Fyn, members of the Src family of tyrosine kinases, are key components of the αβTCR-coupled signaling pathway. While it is generally accepted that both Lck and Fyn positively regulate signal transduction by the αβTCR, recent studies have shown that Lck and Fyn have distinct functions in this signaling pathway, with Lck being a positive regulator and Fyn being a negative regulator of αβTCR signal transduction. To determine whether Lck and Fyn also differentially regulate γδTCR signal transduction, we analyzed γδ T cell development and function in mice with reduced Lck or Fyn expression levels. We found that reducing Lck or Fyn levels altered the strength of the γδTCR signaling response, with low levels of Lck weakening γδTCR signal strength and low levels of Fyn augmenting γδTCR signal strength. These alterations in γδTCR signal strength had profound effects not only on αβ/γδ lineage choice, but also on γδ thymocyte maturation and γδ T cell effector function. These results indicate that the cellular levels of Lck and Fyn play a role in regulating the strength of the γδTCR signaling response at different stages in the life of the γδ T cell.


Introduction
Signaling by the TCR is required at multiple stages in the life of a T cell. In the thymus, TCR signaling is necessary for lineage commitment and repertoire selection, while in the periphery, TCR signaling is necessary for maintenance of the peripheral T cell pool and for activation and differentiation of mature T cells. Lck and Fyn, two members of the Src family of tyrosine kinases (SFKs), are involved in initiating the TCR-coupled signaling cascade [1,2]. Following TCR engagement, Lck and/or Fyn phosphorylate the tyrosines within the ITAMs of the CD3 and TCRf chains. This proximal signaling event leads to the recruitment of other signaling molecules to the TCR signaling complex and to the subsequent activation of signaling pathways that ultimately lead to the nucleus and initiation of gene transcription.
It is generally accepted that both Lck and Fyn positively regulate signal transduction by the abTCR because, in the absence of either one of these SFKs, abTCR signaling responses are impaired following anti-CD3 mAb stimulation [3][4][5][6][7][8][9]. However, it has also been shown that Lck and Fyn localize to different subcellular compartments [10,11] and have different substrates [11,12], suggesting that they have discrete functions during ab T cell activation. This idea is supported by the disparate phenotypes of Lck-and Fyn-deficient mice. In Lck 2/2 mice, thymus cellularity is severely reduced, thymocyte development is almost completely blocked at the CD4 + CD8 + (double-positive; DP) stage, and very few mature abTCR + cells are detected in peripheral lymphoid tissues [12][13][14]. In contrast, Fyn 2/2 mice exhibit a mild defect in ab T cell development, as shown by the fact that Fyn 2/2 thymocytes, when in vitro stimulated, do not flux calcium or proliferate as well as wild-type (WT) thymocytes [3,5]. Despite this signaling defect in the thymus, equivalent numbers of ab T cells are found in the periphery of Fyn 2/2 and WT mice [3,5].
While recent studies have confirmed that Lck functions primarily as a positive regulator of abTCR signaling [15][16][17], evidence is accumulating in support of Fyn acting as a negative regulator of abTCR signaling. First, it has been shown that Fyn is responsible for phosphorylating the adaptor protein, phosphoprotein associated with glycolipid-enriched membranes or PAG, in both resting thymocytes and T cells [11]. Once phosphorylated, PAG then recruits Csk, an inhibitor of SFKs [18]. Recruitment of Csk to phosphorylated PAG is required for optimal Csk kinase activity because, in the absence of Fyn, there is reduced phosphorylation of PAG and reduced Csk kinase activity [11]. Therefore, by indirectly controlling the activity of the inhibitor Csk, Fyn may negatively regulate the activation threshold of ab T cells [11]. It has also been shown that CD8 + T cells from Fyn 2/2 F5 abTCR Tg mice are hyperresponsive in comparison to CD8 + T cells from WT F5 abTCR Tg mice following in vitro stimulation with peptide and APCs [19]. This hyperresponsiveness is manifested as enhanced proliferation, increased IL-2 production and more effective cytolytic activity [19]. CD4 + T cells from Fyn 2/2 DO11.10 abTCR Tg mice, however, do not display increased proliferation compared to CD4 + T cells from WT DO11.10 abTCR Tg mice when stimulated, either in vitro or in vivo, with peptide and APCs [20]. Nonetheless, when activated under the appropriate priming conditions, CD4 + T cells from Fyn 2/2 DO11.10 abTCR Tg mice produce significantly more IL-4 or IFNc than CD4 + T cells from WT DO11.10 abTCR Tg mice [20]. Taken together, these findings suggest that Fyn negatively regulates the abTCR signaling response.
Since studies investigating the functions of Lck and Fyn have focused primarily on ab T cells, it is not known whether their functional dichotomy is observed in only ab T lineage cells or in both ab and cd T lineage cells. Analyses of Lck 2/2 and Lck 2/2 Fyn 2/2 mice have in fact revealed differences in the requirements for these SFKs in ab and cd T cell development. In Lck 2/2 mice, the number of thymic and peripheral cd T cells is only modestly reduced compared to their numbers in WT mice [13,14,21]. Moreover, in Lck 2/2 Fyn 2/2 mice, in which ab T cell development is completely abrogated, a small number of cdTCR + cells do develop and can be detected in secondary lymphoid tissues, the small intestine, and the epidermis [21,22]. These differential requirements for Lck and Fyn in ab and cd T cell development suggest that these SFKs may have different functions in aband cdTCR signal transduction. To investigate this, we evaluated the individual roles of Lck and Fyn in the development and function of cd lineage cells. Here, we report that Lck and Fyn expression levels vary in cd lineage cells depending on their stage in development, with thymic cd T cells expressing relatively high levels of Lck and Fyn and peripheral cd T cells expressing relatively low levels of Lck and Fyn. These differences in the cellular levels of Lck and Fyn play a role in regulating the strength of the cdTCR signaling response at the different developmental stages because, when we reduced Lck or Fyn expression levels by using Lck +/2 and Fyn +/2 mice, we observed significant effects on ab/cd lineage choice, cd thymocyte maturation, and cd T cell effector function. Moreover, because reducing the levels of Lck or Fyn altered the cdTCR signaling response, such that low Lck levels weakened cdTCR signal strength and low Fyn levels augmented cdTCR signal strength, we conclude that Lck and Fyn have similar functions in aband cdTCR signal transduction, with Lck serving to amplify the TCR signal and Fyn serving to dampen the TCR signal.

Expression Pattern of Lck and Fyn in cd Lineage Cells
Although it is generally accepted that cd T cells express Lck and Fyn, this idea is based more on indirect evidence from studies investigating cd T cell development in Lck-or Fyn-deficient mice [13,14,21,23,24] than on a direct demonstration of expression [25]. To resolve this, we developed an intracellular (i.c.) flow cytometric assay to measure and compare the relative levels of Lck and Fyn in ab and cd lineage cells from wild-type (WT) mice. Using Lck 2/2 and Fyn 2/2 cells as negative staining controls, we found that both Lck and Fyn are expressed in DN cd thymocytes and peripheral DN cd T cells (Fig. 1A). On average, DN cd thymocytes expressed Lck and Fyn at higher levels than DP and mature CD4 + thymocytes, whereas peripheral DN cd T cells expressed Lck at levels comparable to those in CD4 + T cells and Fyn at levels lower than those in CD4 + T cells (Fig. 1A). When Lck and Fyn expression levels were compared between thymic and peripheral cd T cells, we found that both SFKs are expressed at significantly higher levels in DN cd thymocytes than in DN cd T cells ( Fig. 1B-C). This finding suggested that the expression levels of both Lck and Fyn are down-regulated once cd T cells emigrate from the thymus to the secondary lymphoid organs. This same phenomenon was also observed for ab lineage cells; however, the degree of reduction in Lck and Fyn expression levels between the thymus and LN was greater for cd lineage cells than ab lineage cells ( Fig. 1C and data not shown). Taken together, these data indicated that immature and mature cd lineage cells express Lck and Fyn and that the expression of these SFKs is dynamic during cd T cell development and maturation.
Polyclonal cd T Cell Development in Lck +/2 and Fyn +/2 Mice Because DN cd thymocytes expressed higher levels of Lck and Fyn than mature DN cd T cells, we sought to determine whether high levels of Lck or Fyn were required for cd lineage commitment and/or development in the thymus. To investigate this, we reduced the expression levels of Lck or Fyn during T cell development by using Lck +/2 and Fyn +/2 mice. To verify that protein expression was reduced in the heterozygous mice, we compared the relative expression levels of Lck and Fyn in immature and mature CD4 + lineage cells from Lck +/2 and Fyn +/2 mice with those from WT mice. As expected, we observed a 50% reduction in Lck expression levels in Lck +/2 mice and a 50% reduction in Fyn expression levels in Fyn +/2 mice ( Fig. 2A). In addition, there was no compensatory increase in the expression of one SFK when expression of the other SFK was reduced ( Fig. 2A).
When cd T cell development was analyzed in Lck +/2 and Fyn +/2 mice, we observed no significant difference in the number of DN cd TCR + cells in the thymus and lymph nodes (LNs) of these mice compared to WT mice (Fig. 2B). Moreover, phenotypic analysis of the DN cd thymocytes and mature DN cd T cells from WT, Lck +/2 and Fyn +/2 mice revealed no appreciable differences in Vc usage, TCRcd surface levels and cell surface phenotype (data not shown), indicating that reducing the expression levels of Lck or Fyn resulted in no apparent defect in the development of polyclonal cd T cells.
Because cd T cell development appeared not to be affected in Lck +/2 and Fyn +/2 mice, we examined the levels of Lck and Fyn in cd lineage cells of the heterozygous mice to determine whether they were reduced by 50% as they were in ab lineage cells ( Fig. 2A and C). We found that Lck levels were indeed reduced by 50% in thymic and peripheral cd lineage cells from Lck +/2 mice (Fig. 2C). Fyn levels, on the other hand, were reduced to ,60% in immature and mature cd lineage cells from Fyn +/2 mice, which is significantly different from the expected 50% (Fig. 2C). Since Fyn expression levels in cd lineage cells were not reduced to the expected 50% in Fyn +/2 mice, we determined whether the thymic precursors in Lck +/2 and Fyn +/2 mice displayed a 50% reduction in Lck and Fyn expression levels, respectively. To accomplish this, we compared Lck and Fyn expression levels in lineage-negative CD44 + CD25 + (DN2) thymocytes from WT, Lck +/2 and Fyn +/2 mice, as this thymocyte subset contains precursors that have the potential to develop into ab or cd lineage cells [26,27]. We found that, while Lck expression levels were reduced to 50% in Lck +/2 DN2 thymocytes, Fyn expression levels were only reduced to 85% in Fyn +/2 DN2 thymocytes ( Fig. 2D and E). These data demonstrated that although Fyn levels are reduced in thymic precursors and cd lineage cells from Fyn +/2 mice, they are not reduced by 50% as they are in ab lineage cells.

Effect of Reducing Lck or Fyn Levels on the Commitment and Development of cd Lineage Cells
Another reason why we may not have observed any defects in cd T cell development in Lck +/2 and Fyn +/2 mice is because thymocyte development and selection are able to compensate for alterations in cdTCR signal transduction, which may result from reductions in Lck or Fyn expression. To address this, we mated a cdTCR transgene onto Lck +/2 and Fyn +/2 genetic backgrounds to determine whether fixing the specificity of the cdTCR revealed defects in ab/cd lineage choice and/or cd T cell development. For Staining of cells from Lck 2/2 and Fyn 2/2 mice are shown as negative controls for i.c. staining of Lck and Fyn, respectively. B. Comparison of the relative expression levels of Lck and Fyn in gated DN cdTCR + thymocytes and LN cells. C. Quantifying the change in the relative expression levels of Lck and Fyn in DN cdTCR + thymocytes and LN cells and, for comparison, CD4 + CD3 + thymocytes and LN cells. Lck and Fyn expression levels in immature and mature subsets were normalized to those of DP thymocytes, as this population had, in every experiment, consistently lower levels of Lck and Fyn than any other thymocyte or T cell subset (see A). Data are presented as fold change relative to DP thymocytes (set to 1). Data are representative of at least 6 independent experiments. Bars represent mean 6 SEM. *p#0.05, **p#0.01, #p#0.001. doi:10.1371/journal.pone.0008899.g001 these experiments, we used the Vc6/Jc1/Cc1 and Vd1/Dd1/ Jd2/Cd transgenic (cdTCR Tg) mouse [28], which we have previously used to study ab/cd lineage choice, cd T cell development and cdTCR signal transduction [29][30][31]. It is important to note that, although Vc6/Vd1 + cd T cells are only generated in the fetal thymus of a WT (non-cdTCR Tg) mouse [28], the Vc6/Vd1 + cd T cells generated in the cdTCR Tg mouse represent adult cd T cells, as they express the panel of cd-biased genes [31] typical of adult but not fetal cd T cell populations [32]. Moreover, one of the advantages of using this cdTCR Tg mouse model to study ab/cd lineage choice is that the rearranged TCRc and -d chains are expressed early during T cell development prior to TCRb expression and, as a consequence, the ab/cd lineage decision is mediated exclusively by the cdTCR [29].
By fixing the specificity of the cdTCR, we observed significant effects on the ab/cd lineage fate decision when the levels of Lck but not Fyn were reduced. In WT cdTCR Tg mice, equivalent numbers of DN cdTCR + thymocytes (cd lineage) and DP thymocytes (ab lineage cells) are generated ( Fig. 3A-B). Reducing Lck expression resulted in a striking 4-fold increase in thymus cell number compared to WT cdTCR Tg mice, most likely due to the significant increase in the percentage of DP thymocytes (Fig. 3A). Consequently, the number of ab lineage cells in Lck +/2 cdTCR Tg mice was significantly higher than that in WT cdTCR Tg mice (Fig. 3B). These findings are consistent with those of a previous study, in which cdTCR-dependent generation of DP thymocytes but not of DN cdTCR + thymocytes was observed in the absence of Lck expression [33]. In contrast, reducing Fyn levels had no effect on total thymus cell number. Nonetheless, we did observe a decrease in both the percentage and number of DP thymocytes in Fyn +/2 cdTCR Tg mice, although these differences were not statistically significant from the percentage and number of DP E. Quantifying the reduction of Lck and Fyn expression levels in DN2 thymocytes from Lck +/2 and Fyn +/2 mice. The MFI of the i.c. levels of Lck and Fyn in DN2 thymocytes from heterozygous mice are expressed as a percentage of the MFI of the i.c. levels of Lck and Fyn in DN2 thymocytes from WT mice. A dashed line marks the expected 50% reduction in WT Lck and Fyn levels. In A, B, C, and E, the bars represent mean 6 SEM. *p#0.05, **p#0.01, #p#0.001. doi:10.1371/journal.pone.0008899.g002 thymocytes in WT cdTCR Tg mice ( Fig. 3A-B). Interestingly, despite the changes in the numbers of DP thymocytes in the heterozygous mice, the numbers of DN cdTCR + thymocytes, or cd lineage cells, in Lck +/2 cdTCR Tg and Fyn +/2 cdTCR Tg mice were comparable to their number in WT cdTCR Tg mice ( Fig. 3A-B).
It has previously been shown that genetic manipulation of cdTCR signal strength affects ab/cd lineage choice in a consistent manner. Namely, when the cdTCR signaling response is strengthened, cd lineage fate is favored and, conversely, when the cdTCR signaling response is weakened, ab lineage fate is favored [29,33]. Our finding that the number of ab lineage cells was significantly increased in Lck +/2 cdTCR Tg mice compared to WT cdTCR Tg mice suggested that reducing Lck levels weakened cdTCR signal strength. Surprisingly, this increase in the number of DP thymocytes was not accompanied by a corresponding decrease in the number of DN cd thymocytes in Lck +/2 cdTCR Tg (Fig. 3B). We reasoned that the number of DN cd thymocytes in Lck +/2 cdTCR Tg mice may not reflect the number of thymocytes that adopted the cd fate but instead reflected an expansion of the thymocytes that already adopted the cd fate. To investigate this, we compared the proliferative status of DN cd thymocytes in WT cdTCR Tg, Lck +/2 cdTCR Tg, and Fyn +/2 cdTCR Tg mice by measuring their expression of the Ki-67 Ag, which is a marker of actively cycling cells [34,35]. We found that the frequency of Ki67 + DN cd thymocytes in Lck +/2 cdTCR Tg mice was significantly higher than the frequency of Ki67 + DN cd thymocytes in WT cdTCR Tg and Fyn +/2 cdTCR Tg mice (Table 1). These data suggested that the number of DN cd thymocytes in Lck +/2 cdTCR Tg mice does not reflect the number of thymocytes that adopted the cd fate.
Next, we compared the expression levels of the cdTCR and CD5 on the surface of DN cd thymocytes from WT cdTCR Tg, Lck +/2 cdTCR Tg, and Fyn +/2 cdTCR Tg mice to gauge the effects of reducing Lck or Fyn levels on the phenotype of the cells choosing the cd lineage. When we examined cdTCR and CD5 surface levels on the DN cd thymocytes that were generated in WT cdTCR Tg, Lck +/2 cdTCR Tg and Fyn +/2 cdTCR Tg mice, we found that DN cd thymocytes in the three genotypes expressed different levels of the cdTCR but equivalent levels of CD5. Specifically, Lck +/2 cd thymocytes expressed significantly higher levels of the cdTCR than WT cd thymocytes, whereas Fyn +/2 cd thymocytes expressed significantly lower levels of the cdTCR than WT cd thymocytes (Fig. 4A). These data suggested that reducing Lck expression levels weakened cdTCR signal strength to the extent that immature thymocytes expressing relatively high levels of the cdTCR adopted the cd fate. Conversely, reducing Fyn expression levels augmented cdTCR signal strength to where immature thymocytes expressing relatively low levels of the cdTCR adopted the cd fate.
To investigate the effects of reducing the levels of Lck or Fyn on the maturation of cd T cells in the thymus and their subsequent ability to migrate to the periphery, we enumerated DN cd T cells in the LNs of WT cdTCR Tg, Lck +/2 cdTCR Tg, and Fyn +/2 cdTCR Tg mice. Compared to WT cdTCR Tg mice, we observed a significant decrease in the number of DN cd T cells in Lck +/2 cdTCR Tg mice but not in the number of DN cd T cells in Fyn +/2 cdTCR Tg mice (Fig. 3C). Consistent with the lower numbers of peripheral cd T cells in Lck +/2 cdTCR Tg mice was the finding that there were fewer DN cd T cells in these mice that expressed CD24, a marker of recent thymic emigrant cd T cells [36], and that there were more cells that expressed CD44, a marker of activated cells, memory cells, and/or cells undergoing homeostatic proliferation [36][37][38] (Fig. 4B). Moreover, we found that the cdTCR and CD5 surface levels that were noted among the DN cdTCR + thymocytes from the three genotypes were maintained on their respective peripheral cd T cells, with the exception that Lck +/2 cd T cells expressed significantly lower levels of CD5 than WT cd T cells (Fig. 4C). Therefore, although DN cd thymocytes  were generated in Lck +/2 cdTCR Tg mice in numbers comparable to WT cdTCR Tg mice, these mice had reduced numbers of mature DN cd T cells. These data indicated that a reduction in Lck levels but not Fyn levels affects the maturation and/or survival of thymic cd T cells.

Effect of Reducing Lck or Fyn Levels on cd T Cell Effector Fate and Function
cd T cell effector fate has been shown to segregate with expression of specific surface antigens, specifically CD122 + and/or CD27 + cd T cells preferentially produce IFNc [39,40], whereas IL-23R + cd T cells preferentially produce IL-17 [41][42][43]. Given these findings, we sought to determine whether reducing the levels of Lck or Fyn altered the ability of a cd T cell to become an IL-17and/or IFNc-producing effector cell. To accomplish this, we chose to use our cdTCR Tg mouse model as it generates, on the WT background, 30-fold more DN cd T cells than non-cdTCR Tg mice [30]. This means that DN cd T cells can be analyzed without the concern that purification by positive selection using an anti-TCRcd mAb may crosslink the cdTCR and, in turn, pre-activate the cd T cell. First, we determined whether there were any differences in the percentages of cd T cells expressing CD122, CD27 or IL-23R among WT cdTCR Tg, Lck +/2 cdTCR Tg, and Fyn +/2 cdTCR Tg mice. We found that both Lck +/2 cdTCR Tg and Fyn +/2 cdTCR Tg mice had similar percentages of CD27 + and CD27 2 cd T cells as WT cdTCR Tg mice, but the percentages of CD27 + cells co-expressing CD122 + in both Lck +/2 cdTCR Tg and Fyn +/2 cdTCR Tg mice was reduced compared to WT cdTCR Tg mice (Fig. 5A). As CD122 expression by cd T cells is induced when the cdTCR interacts with its ligand in the thymus [39], these results suggested that the selection and/or survival of CD122 + cd T cells is impaired when Lck or Fyn levels are reduced. To evaluate IL-23R expression among WT, Lck +/2 , and Fyn +/2 cd T cells, we performed quantitative real-time RT-PCR analysis to detect transcription of IL12RB1 and IL23R, which encode the two subunits of the IL-23R [44]. No significant differences were noted in the relative amounts of IL12RB1 and IL23R transcripts among WT, Lck +/2 , and Fyn +/2 cd T cells (Fig. 5B), indicating that the selection and/or survival of cd T cells with the potential to produce IL-17 is not affected when Lck or Fyn levels are reduced.
Next, we assessed cytokine production by WT, Lck +/2 , and Fyn +/2 cd T cells following CD3 crosslinking. Interestingly, we found that the percentages of WT, Lck +/2 and Fyn +/2 cd T cells producing IFNc at 16 h were equivalent to the percentages of CD122 + CD27 + cd T cells in each mouse ( Fig. 5A and C). It is also important to note that the level of IFNc production, as measured by MFI, was 2 to 3-fold less in Lck +/2 and Fyn +/2 IFNc + cd T cells than in WT IFNc + cd T cells (Fig. 5C and data not shown). Moreover, when we compared the ability of WT, Lck +/2 and Fyn +/2 cd T cells to differentiate into IL-17-producing cells at 16 h, we detected considerably fewer Fyn +/2 IL-17 + cd T cells than Lck +/2 or WT IL-17 + cd T cells (Fig. 5C). Taken together, these data indicated that reducing the levels of Fyn impacts the function of cd T cells that have the potential to become either IL-17-or IFNc-producing effector cells, while reducing the levels of Lck only impacts the function of cd T cells that have the potential to become IFNc-producing effector cells.
Because Lck +/2 and Fyn +/2 cd T cells do not efficiently produce IFNc, it was of interest to determine whether reducing Lck or Fyn levels also affected the ability of cd T cells to produce other cytokines. To test this, we evaluated TNFa production by cd T cells from Lck +/2 cdTCR Tg and Fyn +/2 cdTCR Tg mice, since cd T cells, including those that produce IFNc, have been shown to produce this cytokine [40]. As shown in Fig. 5D, we found that more cd T cells from Lck +/2 cdTCR Tg and Fyn +/2 cdTCR Tg mice than from WT cdTCR Tg mice were producing TNFa. These data indicated that reducing Lck or Fyn levels does not impair the ability of cd T cells to produce TNFa and suggested that the TCR signals required to activate the genetic program for IFNc production are different than those for TNFa production.

Discussion
Since SFKs have both positive and negative roles in receptor signaling, it has been postulated that these kinases function more like rheostats than on/off switches [45]. Our data support this idea, as changes in the cellular levels of Lck or Fyn at different stages in the life of a cd lineage cell affected the strength of the cdTCR signaling response and, in turn, affected ab/cd lineage commitment, cd T cell maturation and cd T effector cell differentiation.
The expression levels of Lck and Fyn change during T cell development and maturation. Immature thymocytes (i.e., DN2 thymocytes), which have the potential to become either ab or cd lineage cells [26,27], expressed relatively high levels of both Lck and Fyn. In thymic cd lineage cells, these high levels of Lck and Fyn were maintained and, not until the cd lineage cells were exported from the thymus, did their Lck and Fyn expression levels decrease. However, in immature ab lineage cells, Lck and Fyn expression levels dramatically declined and, at the DP stage, their levels of Lck and Fyn were extremely low. The low SFK expression in DP thymocytes has also been reported by Olszowy et al. [46], who used quantitative Western blot analysis to measure Lck and Fyn protein levels in thymocyte subsets. Interestingly, Lck and Fyn levels were increased in TCR hi DP (data not shown) and SP thymocytes compared to DP thymocytes, suggesting that positive selection upregulated both Lck and Fyn expression. It is important to note that mature SP thymocytes, after leaving the thymus, downregulated Lck and Fyn expression levels, but not to the levels observed for DP thymocytes nor to the extent observed between thymic and peripheral cd T cells.
To investigate the importance of the quantitative difference in Lck and Fyn expression levels between thymic and peripheral cd T cells, we used Lck +/2 and Fyn +/2 mice to study the effect of reducing Lck or Fyn expression levels on cd T cell development and function. We chose to reduce, as opposed to eliminate, Lck and Fyn expression levels to prevent any compensatory action that one SFK may exhibit in the absence of the other. Although Fyn levels were reduced by 50% in ab lineage cells from Fyn +/2 mice, they were only reduced by 40% in cd lineage cells from the same mice. There are two possible explanations, which are not mutually exclusive, for why Fyn levels were not reduced to the expected 50% in cd lineage cells from Fyn +/2 mice. First, the relatively high levels of Fyn in Fyn +/2 cd thymocytes may be a result of selection, where only cells with high levels of Fyn survive and continue to mature. The second possibility is that the high Fyn levels in cd lineage cells reflect high Fyn levels in a precursor population, such as DN2 thymocytes, which have the developmental potential to give rise to ab and cd lineage cells [26,27]. Indeed, we found that Fyn expression was only reduced by ,15% in DN2 thymocytes from Fyn +/2 mice. Therefore, even though we cannot rule out selection of cd thymocytes with high levels of Fyn, it is conceivable that the relatively high levels of Fyn in cd thymocytes from Fyn +/2 mice may be a direct result of the high levels of Fyn in Fyn +/2 DN2 thymocytes.
The high level of Fyn in thymic precursors highlights the importance of Fyn activity during an early stage of T cell development. As surface TCR complexes are not expressed at this stage, it is possible that Fyn is required for signaling through other receptors. One such receptor may be the IL-7 receptor (IL-7R), as Fyn has been shown to be recruited to this receptor [47,48]. Given this association and that DN2 thymocytes require IL-7R expression and signaling for their survival and proliferation [49][50][51][52][53], it is possible that relatively high levels of Fyn are required for proper IL-7R signaling at this stage.
The first stage in T cell development where we observed regulation of cdTCR signal strength by Lck and Fyn is during ab/cd lineage commitment. We and others have previously demonstrated that TCR signal strength influences the ab/cd lineage decision, with a strong signal favoring cd lineage commitment and a weak signal favoring ab lineage commitment [29,33,54]. By fixing the specificity of the cdTCR, we were able to detect changes in the cdTCR signal response that were not apparent with a polyclonal cdTCR repertoire. Reducing the expression of Lck weakened cdTCR signal strength and resulted in a striking increase in the percentage and number of DP thymocytes in Lck +/2 cdTCR Tg mice compared to WT cdTCR Tg mice. Moreover, weakening of the cdTCR signal was confirmed by the finding that DN cd thymocytes from Lck +/2 cdTCR Tg mice expressed higher levels of the cdTCR than WT DN cd thymocytes. Together, these results indicated that relatively high levels of Lck are required to achieve the appropriate TCR signal response to support the cd lineage choice. Conversely, reducing Fyn expression levels strengthened the cdTCR signaling response, as evidenced by the decrease, albeit not significant, in the number of DP thymocytes and the significant decrease in cdTCR surface expression on DN cd thymocytes from Fyn +/2 cdTCR Tg mice. In contrast to the results of our previous study, in which cd T cell fate was favored over ab T cell fate when cdTCR signal strength was augmented [29], reducing Fyn expression had modest effects on the generation of ab and cd lineage cells. This difference may be attributed to the fact that Fyn expression was reduced by 35% in cd thymocytes from Fyn +/2 cdTCR Tg mice (data not shown), suggesting that immature DN thymocytes expressing low levels of Fyn were unable to survive and develop into ab or cd lineage cells.
Our data also demonstrated that the alterations in cdTCR signal strength by reducing Lck or Fyn expression levels affected cd T cell maturation in the thymus. When cdTCR signal strength was weakened by reducing Lck levels, there was a significant decrease in the number of DN cd T cells in LNs despite normal numbers of cd lineage cells present in the thymus. In support of a maturation defect, the frequency of recent thymic emigrants (CD24 + DN cdTCR + ) was reduced in the LNs of Lck +/2 cdTCR Tg mice. Because of the reduction in thymic output, Lck +/2 cd T cells seemingly underwent homeostatic proliferation in the periphery, evidenced by the increased frequency of cd T cells with an activated/memory phenotype (CD44 + DN cdTCR + ). On the other hand, strengthening the cdTCR signal response by reducing Fyn levels had little effect on thymic maturation. These data not only demonstrate that signaling by the cdTCR is required following cd lineage commitment for cd thymocyte maturation, but also indicate that the cellular levels of Lck in cd thymocytes regulate the cdTCR signaling response at this developmental stage.
Unlike ab thymocytes, cd thymocytes do not need to encounter Ag to mature and emigrate to the periphery [39,55]. However, a recent study demonstrates that Ag encounter in the thymus by cd T cells controls their effector fate, with Ag-experienced cd T cells expressing CD122 and preferentially producing IFNc and Agnaïve cd T cells lacking CD122 expression and preferentially producing IL-17 [39]. In addition to CD122 expression, cd T cell effector fate has been shown to segregate with CD27 expression, in that CD27 + cd T cells produce IFNc and CD27 2 cd T cells produce IL-17 [40]. We found that either weakening or augmenting the cdTCR signal response had an impact on the generation and/or survival of CD122 + CD27 + cd T cells but not on the generation and/or survival of CD122 2 CD27 + cd T cells. Unfortunately, we do not know whether the defect lies in CD122 + cd T cell selection, survival or both, as very few (,0.5%) CD122 + cd T cells are detected in the thymus of WT cdTCR Tg, Lck +/2 cdTCR Tg, and Fyn +/2 cdTCR Tg mice (data not shown). Since Vc6/Vd1 + cd T cells are normally generated in the fetal thymus [28], the paucity of CD122 + cd thymocytes in WT cdTCR Tg mice may be due to decreased or negligible expression of the ligand for the Vc6/Vd1 cdTCR in the adult thymus. Nonetheless, we propose that weakening the cdTCR signal results in the generation of fewer CD122+ cd thymocytes following Ag encounter in the thymus. However, due to homeostatic proliferation in the periphery, the size of the CD122 + cd T cell pool is consequently increased. When the cdTCR signal is augmented, on the other hand, Ag encounter in the thymus may lead to negative selection and, in turn, fewer cd thymocytes and peripheral cd T cells expressing CD122. Surprisingly, in regards to cd T cells with the potential to produce IL-17, we found that neither weakening nor augmenting cdTCR signal strength had an impact on their generation and/or survival, based on the comparable expression of IL23R and IL12RB1 in cd T cells from the three genotypes. These results indicate that cd T cells destined to become IL-17 producers, unlike those destined to become IFNc producers, are not dependent on Lck-or Fyn-mediated signaling for their generation and/or survival, and suggest that the choice to become an IFNc effector versus an IL-17 effector is not a binary fate decision.
Consistent with their being multiple cd effector fates is our finding that, following TCR stimulation, 10% of WT cd T cells produce IFNc, 5% produce IL-17 and 30% produce TNFa [a third of these also produce IFNc, while none produce IL-17 (data not shown)]. Interestingly, the frequency of CD122 + cd T cells but not CD27 + cd T cells was found to be an indicator of IFNc production by cd T cells, regardless of whether the stimulated cd T cells were from WT cdTCR Tg, Lck +/2 cdTCR Tg, and Fyn +/2 cdTCR Tg mice. While our finding is consistent with the findings of the recent study demonstrating that CD122 expression correlates with IFNc production, it is not consistent with those of the other study demonstrating that CD27 expression correlates with IFNc production. An explanation for the difference in results may be how, in each study, cd T cells were stimulated. In our study, as well as that of Jensen et al. [38], cd T cells were stimulated by crosslinking the TCR with anti-CD3 or anti-TCRd mAbs. However, in the study of Ribot et al. [39], cd T cells were first activated and expanded in the presence of anti-CD3 mAb and IL-2 for four days prior to re-stimulation with the phorbol ester, phorbol 12-myristate 13-acetate, and ionomycin. These data, taken together, suggest that CD122 + cd T cells rapidly produce IFNc following TCR activation, whereas CD27 + cd T cells require multiple days post TCR activation to differentiate into IFNc-producing cells. Notably, both weakening and augmenting cdTCR signal strength affected the proficiency by which cd T cells produced IFNc but not TNFa. This finding indicates that cdTCR signal strength regulates the quality of the cytokine effector response of stimulated cd T cells. Moreover, having both weak and strong TCR signals leading to a similar outcome is not novel in T cell biology, as for example it well known that DP thymocytes that receive TCR signals that are too weak or too strong undergo apoptosis [reviewed in 56].
In summary, we have demonstrated that the cellular levels of Lck or Fyn regulate the strength of the cdTCR signaling response. Specifically, we found that reducing Lck levels, thereby weakening cdTCR signal strength, had profound effects on ab/cd lineage commitment, cd thymocyte maturation, and the generation of IFNc-producing effectors. In contrast, when Fyn levels were reduced, thus augmenting cdTCR signal strength, we observed defects in the generation of both IL-17-and IFNc-producing effectors. Thus, this study has revealed that a relatively strong cdTCR signaling response is required following cd lineage commitment for cd T cell maturation and cd T effector cell differentiation.

Ethics Statement
All research involving animals has been conducted according to the relevant national and international guidelines with respect to husbandry, experimentation and welfare. Mouse protocols were approved by the SUNY Upstate Medical University Committee on the Humane Use of Animals.

Flow Cytometry
Flow cytometric analysis for surface antigens was performed by pre-incubating cells with the anti-CD16/CD32 mAb for at least 10 min to block non-specific binding of immunoglobulins to Fc receptors followed by staining with fluorochrome-conjugated Abs against the various surface antigens. I.c. staining for Ki-67 was performed according to the manufacturer's instructions (BD Pharmingen). I.c. staining for Lck and Fyn in addition to IL-17A, TNFa and IFNc was performed by first fixing cells in a final concentration of 1.5% formaldehyde for 10 min at 37uC. Fixed cells were then stained for surface antigens, permeabilized with Perm/Wash Buffer (BD Pharmingen) for 20 min at 4uC, and then stained with mAbs against the appropriate proteins. The mAbs specific for IL-17A, TNFa and IFNc were directly conjugated to a fluorochrome, while FITC-conjugated anti-mouse IgG 2b Ab was used as a secondary reagent to detect the anti-Lck mAb or the anti-Fyn mAb. For all experiments, 0.1-2610 6 cells were collected on a LSR II using FACSDiva software (BD Immunocytometry Systems, San Jose, CA, USA) and analyzed using FlowJo software (Tree Star Inc., Ashland, OR, USA). Dead cells were excluded from analysis based on forward and side scatter profiles.

Cell Separation
DN cd T cells were purified by negative selection from the LNs of WT cdTCR Tg, Lck +/2 cdTCR Tg and Fyn +/2 cdTCR Tg mice using the MACS magnetic bead separation system (Miltenyi Biotec, Auburn, CA, USA). LN cells were stained for 10 min at 4uC with a panel of FITC-labeled mAbs containing anti-CD19, anti-TCRb, anti-CD4, anti-CD8, anti-IA b and anti-DX5 mAbs. Cells were washed, incubated with anti-FITC MACS beads for 15 min at 4uC, and then separated on an autoMACS cell separator, according to manufacturer's directions. The purity of the resulting DN cd T cell populations were typically $99%.

RT-PCR Analysis
RNA was extracted from purified DN cd T cells using the Qiagen RNeasy kit (Valencia, CA, USA). cDNA was then synthesized using Invitrogen's SuperScriptH First-Strand Synthesis System. Quantitative real-time RT-PCR analysis was performed using a Bio-Rad iQ TM 5 Real-time PCR machine (Hercules, CA, USA) according to manufacturer's directions. All of the primer sets for the quantitative real-time RT-PCR analysis, which included GAPDH, IL12RB1, and IL23R, along with the SYBR Green PCR Master Mix were purchased from SABiosciences (Frederick, MD, USA).

Statistical Analysis
Data are presented as mean 6 SEM. Student's t-test was used for all statistical comparisons (Graph Pad Prism or Microsoft Excel software) except for the one evaluating Fyn levels in ab lineage cells, cd lineage cells and DN2 thymocytes from Fyn +/2 mice, in which a x 2 test (Microsoft Excel software) was used. p values less than or equal to 0.05 were considered statistically significant.